TagTime/sourcery.html at master · tagtime/TagTime · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!doctype html>
<!-- Fasciculus hic a sourcery generatus est, numquam manu scriptus.
     Noli emendare: generatio proxima omnia superscribet. -->
<html lang="und">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="color-scheme" content="light dark">
<link rel="icon" href="data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 16 16'%3E%3Ctext y='14' font-size='14'%3E%F0%9F%90%9D%3C/text%3E%3C/svg%3E">
<meta name="description" content="80 prompts · 2026-07-20 – 2026-07-27 · +1,119 −285">
<meta property="og:title" content="tagtime">
<meta property="og:description" content="80 prompts · 2026-07-20 – 2026-07-27 · +1,119 −285">
<title>tagtime</title>
<style>
/* Beeminder hive: honey paper by day, warm black by night, goldenrod
   accents throughout. Light-mode goldenrod is darkened for text contrast;
   dark mode gets the full #FFB300. */
:root {
  color-scheme: light dark;
  --bg: #faf0d8;
  --ink: #1c1508;
  --muted: #77673f;
  --faint: #a8945e;
  --line: #eadbb4;
  --reply-bg: #f4e9cf;
  --reply-ink: #4a3f22;
  --accent: #a97b00;
  --stripe: #ffb300;
  /* Diffstat mark pair, teal for added and red for deleted, validated per
     mode against its page surface (CVD-simulated ΔE >= 10, contrast >= 3:1;
     a plain green fails deutan/protan separation against these reds). */
  --diff-add: #00785a;
  --diff-del: #cf222e;
  /* Provider identity trio for the exchange edge stripe and meta-line chip,
     validated per mode against its page surface (within-trio CVD-simulated
     ΔE >= 8.7, contrast >= 3:1) and against the diffstat pair and accent.
     Identity never rides on color alone: the agent's name sits beside the
     chip in every meta line. */
  --claude: #b02777;
  --codex: #076f9e;
  --copilot: #6a2fc2;
  --measure: 44rem;
  --serif: "Iowan Old Style", Charter, Georgia, "Times New Roman", serif;
  --sans: ui-sans-serif, -apple-system, "Segoe UI", sans-serif;
  --mono: ui-monospace, SFMono-Regular, Menlo, Consolas, monospace;
}
@media (prefers-color-scheme: dark) {
  :root {
    --bg: #16130f;
    --ink: #ece2cc;
    --muted: #a3946f;
    --faint: #5f5540;
    --line: #2d2718;
    --reply-bg: #1f1a12;
    --reply-ink: #c9bc9e;
    --accent: #ffb300;
    --stripe: #ffb300;
    --diff-add: #1fa179;
    --diff-del: #f85149;
    --claude: #d84f97;
    --codex: #1a94b4;
    --copilot: #8a75f0;
  }
}
* { box-sizing: border-box; }
html { background: var(--bg); }
body {
  margin: 0;
  border-top: 4px solid var(--stripe);
  background: var(--bg);
  color: var(--ink);
  font-family: var(--serif);
  font-size: 17px;
  line-height: 1.6;
  text-rendering: optimizeLegibility;
}
main {
  width: min(calc(100% - 2.5rem), var(--measure));
  margin: 0 auto;
  padding: 5rem 0 8rem;
}
.masthead { padding-bottom: 2rem; }
.generator {
  margin: 0 0 1.1rem;
  font-family: var(--sans);
  font-size: .68rem;
  letter-spacing: .06em;
  color: var(--faint);
}
.generator a { color: inherit; text-decoration: none; }
.generator a:hover { color: var(--accent); text-decoration: underline; }
h1 {
  margin: 0;
  font-size: clamp(1.9rem, 5.5vw, 2.9rem);
  font-weight: 600;
  letter-spacing: -0.02em;
  line-height: 1.1;
}
.deck, .repo-path {
  margin: .6rem 0 0;
  color: var(--faint);
  font-family: var(--sans);
  font-size: .74rem;
  letter-spacing: .03em;
  font-variant-numeric: tabular-nums;
}
.repo-path { font-family: var(--mono); font-size: .7rem; overflow-wrap: anywhere; }
.repo-path a { color: var(--muted); text-decoration: none; }
.repo-path a:hover { color: var(--accent); text-decoration: underline; }
.controls {
  margin: .9rem 0 0;
  font-family: var(--sans);
  font-size: .7rem;
  letter-spacing: .04em;
  color: var(--faint);
}
.controls a { color: var(--muted); text-decoration: none; }
.controls a:hover { color: var(--accent); text-decoration: underline; }
/* Minimap: the dialog at a glance, added lines above the midline and
   deleted below, one clickable sliver per prompt. */
.minimap { display: block; height: 48px; margin: 1.6rem 0 0; }
.minimap .hit { fill: transparent; }
.minimap .add { fill: var(--diff-add); }
.minimap .del { fill: var(--diff-del); }
.minimap .nil { fill: var(--line); }
.minimap a:hover .hit { fill: color-mix(in srgb, var(--accent) 20%, transparent); }
/* The reading-progress rail, the device Asterisk magazine runs down its
   left margin: a goldenrod column — the top stripe's vertical sibling —
   fixed to the viewport's left edge, its height the fraction of the page
   scrolled past, invisible until the reader has actually set off (the
   script adds .show past 300px of scroll). One tick per day header hangs
   on the same percent scale, so the bar's tip touches a day's tick at the
   exact moment that day reaches the top of the viewport; a tick's label
   appears on hover and clicking it jumps to the day. The bar's width
   cedes continuously as the viewport narrows toward the text column, and
   below 54rem the marks leave entirely. */
#progress .progress-bar {
  position: fixed;
  top: 0;
  left: 0;
  width: clamp(6px, calc((100vw - var(--measure)) / 2 - 12px), 32px);
  height: 0;
  opacity: 0;
  background: var(--stripe);
  transition: opacity .1s linear;
  z-index: 10;
}
#progress .daymark {
  position: fixed;
  left: 0;
  z-index: 11;
  display: flex;
  margin-top: -12px;
  color: var(--muted);
  font-family: var(--sans);
  font-size: .65rem;
  letter-spacing: .04em;
  font-variant-numeric: tabular-nums;
  text-decoration: none;
  transform: translateX(-100%);
  transition: transform .2s ease;
}
#progress .tick {
  width: 46px;
  height: 12px;
  border-bottom: 1px solid var(--ink);
}
#progress .text {
  margin-left: .5rem;
  padding: .1rem .4rem;
  border: 1px solid var(--line);
  border-radius: .3rem;
  background: var(--bg);
  opacity: 0;
  transition: opacity .2s ease-in;
}
#progress .daymark:hover .text { opacity: 1; }
#progress.show .progress-bar { opacity: 1; }
#progress.show .daymark { transform: none; }
@media (max-width: 54rem) {
  #progress .daymark { display: none; }
}
::selection { background: color-mix(in srgb, var(--accent) 25%, transparent); }
summary a.anchor { color: inherit; text-decoration: none; }
summary a.anchor:hover { text-decoration: underline; }
.exchange:target {
  background: var(--reply-bg);
  border-radius: .45rem;
  padding: 1.1rem 1.25rem;
  margin: 2.6rem -1.25rem 0;
}
.day {
  clear: both;
  margin: 3.8rem 0 0;
  padding-bottom: .4rem;
  border-bottom: 1px solid color-mix(in srgb, var(--accent) 35%, var(--line));
  color: var(--accent);
  font-family: var(--sans);
  font-size: .68rem;
  font-weight: 600;
  letter-spacing: .14em;
  font-variant-numeric: tabular-nums;
}
.exchange { margin: 2.6rem 0 0; clear: both; border-left: 3px solid var(--provider); padding-left: 0.85rem; }
.exchange.claude { --provider: var(--claude); }
.exchange.codex { --provider: var(--codex); }
.exchange.copilot { --provider: var(--copilot); }
/* The prompt's diffstat floats right of the prompt's first lines. Numbers
   wear the metadata ink; polarity lives in the blocks (and in the signs and
   the fixed added-first order). */
.diffstat {
  float: right;
  display: inline-flex;
  align-items: center;
  gap: .5rem;
  margin: .25rem 0 .5rem 1.1rem;
  color: var(--faint);
  font-family: var(--sans);
  font-size: .68rem;
  letter-spacing: .05em;
  font-variant-numeric: tabular-nums;
}
.diffstat .blocks { display: inline-flex; gap: 2px; }
.diffstat .blocks span { width: 7px; height: 7px; border-radius: 2px; }
.diffstat .add { background: var(--diff-add); }
.diffstat .del { background: var(--diff-del); }
.diffstat .nil { background: var(--line); }
pre.prompt {
  margin: 0;
  white-space: pre-wrap;
  overflow-wrap: anywhere;
  tab-size: 4;
  font: inherit;
  font-size: 1.06rem;
  line-height: 1.62;
}
.attachments {
  margin-top: .75rem;
  display: flex;
  flex-wrap: wrap;
  gap: .5rem;
}
.attachments img {
  max-width: 100%;
  max-height: 20rem;
  border: 1px solid var(--line);
  border-radius: .3rem;
}
details { margin: .55rem 0 0; }
summary {
  display: inline-flex;
  align-items: baseline;
  gap: .6rem;
  width: fit-content;
  list-style: none;
  cursor: pointer;
  user-select: none;
  color: var(--faint);
  font-family: var(--sans);
  font-size: .68rem;
  letter-spacing: .05em;
  font-variant-numeric: tabular-nums;
}
summary::-webkit-details-marker { display: none; }
summary::before {
  content: "▸";
  color: var(--accent);
  font-size: .62rem;
  transition: transform .12s ease;
}
details[open] > summary::before { transform: rotate(90deg); }
summary:hover { color: var(--muted); }
/* The agent's marks in the meta line: a small square chip wearing the
   provider color (the same hue as the edge stripe) and the agent's name a
   step louder than the rest of the line — promoted ink, never the mark
   color, so the reading survives any color deficiency. */
.chip { display: inline-block; width: 8px; height: 8px; border-radius: 2px; background: var(--provider); }
.agent { color: var(--muted); font-weight: 600; }
/* INVIOLABLE: machine-generated prose renders only inside a .machine
   container, in the phosphor-terminal style — monospace green on
   near-black, in both color schemes — for maximal distinction from the
   human's serif. The containers are .reply (agent replies) and .ballot
   (multiple-choice questions the agent posed and the option labels it
   wrote, including the ones the human picked). Machine styling and
   collapsing are separate rules: .reply also hides behind the closed
   disclosure, while .ballot is deliberately default-visible — collapsing
   it would orphan the human's choice — but stays in phosphor. */
.machine {
  --m-bg: #060d08;
  --m-ink: #56dd7f;
  --m-bright: #8dffab;
  --m-dim: #2f8a4f;
  --m-line: #1c4b2d;
  font-family: var(--mono);
  background: var(--m-bg);
  color: var(--m-ink);
}
.ballot {
  clear: right; /* its filled box must not run under a floated diffstat */
  margin: 1rem 0 .6rem;
  padding: .65rem .85rem;
  border-radius: .3rem;
  font-size: .8rem;
  line-height: 1.6;
  overflow-wrap: anywhere;
}
.ballot-question {
  color: var(--m-dim);
  font-size: .74rem;
  margin-bottom: .35rem;
  white-space: pre-wrap;
}
.ballot .option { color: var(--m-dim); }
.ballot .option.picked { color: var(--m-bright); }
.exchange > .ballot:first-child { margin-top: 0; }
.reply {
  margin-top: .8rem;
  padding: 1rem 1.15rem;
  border-radius: .35rem;
  font-size: .82rem;
  line-height: 1.6;
  overflow-wrap: anywhere;
}
.reply > :first-child { margin-top: 0; }
.reply > :last-child { margin-bottom: 0; }
.reply p { margin: .8rem 0; }
.reply h2, .reply h3, .reply h4, .reply h5, .reply h6 {
  margin: 1.3rem 0 .55rem;
  line-height: 1.25;
}
.reply h2 { font-size: 1.05rem; }
.reply h3, .reply h4, .reply h5, .reply h6 { font-size: .95rem; }
.reply ul, .reply ol { margin: .75rem 0; padding-left: 1.4rem; }
.reply li + li { margin-top: .25rem; }
.reply blockquote {
  margin: .9rem 0;
  padding-left: .9rem;
  border-left: 2px solid var(--m-line);
  color: var(--m-dim);
}
.reply code { font-size: .95em; }
.reply :not(pre) > code {
  padding: .1em .3em;
  border: 1px solid var(--m-line);
  border-radius: .25rem;
}
.reply pre.code {
  margin: .9rem 0;
  padding: .9rem 1rem;
  overflow-x: auto;
  white-space: pre;
  border: 1px solid var(--m-line);
  border-radius: .3rem;
  background: #0a170e;
}
.reply a { color: var(--m-bright); }
.reply.empty { font-style: italic; color: var(--m-dim); }
@media print {
  :root { --bg: white; --ink: black; --faint: #777; --muted: #555; --line: #bbb; --reply-bg: white; --reply-ink: #333; }
  #progress { display: none; }
  .machine { --m-bg: white; --m-ink: #14572e; --m-bright: #14572e; --m-dim: #4d7a5d; --m-line: #9dbfa8; background: white; border: 1px solid var(--m-line); }
  .reply pre.code { background: white; }
  main { width: 100%; padding: 0; }
  .day { break-after: avoid; }
  .exchange { break-inside: avoid; }
  details > .reply { display: block !important; }
  summary::before { content: ""; }
}
</style>
</head>
<body>
<div id="progress">
<div class="progress-bar" aria-hidden="true"></div>
<nav class="daymarks" aria-label="Index dierum">
<a class="daymark" href="#d2026-07-20"><span class="tick"></span><span class="text">2026-07-20 Monday</span></a>
<a class="daymark" href="#d2026-07-21"><span class="tick"></span><span class="text">2026-07-21 Tuesday</span></a>
<a class="daymark" href="#d2026-07-22"><span class="tick"></span><span class="text">2026-07-22 Wednesday</span></a>
<a class="daymark" href="#d2026-07-23"><span class="tick"></span><span class="text">2026-07-23 Thursday</span></a>
<a class="daymark" href="#d2026-07-24"><span class="tick"></span><span class="text">2026-07-24 Friday</span></a>
<a class="daymark" href="#d2026-07-25"><span class="tick"></span><span class="text">2026-07-25 Saturday</span></a>
<a class="daymark" href="#d2026-07-26"><span class="tick"></span><span class="text">2026-07-26 Sunday</span></a>
<a class="daymark" href="#d2026-07-27"><span class="tick"></span><span class="text">2026-07-27 Monday</span></a>
</nav>
</div>
<main>
<header class="masthead">
  <p class="generator"><a href="https://github.com/beeminder/sourcery">generated by sourcery</a></p>
  <h1>tagtime</h1>
  <p class="deck">80 prompts · 2026-07-20 – 2026-07-27 · +1,119 −285</p>
  <p class="repo-path"><a href="https://github.com/dreeves/TagTime">github.com/dreeves/TagTime</a></p>
  <p class="controls"><a href="#" data-omnia="open">expand all</a> · <a href="#" data-omnia="close">collapse all</a></p>
  <svg class="minimap" viewBox="0 0 240 64" preserveAspectRatio="none" style="width:min(100%, 480px)" aria-label="Tabula dialogi, rogatio post rogationem"><a href="#p1"><title>17:38</title><rect class="hit" x="0" y="0" width="3" height="64"/><rect class="nil" x="0" y="31" width="2" height="2"/></a><a href="#p2"><title>18:13 +21 −0</title><rect class="hit" x="3" y="0" width="3" height="64"/><rect class="add" x="3" y="20.8" width="2" height="11.2"/></a><a href="#p3"><title>00:42</title><rect class="hit" x="6" y="0" width="3" height="64"/><rect class="nil" x="6" y="31" width="2" height="2"/></a><a href="#p4"><title>01:16 +10 −0</title><rect class="hit" x="9" y="0" width="3" height="64"/><rect class="add" x="9" y="24.3" width="2" height="7.7"/></a><a href="#p5"><title>11:22</title><rect class="hit" x="12" y="0" width="3" height="64"/><rect class="nil" x="12" y="31" width="2" height="2"/></a><a href="#p6"><title>11:59 +7 −10</title><rect class="hit" x="15" y="0" width="3" height="64"/><rect class="add" x="15" y="25.6" width="2" height="6.4"/><rect class="del" x="15" y="32" width="2" height="7.7"/></a><a href="#p7"><title>13:28</title><rect class="hit" x="18" y="0" width="3" height="64"/><rect class="nil" x="18" y="31" width="2" height="2"/></a><a href="#p8"><title>13:32</title><rect class="hit" x="21" y="0" width="3" height="64"/><rect class="nil" x="21" y="31" width="2" height="2"/></a><a href="#p9"><title>13:33</title><rect class="hit" x="24" y="0" width="3" height="64"/><rect class="nil" x="24" y="31" width="2" height="2"/></a><a href="#p10"><title>14:47</title><rect class="hit" x="27" y="0" width="3" height="64"/><rect class="nil" x="27" y="31" width="2" height="2"/></a><a href="#p11"><title>14:57</title><rect class="hit" x="30" y="0" width="3" height="64"/><rect class="nil" x="30" y="31" width="2" height="2"/></a><a href="#p12"><title>15:03 +16 −7</title><rect class="hit" x="33" y="0" width="3" height="64"/><rect class="add" x="33" y="22.3" width="2" height="9.7"/><rect class="del" x="33" y="32" width="2" height="6.4"/></a><a href="#p13"><title>15:07 +33 −22</title><rect class="hit" x="36" y="0" width="3" height="64"/><rect class="add" x="36" y="18" width="2" height="14"/><rect class="del" x="36" y="32" width="2" height="11.4"/></a><a href="#p14"><title>15:23</title><rect class="hit" x="39" y="0" width="3" height="64"/><rect class="nil" x="39" y="31" width="2" height="2"/></a><a href="#p15"><title>15:28</title><rect class="hit" x="42" y="0" width="3" height="64"/><rect class="nil" x="42" y="31" width="2" height="2"/></a><a href="#p16"><title>15:33 +29 −0</title><rect class="hit" x="45" y="0" width="3" height="64"/><rect class="add" x="45" y="18.9" width="2" height="13.1"/></a><a href="#p17"><title>15:34 +42 −15</title><rect class="hit" x="48" y="0" width="3" height="64"/><rect class="add" x="48" y="16.2" width="2" height="15.8"/><rect class="del" x="48" y="32" width="2" height="9.4"/></a><a href="#p18"><title>16:04 +8 −6</title><rect class="hit" x="51" y="0" width="3" height="64"/><rect class="add" x="51" y="25.1" width="2" height="6.9"/><rect class="del" x="51" y="32" width="2" height="6"/></a><a href="#p19"><title>16:05 +4 −0</title><rect class="hit" x="54" y="0" width="3" height="64"/><rect class="add" x="54" y="27.1" width="2" height="4.9"/></a><a href="#p20"><title>16:10 +1 −1</title><rect class="hit" x="57" y="0" width="3" height="64"/><rect class="add" x="57" y="29.6" width="2" height="2.4"/><rect class="del" x="57" y="32" width="2" height="2.4"/></a><a href="#p21"><title>16:13 +13 −0</title><rect class="hit" x="60" y="0" width="3" height="64"/><rect class="add" x="60" y="23.2" width="2" height="8.8"/></a><a href="#p22"><title>19:10 +12 −3</title><rect class="hit" x="63" y="0" width="3" height="64"/><rect class="add" x="63" y="23.6" width="2" height="8.4"/><rect class="del" x="63" y="32" width="2" height="4.2"/></a><a href="#p23"><title>21:27 +6 −2</title><rect class="hit" x="66" y="0" width="3" height="64"/><rect class="add" x="66" y="26" width="2" height="6"/><rect class="del" x="66" y="32" width="2" height="3.4"/></a><a href="#p24"><title>22:01</title><rect class="hit" x="69" y="0" width="3" height="64"/><rect class="nil" x="69" y="31" width="2" height="2"/></a><a href="#p25"><title>23:18</title><rect class="hit" x="72" y="0" width="3" height="64"/><rect class="nil" x="72" y="31" width="2" height="2"/></a><a href="#p26"><title>23:33</title><rect class="hit" x="75" y="0" width="3" height="64"/><rect class="nil" x="75" y="31" width="2" height="2"/></a><a href="#p27"><title>09:57</title><rect class="hit" x="78" y="0" width="3" height="64"/><rect class="nil" x="78" y="31" width="2" height="2"/></a><a href="#p28"><title>10:12</title><rect class="hit" x="81" y="0" width="3" height="64"/><rect class="nil" x="81" y="31" width="2" height="2"/></a><a href="#p29"><title>10:22</title><rect class="hit" x="84" y="0" width="3" height="64"/><rect class="nil" x="84" y="31" width="2" height="2"/></a><a href="#p30"><title>10:35 +53 −49</title><rect class="hit" x="87" y="0" width="3" height="64"/><rect class="add" x="87" y="14.3" width="2" height="17.7"/><rect class="del" x="87" y="32" width="2" height="17"/></a><a href="#p31"><title>11:07 +10 −24</title><rect class="hit" x="90" y="0" width="3" height="64"/><rect class="add" x="90" y="24.3" width="2" height="7.7"/><rect class="del" x="90" y="32" width="2" height="11.9"/></a><a href="#p32"><title>11:37</title><rect class="hit" x="93" y="0" width="3" height="64"/><rect class="nil" x="93" y="31" width="2" height="2"/></a><a href="#p33"><title>14:10 +11 −4</title><rect class="hit" x="96" y="0" width="3" height="64"/><rect class="add" x="96" y="23.9" width="2" height="8.1"/><rect class="del" x="96" y="32" width="2" height="4.9"/></a><a href="#p34"><title>15:18 +27 −0</title><rect class="hit" x="99" y="0" width="3" height="64"/><rect class="add" x="99" y="19.4" width="2" height="12.6"/></a><a href="#p35"><title>23:40</title><rect class="hit" x="102" y="0" width="3" height="64"/><rect class="nil" x="102" y="31" width="2" height="2"/></a><a href="#p36"><title>00:40</title><rect class="hit" x="105" y="0" width="3" height="64"/><rect class="nil" x="105" y="31" width="2" height="2"/></a><a href="#p37"><title>00:42</title><rect class="hit" x="108" y="0" width="3" height="64"/><rect class="nil" x="108" y="31" width="2" height="2"/></a><a href="#p38"><title>01:36 +152 −2</title><rect class="hit" x="111" y="0" width="3" height="64"/><rect class="add" x="111" y="2" width="2" height="30"/><rect class="del" x="111" y="32" width="2" height="3.4"/></a><a href="#p39"><title>03:06</title><rect class="hit" x="114" y="0" width="3" height="64"/><rect class="nil" x="114" y="31" width="2" height="2"/></a><a href="#p40"><title>03:12</title><rect class="hit" x="117" y="0" width="3" height="64"/><rect class="nil" x="117" y="31" width="2" height="2"/></a><a href="#p41"><title>12:17 +21 −1</title><rect class="hit" x="120" y="0" width="3" height="64"/><rect class="add" x="120" y="20.8" width="2" height="11.2"/><rect class="del" x="120" y="32" width="2" height="2.4"/></a><a href="#p42"><title>12:22</title><rect class="hit" x="123" y="0" width="3" height="64"/><rect class="nil" x="123" y="31" width="2" height="2"/></a><a href="#p43"><title>12:24 +5 −1</title><rect class="hit" x="126" y="0" width="3" height="64"/><rect class="add" x="126" y="26.6" width="2" height="5.4"/><rect class="del" x="126" y="32" width="2" height="2.4"/></a><a href="#p44"><title>12:27</title><rect class="hit" x="129" y="0" width="3" height="64"/><rect class="nil" x="129" y="31" width="2" height="2"/></a><a href="#p45"><title>12:28 +71 −24</title><rect class="hit" x="132" y="0" width="3" height="64"/><rect class="add" x="132" y="11.5" width="2" height="20.5"/><rect class="del" x="132" y="32" width="2" height="11.9"/></a><a href="#p46"><title>12:39 +3 −0</title><rect class="hit" x="135" y="0" width="3" height="64"/><rect class="add" x="135" y="27.8" width="2" height="4.2"/></a><a href="#p47"><title>14:52 +29 −1</title><rect class="hit" x="138" y="0" width="3" height="64"/><rect class="add" x="138" y="18.9" width="2" height="13.1"/><rect class="del" x="138" y="32" width="2" height="2.4"/></a><a href="#p48"><title>15:23 +55 −3</title><rect class="hit" x="141" y="0" width="3" height="64"/><rect class="add" x="141" y="14" width="2" height="18"/><rect class="del" x="141" y="32" width="2" height="4.2"/></a><a href="#p49"><title>15:29</title><rect class="hit" x="144" y="0" width="3" height="64"/><rect class="nil" x="144" y="31" width="2" height="2"/></a><a href="#p50"><title>16:00 +1 −1</title><rect class="hit" x="147" y="0" width="3" height="64"/><rect class="add" x="147" y="29.6" width="2" height="2.4"/><rect class="del" x="147" y="32" width="2" height="2.4"/></a><a href="#p51"><title>16:11 +24 −9</title><rect class="hit" x="150" y="0" width="3" height="64"/><rect class="add" x="150" y="20.1" width="2" height="11.9"/><rect class="del" x="150" y="32" width="2" height="7.3"/></a><a href="#p52"><title>16:17</title><rect class="hit" x="153" y="0" width="3" height="64"/><rect class="nil" x="153" y="31" width="2" height="2"/></a><a href="#p53"><title>16:19</title><rect class="hit" x="156" y="0" width="3" height="64"/><rect class="nil" x="156" y="31" width="2" height="2"/></a><a href="#p54"><title>16:24 +6 −1</title><rect class="hit" x="159" y="0" width="3" height="64"/><rect class="add" x="159" y="26" width="2" height="6"/><rect class="del" x="159" y="32" width="2" height="2.4"/></a><a href="#p55"><title>16:31 +11 −4</title><rect class="hit" x="162" y="0" width="3" height="64"/><rect class="add" x="162" y="23.9" width="2" height="8.1"/><rect class="del" x="162" y="32" width="2" height="4.9"/></a><a href="#p56"><title>17:05</title><rect class="hit" x="165" y="0" width="3" height="64"/><rect class="nil" x="165" y="31" width="2" height="2"/></a><a href="#p57"><title>14:13</title><rect class="hit" x="168" y="0" width="3" height="64"/><rect class="nil" x="168" y="31" width="2" height="2"/></a><a href="#p58"><title>14:19</title><rect class="hit" x="171" y="0" width="3" height="64"/><rect class="nil" x="171" y="31" width="2" height="2"/></a><a href="#p59"><title>14:42</title><rect class="hit" x="174" y="0" width="3" height="64"/><rect class="nil" x="174" y="31" width="2" height="2"/></a><a href="#p60"><title>14:56</title><rect class="hit" x="177" y="0" width="3" height="64"/><rect class="nil" x="177" y="31" width="2" height="2"/></a><a href="#p61"><title>16:02</title><rect class="hit" x="180" y="0" width="3" height="64"/><rect class="nil" x="180" y="31" width="2" height="2"/></a><a href="#p62"><title>16:03</title><rect class="hit" x="183" y="0" width="3" height="64"/><rect class="nil" x="183" y="31" width="2" height="2"/></a><a href="#p63"><title>16:13</title><rect class="hit" x="186" y="0" width="3" height="64"/><rect class="nil" x="186" y="31" width="2" height="2"/></a><a href="#p64"><title>17:44</title><rect class="hit" x="189" y="0" width="3" height="64"/><rect class="nil" x="189" y="31" width="2" height="2"/></a><a href="#p65"><title>18:09 +104 −31</title><rect class="hit" x="192" y="0" width="3" height="64"/><rect class="add" x="192" y="7.2" width="2" height="24.8"/><rect class="del" x="192" y="32" width="2" height="13.5"/></a><a href="#p66"><title>19:40</title><rect class="hit" x="195" y="0" width="3" height="64"/><rect class="nil" x="195" y="31" width="2" height="2"/></a><a href="#p67"><title>21:20</title><rect class="hit" x="198" y="0" width="3" height="64"/><rect class="nil" x="198" y="31" width="2" height="2"/></a><a href="#p68"><title>21:28</title><rect class="hit" x="201" y="0" width="3" height="64"/><rect class="nil" x="201" y="31" width="2" height="2"/></a><a href="#p69"><title>22:13 +58 −34</title><rect class="hit" x="204" y="0" width="3" height="64"/><rect class="add" x="204" y="13.5" width="2" height="18.5"/><rect class="del" x="204" y="32" width="2" height="14.2"/></a><a href="#p70"><title>23:01</title><rect class="hit" x="207" y="0" width="3" height="64"/><rect class="nil" x="207" y="31" width="2" height="2"/></a><a href="#p71"><title>23:03</title><rect class="hit" x="210" y="0" width="3" height="64"/><rect class="nil" x="210" y="31" width="2" height="2"/></a><a href="#p72"><title>23:37 +117 −1</title><rect class="hit" x="213" y="0" width="3" height="64"/><rect class="add" x="213" y="5.7" width="2" height="26.3"/><rect class="del" x="213" y="32" width="2" height="2.4"/></a><a href="#p73"><title>23:52</title><rect class="hit" x="216" y="0" width="3" height="64"/><rect class="nil" x="216" y="31" width="2" height="2"/></a><a href="#p74"><title>23:56</title><rect class="hit" x="219" y="0" width="3" height="64"/><rect class="nil" x="219" y="31" width="2" height="2"/></a><a href="#p75"><title>10:59 +19 −0</title><rect class="hit" x="222" y="0" width="3" height="64"/><rect class="add" x="222" y="21.4" width="2" height="10.6"/></a><a href="#p76"><title>15:14</title><rect class="hit" x="225" y="0" width="3" height="64"/><rect class="nil" x="225" y="31" width="2" height="2"/></a><a href="#p77"><title>15:17 +49 −2</title><rect class="hit" x="228" y="0" width="3" height="64"/><rect class="add" x="228" y="15" width="2" height="17"/><rect class="del" x="228" y="32" width="2" height="3.4"/></a><a href="#p78"><title>15:31</title><rect class="hit" x="231" y="0" width="3" height="64"/><rect class="nil" x="231" y="31" width="2" height="2"/></a><a href="#p79"><title>15:33 +59 −27</title><rect class="hit" x="234" y="0" width="3" height="64"/><rect class="add" x="234" y="13.3" width="2" height="18.7"/><rect class="del" x="234" y="32" width="2" height="12.6"/></a><a href="#p80"><title>15:43 +32 −0</title><rect class="hit" x="237" y="0" width="3" height="64"/><rect class="add" x="237" y="18.2" width="2" height="13.8"/></a></svg>
</header>
<h2 class="day" id="d2026-07-20"><time datetime="2026-07-20">2026-07-20 Monday</time></h2>
<article class="exchange claude" id="p1">
<pre class="prompt">every since the last upgrade to macOS i'm having a problem where the tagtime window can take many seconds to pop up after i hear the ping. any ideas?</pre>
<details>
<summary><a class="anchor" href="#p1"><time datetime="2026-07-21T00:38:21.040000+00:00">17:38</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 7m23s · 33m27s wall-clock time</span></summary>
<div class="reply machine"><p>Let me look at how the popup actually gets launched on macOS in your setup, then I can reason about what the OS upgrade might have changed.</p>
<p>The popup goes through [iterm2-tagtime](iterm2-tagtime) — let me see what that does.</p>
<p>pmset log is blocked in the sandbox — let me check the kernel log for a wake event around 11:01 instead.</p>
<p>Findings, from tracing today&#x27;s actual pings through the unified log:</p>
<p><strong>Every popup has ~2.5s of fixed overhead, and about 1s of it is the sound itself.</strong> [tagtimed.pl:99](tagtimed.pl#L99) plays the ping sound with a blocking <code>system(&quot;afplay ...&quot;)</code> and only starts [launch.pl](launch.pl#L103) after afplay exits — and afplay measures 0.94s of wall time to play the 0.08s blip (CoreAudio spin-up/teardown overhead). Then the chain is: perl startup → [iterm2-tagtime](iterm2-tagtime) → <code>open -a iTerm</code> (~0.4s, redundant since the AppleScript&#x27;s <code>activate</code> does the same thing) → osascript startup (~0.4s, which now includes an XProtect YARA scan — newer macOS malware-scans every osascript invocation, visible in the log as <code>XPScripts.yr</code> — plus TCC checks) → <code>create window with profile</code> (median 0.46s, and iTerm2 spends ~0.5s of that re-verifying Automation permission per window).</p>
<p><strong>The &quot;many seconds&quot; cases correlate with the screen locking.</strong> Out of 21 pings in today&#x27;s log, one took 18.7s for iTerm2 to create the window (11:01:44). The log shows <code>kCGSDisplayWillSleep</code> and the screensaver/lock engaging at 11:01:41 — window creation stalls while the display is asleep or locking. So the worst delays are macOS deferring window work during lock/display-sleep transitions, not anything in TagTime.</p>
<p>Caveats on the evidence: the unified log only retains about a day at the level needed for this, so I can&#x27;t compare against pre-upgrade behavior directly, and I have exactly one many-second sample. The per-exec XProtect script scanning and heavier per-event TCC verification are the plausible upgrade-related additions to the baseline. (Also note: pings that fire while a previous ping window is still open only pop up after you answer the earlier one — [launch.pl](launch.pl#L58) serializes them by design — which can read as popup lag if that&#x27;s the case you&#x27;re noticing.)</p>
<p>Ideas, in order of payoff-per-effort:</p>
<ol><li><strong>Start the popup before (or concurrent with) the sound.</strong> Either swap lines so <code>launch.pl quiet &amp;</code> fires before the afplay call in [tagtimed.pl:97-104](tagtimed.pl#L97-L104), or append <code>&amp;</code> to <code>$playsound</code> in [settings.pl:21](settings.pl#L21). Saves ~0.9s of perceived delay.</li><li><strong>Drop <code>open -a iTerm</code> from [iterm2-tagtime:3](iterm2-tagtime#L3).</strong> The <code>activate</code> in the AppleScript launches iTerm if needed. Saves ~0.4s.</li><li>The screen-lock stall has no TagTime-side fix; reducing display-sleep aggressiveness would reduce how often you hit it.</li><li>If you observe a many-second delay while the display is on and unlocked, that would contradict the pattern above — in that case it&#x27;d be worth adding timestamps to launch.pl to catch a bad case in the wild.</li></ol>
<p>Items 1 and 2 together cut the typical sound-to-window gap roughly in half. Say the word and I&#x27;ll make those edits.</p></div>
</details>
</article>
<article class="exchange claude" id="p2">
<div class="diffstat">+21 −0 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span></span></div>
<pre class="prompt">yes, it happens with the display on and unlocked</pre>
<details>
<summary><a class="anchor" href="#p2"><time datetime="2026-07-21T01:13:59.897000+00:00">18:13</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 5m21s · 12m18s wall-clock time</span></summary>
<div class="reply machine"><p>That rules out the screen-lock explanation for the cases you&#x27;re hitting. Before instrumenting, let me squeeze today&#x27;s log harder: those medium ping→send lags (6–70s) that I attributed to queuing behind an unanswered window — I never actually verified a window was open at those moments. The Apple-event log lets me reconstruct each popup session&#x27;s lifetime (osascript polls every second while its window exists), so I can check which lags were genuine latency and which were queuing.</p>
<p>Instrumentation is in place and verified. What I did:</p>
<ul><li>[launch.pl](launch.pl#L105-L121): logs <code>launch ping=&lt;t&gt;</code> immediately before spawning the popup and <code>closed ping=&lt;t&gt;</code> when the window closes, with millisecond timestamps, to <code>tmp/popup-timing.log</code> (gitignored). A <code>tslog</code> sub does the appending; failures print in the file&#x27;s existing SYSERR idiom rather than dying, so a logging problem can&#x27;t block a popup.</li><li>[iterm2-tagtime](iterm2-tagtime#L3-L11): logs <code>xt-start</code> (bash spawned), <code>xt-opened</code> (after <code>open -a iTerm</code>), and <code>xt-exit</code>. The AppleScript body is untouched.</li><li>Both files are re-executed fresh on every ping, so this is live now — no daemon restart needed. <code>perl -c</code> and <code>bash -n</code> pass, and two end-to-end runs confirmed the log lines fire. (First test ran the window command <code>true</code>, which exits so fast the AppleScript&#x27;s window reference came back as <code>missing value</code> and errored — an artifact of the instant-exit test command, not of the edits; a second test with <code>sleep 3</code> ran the real path cleanly: 0.18s to <code>open</code>, window created and closed normally.)</li></ul>
<p>These log lines, combined with what the unified log records on its own (the create-window Apple event send/reply), split every future popup into: scheduled ping → afplay+perl startup → bash spawn → <code>open -a</code> → osascript startup → iTerm2 window creation. The <code>closed</code> lines also let me distinguish real latency from pings queued behind an unanswered window automatically.</p>
<p>So: nothing you need to do except live your life. Next time you notice a slow popup with the display on and unlocked, come back (a rough memory of when helps but isn&#x27;t required) and I&#x27;ll pull the offending ping&#x27;s segment breakdown from the timing log. One caveat: the gap between the scheduled ping time and the <code>launch</code> line includes afplay running inside the daemon — if the missing seconds turn out to hide there, the next step is instrumenting [tagtimed.pl](tagtimed.pl) itself, which does require restarting the daemon.</p>
<p>I&#x27;ve deliberately not applied the two speed-ups (reordering the sound relative to the popup, dropping the redundant <code>open -a</code>) so they don&#x27;t contaminate the measurements. They&#x27;re still worth doing once the slow case is diagnosed.</p></div>
</details>
</article>
<h2 class="day" id="d2026-07-21"><time datetime="2026-07-21">2026-07-21 Tuesday</time></h2>
<article class="exchange claude" id="p3">
<pre class="prompt">sounds good. aaaaaand, it just happened. alt-tabbing to iterm2 seems to nudge the window to finally appear. note i'm not alt-tabbing to the red window, just that alt-tabbing to the app seems to trigger the red window to finally launch.</pre>
<details>
<summary><a class="anchor" href="#p3"><time datetime="2026-07-21T07:42:13.910000+00:00">00:42</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 2m37s</span></summary>
<div class="reply machine"><p>That&#x27;s a strong clue — window creation stalling until iTerm2 gets focus points at App Nap or macOS&#x27;s focus-stealing deferral. And the instrumentation should have caught this one. Let me pull the timing log.</p></div>
</details>
</article>
<article class="exchange claude" id="p4">
<div class="diffstat">+10 −0 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span></span></div>
<pre class="prompt">(again just now, btw, with less delay because i more quickly alt-tabbed to iTerm2)</pre>
<details>
<summary><a class="anchor" href="#p4"><time datetime="2026-07-21T08:16:22.310000+00:00">01:16</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 6m2s · 49m57s wall-clock time</span></summary>
<div class="reply machine"><p>Root cause confirmed and a fix is deployed. Summary:</p>
<p><strong>What&#x27;s happening.</strong> The window is never late — it&#x27;s *hidden*. iTerm2 creates it in ~0.5s every time, but macOS 26&#x27;s focus-stealing prevention (visible in the WindowServer log as &quot;DeferringManager&quot;) defers or drops activation requests from background processes. In the 23:40 incident, iTerm went frontmost, the window was created, and VS Code re-took frontmost 1.5 seconds later — leaving the red window buried behind your windows. Then the queueing in [launch.pl](launch.pl#L36-L39) amplifies it: subsequent pings beep but can&#x27;t open windows until the buried one is answered, so you hear pings with nothing appearing. Alt-tabbing to iTerm2 is a *user-initiated* activation, which macOS always honors — that&#x27;s why it &quot;nudges&quot; the window out. Apple DTS says there&#x27;s no reliable workaround and treats it as a bug (<a href="https://developer.apple.com/forums/thread/807805">Apple Developer Forums thread</a>, FB21087054); a Safari-flavored variant of the same Tahoe focus chaos is described <a href="https://dev.to/achiya-automation/i-had-a-focus-guard-macos-tahoe-walked-through-it-three-times-1chp">here</a>.</p>
<p><strong>The fix</strong>, in [iterm2-tagtime](iterm2-tagtime#L21-L33): the existing once-per-second poll loop now re-issues <code>activate</code> + <code>select newWindow</code> on its first 5 iterations. A single activation can be swallowed by the deferral, but one of several attempts over ~6 seconds lands as soon as the typing shield lapses — same mechanism as your alt-tab, minus you. Per the anti-magic rule, flagging the one if-statement this adds: <code>if raiseTries &gt; 0</code> bounds the re-raising, because raising forever would make deliberately ignoring a ping impossible (the 19:54 window sat 3.5 hours during afk — an eternal re-activate loop there would be unusable). Bounded, it approximates the pre-upgrade semantics: window pops on top once, then leaves you alone. On quals: OS focus arbitration isn&#x27;t testable headlessly, so validation is empirical — the instrumentation stays in place and your next several pings are the quals.</p>
<p><strong>What to expect / next steps:</strong></p>
<ul><li>Next pings should surface within ~6s of the beep even while you&#x27;re typing elsewhere. If one still stays buried, tell me — escalations in reserve: a second LaunchServices <code>open -a iTerm</code> after window creation (the community workaround from the Apple thread), the <a href="https://iterm2.com/appcasts/testing_changes3.txt">iTerm2 3.7.0 beta</a>, or extending <code>raiseTries</code>.</li><li>Note <code>open -a iTerm</code> is no longer redundant-looking dead weight — it&#x27;s an activation signal — so I&#x27;ve dropped my earlier suggestion to remove it.</li><li>Still pending your go-ahead: the ~1s afplay speedup (backgrounding the sound in [settings.pl:21](settings.pl#L21), or reordering in [tagtimed.pl](tagtimed.pl#L97-L104), which requires restarting the daemon).</li><li>Once you&#x27;re satisfied the fix holds, I&#x27;ll strip the instrumentation from both files.</li></ul>
<p>Sources: <a href="https://developer.apple.com/forums/thread/807805">Apple Developer Forums — Activating application from Terminal occasionally fails on macOS 26</a>, <a href="https://dev.to/achiya-automation/i-had-a-focus-guard-macos-tahoe-walked-through-it-three-times-1chp">dev.to — macOS Tahoe focus guard</a>, <a href="https://iterm2.com/downloads.html">iTerm2 downloads</a>, <a href="https://iterm2.com/appcasts/testing_changes3.txt">iTerm2 3.7 beta changelog</a>.</p></div>
</details>
</article>
<article class="exchange claude" id="p5">
<pre class="prompt">ok, it just happened again just now</pre>
<details>
<summary><a class="anchor" href="#p5"><time datetime="2026-07-21T18:22:28.490000+00:00">11:22</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 4m23s</span></summary>
<div class="reply machine"><p>Let me pull the data for this latest ping.</p></div>
</details>
</article>
<article class="exchange claude" id="p6">
<div class="diffstat">+7 −10 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="del"></span><span class="del"></span><span class="nil"></span></span></div>
<div class="ballot machine">
<div class="ballot-question">How aggressively should the ping window fight for attention when macOS keeps restoring focus to the app you're typing in?</div>
<div class="option">· Re-raise until answered (Recommended)</div>
<div class="option">· Longer burst, then give up</div>
<div class="option">· Burst + notification fallback</div>
</div>
<pre class="prompt">what's the maximally aggressive thing we can do?</pre>
<details>
<summary><a class="anchor" href="#p6"><time datetime="2026-07-21T18:59:32.634000+00:00">11:59</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m9s · 3m22s wall-clock time</span></summary>
<div class="reply machine"><p>Deployed and tested. The final version of the loop in [iterm2-tagtime](iterm2-tagtime#L21-L32): every second, as long as the window exists, <code>activate</code> + <code>select newWindow</code>. No counter, no give-up — the <code>raiseTries</code> if-statement from the interim version is gone entirely, so this is both the most aggressive and the least conditional form of the code. <code>bash -n</code> passes and an end-to-end run with a 6-second window exited cleanly.</p>
<p>What to expect from the next pings:</p>
<ul><li>If you&#x27;re not typing when the ping fires: window pops on top immediately, as pre-upgrade.</li><li>If you are typing: macOS will keep restoring your app for as long as you keep typing (with a once-per-second flicker as iTerm repeatedly goes front), and the red window lands on top with focus the instant you pause. Nothing can sit buried for minutes anymore, and a forgotten open window can no longer silently starve the queue — it will re-assert itself every second until answered.</li><li>The flip side, by your choice: an unanswered ping makes the rest of the machine effectively unusable until you answer it.</li></ul>
<p>Still open, for whenever: the ~1s afplay speedup in [settings.pl:21](settings.pl#L21) or [tagtimed.pl](tagtimed.pl#L97-L104) (the latter needs a daemon restart), and stripping the instrumentation out of [launch.pl](launch.pl) and [iterm2-tagtime](iterm2-tagtime) once you&#x27;re satisfied the fix holds. Report back after a few natural pings — especially one that catches you mid-typing.</p></div>
</details>
</article>
<article class="exchange claude" id="p7">
<pre class="prompt">while we await the next ping, can we think about the ideal way to do this? when a ping pings it needs to be completely unmissable and completely interrupt you no matter what you're doing. if you really need to postpone answering the ping, there should be a low-friction way to do that but it shouldn't go out of sight out of mind. we need a reliable way to find out from the user what they were doing at that exact moment in time.

sky-pie: construct a nice html page with screenshots from all connected displays at the moment of each ping.</pre>
<details>
<summary><a class="anchor" href="#p7"><time datetime="2026-07-21T20:28:06.724000+00:00">13:28</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 2m53s</span></summary>
<div class="reply machine"><p>Good framing — three separable problems: unmissability, principled postponing, and ground-truth capture. The sky-pie part is actually buildable right now, so let me think through the design while checking feasibility facts. First: something already ran <code>/usr/sbin/screencapture</code> under Terminal at 17:39:56 yesterday (it showed up in the TCC log, allowed) — suspiciously close to a ping. Let me check whether you already have capture wired somewhere, see the answer flow, and test multi-display capture mechanics:</p></div>
</details>
</article>
<article class="exchange claude" id="p8">
<pre class="prompt">how did you know about scrumbs?</pre>
<details>
<summary><a class="anchor" href="#p8"><time datetime="2026-07-21T20:32:25.457000+00:00">13:32</time></a> <span class="chip"></span> <span class="agent">Claude Code</span></summary>
<div class="reply machine empty"><p>No response.</p></div>
</details>
</article>
<article class="exchange claude" id="p9">
<pre class="prompt">we'll get to the screencap stuff later. request for explanation of where you saw scrumbs</pre>
<details>
<summary><a class="anchor" href="#p9"><time datetime="2026-07-21T20:33:26.360000+00:00">13:33</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 11s</span></summary>
<div class="reply machine"><p>From your own crontab. In my previous set of commands I ran:</p>
<pre class="code"><code>crontab -l 2&gt;/dev/null | grep -i &quot;screen\|capture&quot;</code></pre>
<p>which matched this commented-out line:</p>
<pre class="code"><code># * * * * * /Users/dreeves/Dropbox/bin/screencap.sh # replaced by scrumbs</code></pre>
<p>The trailing comment — &quot;replaced by scrumbs&quot; — is the only place I&#x27;ve seen the word. I don&#x27;t know anything else about it.</p>
<p>Why I was looking there: yesterday&#x27;s TCC log had <code>/usr/sbin/screencapture</code> running with Terminal as the responsible process at 17:39:56, seconds before a ping, so before designing screenshot capture I wanted to know whether some screenshot mechanism already existed on your machine. The crontab grep was part of that check, along with grepping this repo (no hits) and counting displays.</p>
<p>Holding off on all the screencap work per your note. The design discussion from your message is still on the table whenever you want to pick it up — and the aggressive re-raise fix is live in the meantime, awaiting its first natural mid-typing ping.</p></div>
</details>
</article>
<article class="exchange claude" id="p10">
<pre class="prompt">i think we've solved the delay problem, which was the most important thing. but what actually changed here? i don't see anything in the code except your tslog instrumentation.</pre>
<details>
<summary><a class="anchor" href="#p10"><time datetime="2026-07-21T21:47:46.041000+00:00">14:47</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m11s · 1m12s wall-clock time</span></summary>
<div class="reply machine"><p>The fix isn&#x27;t in [launch.pl](launch.pl) — that file only got instrumentation. The actual change is in [iterm2-tagtime](iterm2-tagtime), the AppleScript wrapper. Here&#x27;s the diff:</p>
<p>Empty diff — that needs explaining. Let me check whether the file is even tracked, and confirm its current contents:</p>
<p>Mystery solved: <strong>iterm2-tagtime is in your .gitignore</strong> (line 21), so it never appears in git status or git diff — your review-the-diff-in-the-IDE workflow has a blind spot for it, and that&#x27;s worth knowing about generally: I edited a file your review process doesn&#x27;t surface.</p>
<p>What actually changed, in full:</p>
<ul><li><strong>[iterm2-tagtime](iterm2-tagtime) — the actual fix</strong> (gitignored, so review it by opening the file; the current content is in the Read output above). The AppleScript&#x27;s wait loop used to be just <code>repeat while exists newWindow / delay 1</code>. It now runs <code>activate</code> + <code>select newWindow</code> on every one-second iteration until the window closes, plus the explanatory comment. Nothing else about the pipeline changed — the window was always being *created* in ~0.5s; macOS 26 was letting your active app keep it buried, and this loop re-fronts it until it sticks. The top of the file also gained the <code>xt-*</code> timing instrumentation (same temporary status as tslog).</li><li><strong>[launch.pl](launch.pl)</strong> — instrumentation only, as you saw.</li><li><strong>tmp/popup-timing.log</strong> — generated timing data (under gitignored <code>tmp</code>).</li></ul>
<p>That&#x27;s every repo file I&#x27;ve touched. Also for the record: settings.pl and terminal-tagtime are likewise gitignored, so the same blind spot applies if we later do the afplay speedup in settings.pl — I&#x27;ll call it out explicitly whenever an edit lands in an ignored file.</p></div>
</details>
</article>
<article class="exchange claude" id="p11">
<pre class="prompt">i see. ok, this is working. would it be just as easy to force-steal focus the first time, so the window pops up, but not doing it repeatedly? and is it an option to make the window pop up without actually stealing focus?</pre>
<details>
<summary><a class="anchor" href="#p11"><time datetime="2026-07-21T21:57:45.239000+00:00">14:57</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m9s</span></summary>
<div class="reply machine"><p>Both good questions, and the diagnosis we did answers them fairly precisely.</p>
<p><strong>&quot;Force-steal once so it pops, then stop&quot;</strong> — a true one-shot force doesn&#x27;t exist on macOS 26; that&#x27;s the root cause of this whole bug. Every single steal we observed *succeeded* and was then reverted by the OS within 60–1100ms in favor of the app you were typing in. There is no API that says &quot;activate and make it stick.&quot; What *is* easy is the honest version of what you&#x27;re describing: <strong>retry until one steal survives, then stop fighting</strong>. The raise loop exits as soon as iTerm2 is genuinely frontmost instead of continuing until you answer:</p>
<pre class="code"><code data-language="applescript">  repeat while (exists newWindow) and not frontmost
    activate
    select newWindow
    delay 1
  end repeat
  repeat while exists newWindow
    delay 1
  end repeat</code></pre>
<p>Same size as the current code. The behavioral difference from what&#x27;s deployed now: once the window has landed in front of you, tabbing away parks it — no re-nagging. That reintroduces the tab-away-and-forget failure (this morning&#x27;s 11:22 window sat 22 minutes exactly that way, and a forgotten window silently blocks all subsequent pings until answered). One caveat: roughly one revert in five came slower than my 1-second check cadence, so occasionally the loop would believe the steal stuck when it didn&#x27;t; checking <code>frontmost</code> on two consecutive iterations before quitting would close that gap.</p>
<p><strong>&quot;Pop up visible without stealing focus at all&quot;</strong> — not with iTerm2 windows. A regular window&#x27;s place in the stacking order is tied to app activation: the only way a background app&#x27;s normal window gets above the active app&#x27;s windows is activation, which is exactly what macOS 26 now polices. iTerm2&#x27;s floating-window feature applies only to hotkey windows, not scripted ones. So within the current architecture, no.</p>
<p>But it *is* possible — and it&#x27;s the architecture I&#x27;d call the actual ideal, matching your &quot;completely unmissable&quot; spec from earlier: a tiny native helper (~150 lines of Swift, compiled once, no dependencies) that shows a <strong>non-activating panel at an elevated window level</strong> — the tier that things like Spotlight occupy. Properties that fall out of that design for free: visible above everything including fullscreen apps, on all displays and Spaces, regardless of which app is active; takes zero keyboard focus, so your typing stream is never touched; yet a click (or global hotkey) lets you type the answer into it without any app switch. It splits the two things macOS currently welds together — *visual* interruption and *input* interruption — which is precisely the distinction your two questions are circling. The snooze-to-countdown-badge idea from your ideal-design message fits naturally there too, and it would replace all of the AppleScript activation warfare rather than adding to it.</p>
<p>So the menu, in increasing order of build effort: (a) keep current max-aggression, (b) the retry-until-stuck variant above (few minutes), (c) the native panel (a real but small project). Say which and I&#x27;ll do it — and if it&#x27;s (c), I&#x27;d start with a minimal panel that just proves unmissability-without-focus-theft on your setup before wiring it into the ping flow.</p></div>
</details>
</article>
<article class="exchange claude" id="p12">
<div class="diffstat">+16 −7 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span><span class="nil"></span></span></div>
<pre class="prompt">verbosity minus minus.
ok, let's do (b)</pre>
<details>
<summary><a class="anchor" href="#p12"><time datetime="2026-07-21T22:03:32.623000+00:00">15:03</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 30s · 1m46s wall-clock time</span></summary>
<div class="reply machine"><p>Done. [iterm2-tagtime](iterm2-tagtime) now re-raises each second only until iTerm2 stays frontmost for 2 consecutive polls, then goes quiet. Syntax-checked and test-ran clean (again in the gitignored file, so no diff will show). Tradeoff now live: a window you tab away from won&#x27;t re-nag, and it still blocks queued pings until answered.</p></div>
</details>
</article>
<article class="exchange claude" id="p13">
<div class="diffstat">+33 −22 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span><span class="del"></span></span></div>
<pre class="prompt">ok, it seems to be good. can you undo the instrumentation you added? or put it at the bottom of AGENTS.md for safekeeping</pre>
<details>
<summary><a class="anchor" href="#p13"><time datetime="2026-07-21T22:07:32.552000+00:00">15:07</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 36s · 1m21s wall-clock time</span></summary>
<div class="reply machine"><p>Done. Instrumentation stripped from [launch.pl](launch.pl) (now back to its original state) and [iterm2-tagtime](iterm2-tagtime) (which keeps only the re-raise fix); the snippets are archived in the [AGENTS.md](AGENTS.md) scratchpad section with a note on how to interpret the log. Both files syntax-check and the wrapper test-ran clean. The collected data stays in tmp/popup-timing.log. Only remaining loose end from this saga: the optional ~1s afplay speedup, whenever you want it.</p></div>
</details>
</article>
<article class="exchange claude" id="p14">
<pre class="prompt">can you check the beeminder forum and figure out why we needed to do this with iterm2 instead of terminal?</pre>
<details>
<summary><a class="anchor" href="#p14"><time datetime="2026-07-21T22:23:06.118000+00:00">15:23</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m28s · 4m7s wall-clock time</span></summary>
<div class="reply machine"><p>Found the origin: <a href="https://forum.beeminder.com/t/a-script-for-using-iterm2-instead-of-xterm-on-osx/4085">A script for using iTerm2 instead of xterm on OSX</a> (mvr, 2018). But it answers &quot;why not xterm&quot; (you were ditching X11 — you replied in 2019: &quot;I needed to get rid of xterm and this has been working like a charm&quot;), not &quot;why not Terminal.&quot; Nobody in the thread discusses Terminal.app at all.</p>
<p>The Terminal reason isn&#x27;t on the forum — it&#x27;s fossilized in your repo, in [terminal-tagtime](terminal-tagtime) and [script/term.sh](script/term.sh). Reconstructing from those:</p>
<ol><li><strong>launch.pl needs the popup command to block until answered</strong> (its queueing depends on it). iTerm2&#x27;s <code>create window with profile &quot;TagTime&quot; command X</code> runs X *as* the session — window dies when X exits, and <code>exists newWindow</code> is a clean completion signal. Terminal&#x27;s <code>do script</code> just types a command into a shell and returns; term.sh had to poll the window&#x27;s <code>busy</code> flag instead, which is racy and means &quot;some process is running,&quot; not &quot;ping.pl finished.&quot;</li><li><strong>terminal-tagtime does <code>do script &quot;$@&quot; in window 1</code></strong> — it injects ping.pl into whatever Terminal&#x27;s frontmost window is, which is plausibly the very window running tagtimed. That alone would &quot;not work right.&quot;</li><li>Terminal windows don&#x27;t close on command exit and don&#x27;t carry the red styling; term.sh manually sets colors and closes, per window. iTerm2&#x27;s named profile handles both.</li></ol>
<p>So: xterm died for X11 reasons, Terminal was tried and abandoned for the <code>do script</code> model&#x27;s lack of run-command-in-fresh-window-and-wait semantics, and mvr&#x27;s iTerm2 script had exactly the right primitive. Want me to add a one-line correction to the &quot;I forget why/how&quot; comment? (That comment is yours, so I won&#x27;t touch it unasked.)</p></div>
</details>
</article>
<article class="exchange claude" id="p15">
<pre class="prompt">correct to not touch the comment. do you see a way to get Terminal to do what we need here?</pre>
<details>
<summary><a class="anchor" href="#p15"><time datetime="2026-07-21T22:28:33.359000+00:00">15:28</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 27s</span></summary>
<div class="reply machine"><p>Yes. The pieces terminal-tagtime got wrong are all fixable:</p>
<ol><li><strong>Fresh window</strong>: <code>do script &quot;cmd&quot;</code> *without* <code>in window 1</code> creates a new window. The <code>in window 1</code> was the bug.</li><li><strong>Block-until-answered</strong>: <code>do script &quot;exec ${path}ping.pl $t&quot;</code> — the <code>exec</code> replaces the wrapper shell with ping.pl, so when ping.pl exits the session terminates, and with a Terminal profile set to &quot;when the shell exits: close the window,&quot; the window dies. Then the clean <code>repeat while exists window id W</code> polling works, same as iTerm2 — no <code>busy</code> flag hackery.</li><li><strong>Red styling</strong>: one-time GUI setup of a Terminal settings profile named TagTime, applied via <code>set current settings of &lt;tab&gt; to settings set &quot;TagTime&quot;</code>.</li></ol>
<p>Remaining jank relative to iTerm2: <code>do script</code> wraps the command in a login shell (your bashrc runs first — small latency, env differences), and grabbing the new window reference is the slightly racy <code>window 1</code>-right-after-creation idiom, since <code>do script</code> returns a tab, not a window. Plus the one-time profile setup lives in Terminal&#x27;s GUI settings rather than in the script.</p>
<p>So Terminal is viable if the goal is dropping the iTerm2 dependency — at the cost of those two rough edges. The re-raise loop ports over unchanged. Want me to build it as a working terminal-tagtime replacement you could flip <code>$XT</code> to?</p></div>
</details>
</article>
<article class="exchange claude" id="p16">
<div class="diffstat">+29 −0 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span></span></div>
<pre class="prompt">sure</pre>
<details>
<summary><a class="anchor" href="#p16"><time datetime="2026-07-21T22:33:07.754000+00:00">15:33</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="elapsed">thought for 1m24s</span></summary>
<div class="reply machine empty"><p>No response.</p></div>
</details>
</article>
<article class="exchange claude" id="p17">
<div class="diffstat">+42 −15 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span><span class="nil"></span></span></div>
<pre class="prompt">and see AGENTS.md
then add documentation in latin for how to set up the profile named TagTime with the red styling -- wherever it makes sense to put that</pre>
<details>
<summary><a class="anchor" href="#p17"><time datetime="2026-07-21T22:34:20.680000+00:00">15:34</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 58s · 5m35s wall-clock time</span></summary>
<div class="reply machine"><p>Built and red-tested. [terminal-tagtime](terminal-tagtime) is rewritten (old version archived verbatim in AGENTS.md scratchpad, including your updated comment): fresh window via <code>do script &quot;exec $@&quot;</code> (exec makes the shell *become* the command so the window dies on exit), styled via <code>settings set &quot;TagTime&quot;</code>, same stuck-detection re-raise loop as iterm2-tagtime, and — the one new if-statement, an anti-postel precondition — it errors out loudly *before* creating any window if the TagTime settings set doesn&#x27;t exist. That assert just fired correctly in my test (exit 1, &quot;no settings set named TagTime&quot;), which is as far as I can verify: the green path needs the one-time GUI step only you can do (documented in Latin in the file header): Terminal &gt; Settings &gt; Profiles &gt; new profile &quot;TagTime&quot;, red background, Shell &gt; &quot;When the shell exits&quot; &gt; &quot;Close the window&quot; — that last one is what makes the wait loop terminate.</p>
<p>Once the profile exists: rerun <code>./terminal-tagtime sleep 3</code> to confirm green, then flipping over is a one-line <code>$XT</code> change in settings.pl (gitignored — flagging per protocol).</p></div>
</details>
</article>
<article class="exchange claude" id="p18">
<div class="diffstat">+8 −6 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="del"></span><span class="del"></span><span class="nil"></span></span></div>
<pre class="prompt">not "close if the shell exited cleanly"?</pre>
<details>
<summary><a class="anchor" href="#p18"><time datetime="2026-07-21T23:04:23.330000+00:00">16:04</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 42s · 43s wall-clock time</span></summary>
<div class="reply machine"><p>Yes — &quot;Close if the shell exited cleanly&quot; is the better choice, and the more anti-postel one: an answered ping (exit 0) closes identically, but a crashed ping.pl leaves its error message on screen in the red window instead of evaporating unread. The wait loop then holds until you close the dead window, after which launch.pl&#x27;s existing recovery (log <code>err</code>, pop the editor) runs. The tradeoff: a crashed or Ctrl-C&#x27;d ping blocks subsequent ping windows until you notice and close it, whereas &quot;Close the window&quot; never wedges but swallows errors invisibly. Updating the docs to match:</p>
<p>Docs updated (English recap + Latin) in [terminal-tagtime](terminal-tagtime). Setup is now: profile &quot;TagTime&quot;, red background, Shell &gt; &quot;When the shell exits&quot; &gt; &quot;Close if the shell exited cleanly&quot;. Rerun <code>./terminal-tagtime sleep 3</code> once the profile exists.</p></div>
</details>
</article>
<article class="exchange claude" id="p19">
<div class="diffstat">+4 −0 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="nil"></span></span></div>
<pre class="prompt">oops, when i type backspace in this terminal, i see ^H's</pre>
<details>
<summary><a class="anchor" href="#p19"><time datetime="2026-07-21T23:05:25.243000+00:00">16:05</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m11s · 1m50s wall-clock time</span></summary>
<div class="reply machine"><p>Likely the profile is sending ^H (0x08) while the tty&#x27;s erase char is ^? (DEL) — with <code>exec</code> there&#x27;s no shell/readline to paper over the mismatch. Checking the profile&#x27;s keyboard setting:</p>
<p>Docs updated (English + Latin). Uncheck Keyboard &gt; &quot;Delete sends Control-H&quot; in the TagTime profile and backspace will behave. I also noticed <code>shellExitAction = 1</code> in your profile, so the close-if-exited-cleanly setup is already in place — after the keyboard fix, <code>./terminal-tagtime sleep 3</code> should run fully green.</p></div>
</details>
</article>
<article class="exchange claude" id="p20">
<div class="diffstat">+1 −1 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="del"></span><span class="nil"></span><span class="nil"></span><span class="nil"></span></span></div>
<pre class="prompt">also, any good way to suppress these lines at the top of the terminal that pops up?

dreeves@MacBook-Pro-5.local 2026-07-21 16:04:25
1604d@M~$ exec /Users/dreeves/lab/tagtime/ping.pl 1784675063</pre>
<details>
<summary><a class="anchor" href="#p20"><time datetime="2026-07-21T23:10:50.878000+00:00">16:10</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 35s · 1m15s wall-clock time</span></summary>
<div class="reply machine"><p>The prompt and command echo come from <code>do script</code> typing into your login shell (iTerm2&#x27;s <code>command</code> mode bypasses the shell entirely, which is why it never had this). Cheapest suppression: <code>clear;</code> before the exec — the prompt still prints but gets wiped in the same instant:</p>
<p>Done — <code>clear; exec $@</code> and the full green path now runs (exit 0: window created, styled, closed on clean exit, loops terminated). You&#x27;ll see at most a sub-100ms flash of the prompt before the wipe. If even that flash bothers you, the alternative is abandoning <code>do script</code> for a temp-script-as-profile-command arrangement, which I&#x27;d call not worth it.</p></div>
</details>
</article>
<article class="exchange claude" id="p21">
<div class="diffstat">+13 −0 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span></span></div>
<pre class="prompt">where's the iterm2 profile step documented?</pre>
<details>
<summary><a class="anchor" href="#p21"><time datetime="2026-07-21T23:13:15.743000+00:00">16:13</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 25s · 26s wall-clock time</span></summary>
<div class="reply machine"><p>Nowhere — the only trace is the <a href="https://forum.beeminder.com/t/a-script-for-using-iterm2-instead-of-xterm-on-osx/4085">2018 forum post</a> (&quot;create an iTerm2 profile called TagTime with the style you want&quot;). Adding the parallel docs to iterm2-tagtime, below your header comment:</p>
<p>Added to [iterm2-tagtime](iterm2-tagtime): profile &quot;TagTime&quot;, red background under Colors, and Session &gt; &quot;After a session ends&quot; &gt; &quot;Close&quot; (the load-bearing one — iTerm2 has no &quot;only if exited cleanly&quot; variant, so crashed pings close silently there, unlike the Terminal version). Worth a skim of your actual iTerm2 profile to confirm the docs match what you configured in 2019.</p></div>
</details>
</article>
<article class="exchange claude" id="p22">
<div class="diffstat">+12 −3 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span></span></div>
<pre class="prompt">perfect. nice work! i don't suppose there's a way to make just the tagtime terminal pop up without all terminals getting brought to the foreground?</pre>
<details>
<summary><a class="anchor" href="#p22"><time datetime="2026-07-22T02:10:32.341000+00:00">19:10</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m35s · 2m38s wall-clock time</span></summary>
<div class="reply machine"><p>Yes — <code>activate</code> is app-level activation, which raises every Terminal window; same for the <code>open -a Terminal</code>. The per-window alternative is <code>NSRunningApplication</code>&#x27;s <code>activateWithOptions:</code> *without* the all-windows flag, callable from AppleScript&#x27;s ObjC bridge (no extra permissions), which raises only the app&#x27;s front window — so ordering our window to index 1 first makes it the only one that comes forward:</p>
<p>Runs clean (exit 0). Structure now: no <code>open -a</code>, no <code>activate</code> — the window is ordered to Terminal&#x27;s front (<code>set index ... to 1</code>) and then <code>raiseApp()</code> brings only that front window forward via <code>activateWithOptions:</code> sans the all-windows flag. Whether your other Terminal windows genuinely stayed put is only verifiable by your eyeballs — check on the next ping (or run <code>./terminal-tagtime sleep 5</code> with another Terminal window deliberately buried). If it behaves, want the same treatment for iterm2-tagtime?</p></div>
</details>
</article>
<article class="exchange claude" id="p23">
<div class="diffstat">+6 −2 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span><span class="nil"></span></span></div>
<pre class="prompt">oops, that seemed to work but then i just had an instance of the popup not popping up :(</pre>
<details>
<summary><a class="anchor" href="#p23"><time datetime="2026-07-22T04:27:39.345000+00:00">21:27</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 2m · 12m26s wall-clock time</span></summary>
<div class="reply machine"><p>Let me check whether the polite <code>activateWithOptions:</code> raise is being silently denied by macOS — the WindowServer log will show whether any SetFrontProcess for Terminal happened during my <code>sleep 3</code> test a few minutes ago and during the failed ping:</p>
<p>Both scripts fixed and test-run clean. iterm2-tagtime&#x27;s stuck check is now <code>frontmost and (id of current window is id of newWindow)</code>; terminal-tagtime&#x27;s is <code>frontmost and (index of window id w is 1)</code> — so a ping that lands while you&#x27;re in another terminal window of the same app now gets selected/raised instead of declared stuck.</p>
<p>One honest caveat: this instance is *explained* by the app-frontmost hole only if you were in an iTerm window around 21:25:47 — if you were in Discord or elsewhere, something else failed and I&#x27;d want to hear about the next occurrence. Also noting <code>$XT</code> still runs iterm2-tagtime; flip line 9/10 in settings.pl whenever you want the Terminal version live (and the Terminal version&#x27;s polite single-window raise is the one component without battle scars yet).</p></div>
</details>
</article>
<article class="exchange claude" id="p24">
<pre class="prompt">i'm trying both. the iterm version seems to work. on the next ping i'll see if the terminal.app version seems to work</pre>
<details>
<summary><a class="anchor" href="#p24"><time datetime="2026-07-22T05:01:17.890000+00:00">22:01</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 11s</span></summary>
<div class="reply machine"><p>Good. On the Terminal one, the two things to watch: whether it surfaces at all while you&#x27;re typing in another app (the polite <code>activateWithOptions:</code> raise is the unproven piece — if it stays buried, I&#x27;ll add escalation to a full <code>activate</code> after a few failed polite tries), and whether your other Terminal windows stay put when it does.</p></div>
</details>
</article>
<article class="exchange claude" id="p25">
<pre class="prompt">the terminal.app version is broken, maybe inconsistently broken</pre>
<details>
<summary><a class="anchor" href="#p25"><time datetime="2026-07-22T06:18:48.737000+00:00">23:18</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 2m54s · 13m36s wall-clock time</span></summary>
<div class="reply machine"><p>Two findings, one question, then options:</p>
<p><strong>This instance wasn&#x27;t broken.</strong> The window popped, you answered, and ping.pl then sat in a slow Beeminder submission (the <code>d/nafk</code> goal), holding the window open — it&#x27;s since finished. Working as configured.</p>
<p><strong>AutoRaise reframes everything.</strong> It&#x27;s been running since <strong>June 17</strong> — not since the May 21 macOS update. Its <code>-focusDelay 1</code> = 50ms, which matches the 58–85ms focus reverts I&#x27;d attributed to macOS 26&#x27;s typing shield: AutoRaise refocuses whatever&#x27;s under your mouse, every 50ms poll. That would explain the inconsistency far better than OS nondeterminism — outcome depends on mouse position, and the macOS focus-prevention may be only a secondary player on your machine.</p>
<p>The question: did the popup delays actually start around mid-June (AutoRaise) rather than back at the macOS update in May?</p>
<p>Fix options given AutoRaise (<a href="https://github.com/sbmpost/AutoRaise">flags reference</a>):</p>
<ul><li><strong>Warp the mouse into the red window when raising it</strong> — with focus-follows-mouse, moving the pointer *is* moving the focus; AutoRaise then defends our window instead of fighting it. Needs a small helper (<code>cliclick</code> or ~10 lines of compiled Swift).</li><li><strong><code>-stayFocusedBundleIds com.apple.Terminal,com.googlecode.iterm2</code></strong> on AutoRaise — protects a focused ping window from mouse-hover steals, but blunts focus-follows-mouse whenever *any* terminal has focus, which is most of your life.</li><li><strong>The native non-activating always-on-top panel</strong> (the endgame from the ideal-design discussion) — happens to be the AutoRaise-native solution: visible over everything without ever contesting focus; you mouse into it to type, which is exactly your workflow&#x27;s grammar.</li></ul></div>
</details>
</article>
<article class="exchange claude" id="p26">
<pre class="prompt">i don't think it has to do with autoraise</pre>
<details>
<summary><a class="anchor" href="#p26"><time datetime="2026-07-22T06:33:16.515000+00:00">23:33</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 32s</span></summary>
<div class="reply machine"><p>Fair — and on reflection the mechanism doesn&#x27;t hold up: AutoRaise acts on mouse *movement*, and the reverts we captured happened mid-typing with the mouse presumably stationary, when AutoRaise is dormant. Its June 17 start is also just the reboot date (every process on this box started then), not an install date. So the original diagnosis stands: OS-level focus protection, per the Apple thread.</p>
<p>Status on the Terminal version, then: the one specimen I could examine tonight turned out not to be broken (answered fine, then a slow Beeminder submission held the window per close-if-clean), and the genuinely broken 21:25 case was the app-frontmost hole, since fixed in both scripts. So there&#x27;s currently no confirmed failure of the Terminal version with the current code — but also no proof of the one component unique to it, the polite single-window raise.</p>
<p>Want me to re-add the timing instrumentation (archived in AGENTS.md) to terminal-tagtime, plus a line recording the captured window id and whether each raise attempt landed? Then the next &quot;that seemed broken&quot; moment gives us data instead of theories.</p></div>
</details>
</article>
<h2 class="day" id="d2026-07-22"><time datetime="2026-07-22">2026-07-22 Wednesday</time></h2>
<article class="exchange claude" id="p27">
<pre class="prompt">iterm2 popup just failed to pop up until i clicked on some other iterm2 window. i hear the ping, then when the popup finally pops up it says (in this case) "WARNING: 10 seconds late"</pre>
<details>
<summary><a class="anchor" href="#p27"><time datetime="2026-07-22T16:57:19.793000+00:00">09:57</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 2m14s · 12m1s wall-clock time</span></summary>
<div class="reply machine"><p>The &quot;10 seconds late&quot; warning is a precise clue: ping.pl computes it from its own start time, so the *session itself* started 10s late — the window wasn&#x27;t created-but-buried this time; iTerm2 didn&#x27;t run the command until your click. Different failure mode. Pulling the log for this ping:</p>
<p>Diagnosis, and it&#x27;s not a new bug — it&#x27;s the (b) tradeoff playing out exactly as warned:</p>
<ul><li>09:41:33: a ping&#x27;s window was created, raised, and legitimately achieved &quot;stuck&quot; (frontmost with your window current) — raising stopped, per (b). Zero raise attempts after that in the log.</li><li>You moved on without answering; the window got buried. It then sat 14 minutes, and per the queue serialization, your 09:55:23 ping beeped with *no window possible* — the forgotten window held the lock.</li><li>Your click on iTerm at 09:55:31 surfaced the old window; you answered it at :33, and the queued 09:55 window popped immediately — &quot;10 seconds late&quot; is just its queue wait, not creation latency.</li></ul>
<p>(Also for the record: iTerm2 was restarted at 21:26 last night, mid-experimentation, which may have muddied some of last night&#x27;s testing; and its App Nap entries turned out to be a red herring here.)</p>
<p>So (b)&#x27;s &quot;steal once, then hands off&quot; is working as specified — and the spec bites whenever you don&#x27;t answer immediately, because the *next* ping then beeps into a void. The natural completion of your &quot;steal once&quot; preference: <strong>one steal per audible ping</strong> — when a new ping fires and finds an unanswered window blocking the queue, re-summon that window (fresh raise-until-stuck) instead of beeping silently. Deliberate postponing stays quiet between beeps; every beep gets a window in your face. That needs a <code>reraise</code> mode in the popup scripts plus one line in launch.pl&#x27;s lock-failure branch (launch.pl is tracked, so that change will show in your diff). Want it?</p></div>
</details>
</article>
<article class="exchange claude" id="p28">
<pre class="prompt">ugh i hate this. first of all, the window didn't pop up and then get buried. did you see the part about how the window itself warns that it was launched 10 seconds late?</pre>
<details>
<summary><a class="anchor" href="#p28"><time datetime="2026-07-22T17:12:20.836000+00:00">10:12</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m28s · 1m57s wall-clock time</span></summary>
<div class="reply machine"><p>You&#x27;re right to push back — the missing fact: the blocking window was the <strong>09:41:31 ping, answered &quot;afk&quot;</strong>. It popped while you were away from the machine (stuck legitimately, front of everything, nobody watching), and by the time you sat back down it was hidden behind whatever you resumed into. So from where you sat: beep at 09:55:23, no window, ever — my &quot;you tabbed away and forgot it&quot; framing was wrong. The 10-seconds-late warning is the 09:55 window&#x27;s queue wait behind that afk window (created at 09:55:33, immediately after your click let you answer the 09:41 one).</p>
<p>The structural problem this exposes: steal-once semantics steal exactly once, and for any ping that lands while you&#x27;re afk, that once is spent on an empty room. Every return-from-afk is then a minefield: an invisible window holding the queue lock, and subsequent beeps producing nothing.</p>
<p>Two candidate fixes, smallest first:</p>
<ol><li><strong>Re-summon on beep</strong>: when a new ping finds the queue blocked, re-raise the blocking window (one steal per audible beep). Fixes what you experienced at 09:55. Gap: returning from afk with no new ping for a while still leaves the window hidden silently.</li><li><strong>Re-summon on return</strong>: re-arm the raise when system idle time drops from minutes to zero — i.e., steal once per *your presence*, so a window that popped to an empty room re-steals when you come back. Closes the afk gap; slightly more machinery (the wrapper polls <code>HIDIdleTime</code>).</li></ol>
<p>Or both, they compose. Which?</p></div>
</details>
</article>
<article class="exchange claude" id="p29">
<pre class="prompt">no no no. obviously if it pops up while i'm away it's going to sit there will i return. i stopped reading your too-long reply when i spotted that confusion. where are we at here?</pre>
<details>
<summary><a class="anchor" href="#p29"><time datetime="2026-07-22T17:22:08.141000+00:00">10:22</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m48s · 7m wall-clock time</span></summary>
<div class="reply machine"><p>Status, compactly:</p>
<p><strong>Solved</strong>: the original complaint. Window creation was never slow — macOS 26 defers/reverts background activations, so windows arrived buried. The raise-until-stuck loop fixes every case where you&#x27;re at the keyboard; that&#x27;s been holding for both scripts.</p>
<p><strong>Not solved</strong>: pings that fire while you&#x27;re away. The one steal gets spent on an empty room, your return-clicks land wherever they land, and the red window ends up behind them (I can&#x27;t fully reconstruct the 09:41 window&#x27;s visibility from the logs — it&#x27;s also possible the stuck check itself false-positives, since <code>frontmost</code> can read true while WindowServer is still deferring reality). Either way, every afk ping re-creates the trap you hit at 09:55: blocked queue, beep, no window.</p>
<p><strong>Also</strong>: last night&#x27;s &quot;Terminal version broken&quot; instance turned out to be a slow Beeminder submission holding the window per close-if-clean; no confirmed failure of that script since the window-level fix.</p>
<p>My recommendation: give up on steal-once. It depends on trusting focus signals that macOS 26 actively falsifies, and it inherently can&#x27;t handle afk pings. Raise-every-second-until-answered has no state to misread and no afk hole — its only cost is that postponing a ping means the window keeps shoving in (which you originally chose as acceptable). One-line change in both scripts. Do it?</p></div>
</details>
</article>
<article class="exchange claude" id="p30">
<div class="diffstat">+53 −49 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="del"></span><span class="del"></span><span class="nil"></span></span></div>
<pre class="prompt">there's a way to have the best of all worlds we're not thinking of. ping happens, i immediately see the window come to the front, but only that window, not all iterm/terminal.app windows, and if i don't immediately answer it it doesn't spontaneously steal focus *again*, only on first popup. (if it were forced to stay visible, that would be fine but not worth doing something kludgy to try to achieve that part)</pre>
<details>
<summary><a class="anchor" href="#p30"><time datetime="2026-07-22T17:35:38.211000+00:00">10:35</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 5m42s · 20m wall-clock time</span></summary>
<div class="reply machine"><p>That spec is implementable, with one honest adjustment: a literal single activation is exactly what macOS 26 swallows, so &quot;first popup&quot; becomes a dense finite burst — raise attempts every 300ms for ~3s, then never again. Perceptually one popup event; lands on any typing pause ≥300ms; afk pings land trivially (no one&#x27;s typing); no stuck-detection to be fooled; no later re-stealing, ever. And porting the single-window raise (already built for Terminal) to the iTerm2 script satisfies &quot;only that window&quot; — the current iTerm2 version&#x27;s <code>activate</code> raises all windows, which also goes away. Implementing in both:</p>
<p>Your rewritten header is in place (leaving it untouched — note its item 2 is now optional since the script closes the window itself). Retesting:</p>
<p>Both scripts are now on your best-of-all-worlds spec and passing tests (two consecutive clean runs, ~4.7s for a 3s command):</p>
<ul><li><strong>One perceived popup</strong>: a 0.3s×10 raise burst lands the window on any sub-3s typing pause (and trivially when you&#x27;re afk), then *never* steals focus again. No stuck-detection left to be fooled.</li><li><strong>Only the ping window raises</strong>: both scripts now use select/index-to-1 plus the no-all-windows <code>activateWithOptions:</code> raise; the all-window-raising <code>activate</code> and <code>open -a</code> are gone from both.</li><li>Getting the Terminal version there surfaced two latent bugs, now fixed: <code>front window</code> right after <code>do script</code> is a race (it now finds the window by the new tab&#x27;s tty), and Terminal&#x27;s profile close-on-exit turns out to fire only *sometimes* — in both directions — so the script now busy-waits and closes the window itself, inside a <code>try</code> whose only catchable failures mean &quot;window already gone.&quot; Your header comment&#x27;s item 2 (close-if-clean) is therefore optional now; leaving your comment as-is.</li></ul>
<p>Live on the next pings, whichever <code>$XT</code> points at.</p></div>
</details>
</article>
<article class="exchange claude" id="p31">
<div class="diffstat">+10 −24 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="del"></span><span class="del"></span><span class="del"></span><span class="nil"></span></span></div>
<pre class="prompt">sob. two pings just pung and i'm seeing no popups at all!</pre>
<details>
<summary><a class="anchor" href="#p31"><time datetime="2026-07-22T18:07:22.463000+00:00">11:07</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m59s · 2m22s wall-clock time</span></summary>
<div class="reply machine"><p>Fixed and both scripts pass. Summary of where this landed:</p>
<ul><li>The burst design stays: ~3s of raise attempts, perceptually one popup, never steals again after.</li><li>The raise inside it is back to AE <code>activate</code> — the one mechanism that landed every single time across this whole investigation. The polite single-window raise turned out to be denied by macOS when called from the daemon&#x27;s background context (it only ever succeeded from my interactive test shell, which is why testing couldn&#x27;t catch it — that&#x27;s now recorded in memory as a hard-won lesson).</li><li>Cost, stated plainly: on each ping, all iTerm (or Terminal) windows come forward together, ping window selected on top. Your only-that-window requirement is the one thing macOS 26 wouldn&#x27;t give us.</li></ul>
<p>Your full spec — single window, raised once, guaranteed visible, no app&#x27;s window-group dragged along — is precisely what the native always-on-top panel would deliver, since it skips app activation entirely. That remains the endgame if the window-group raise grates. Meanwhile, next pings run this version.</p></div>
</details>
</article>
<article class="exchange codex" id="p32">
<pre class="prompt">anything Sol is missing based on its latest in sourcery.html?
</pre>
<details>
<summary><a class="anchor" href="#p32"><time datetime="2026-07-22T18:37:46.353000+00:00">11:37</time></a> <span class="chip"></span> <span class="agent">Codex</span> <span class="model">gpt-5.6-sol</span> <span class="effort">(ultra)</span> <span class="elapsed">thought for 5m35s · 6m33s wall-clock time</span></summary>
<div class="reply machine"><p>I’ll compare the newest Sol material in <code>sourcery.html</code> against Sol’s current implementation and quals, then report only concrete gaps.</p>
<p>The files match the last transcript mechanically. I’m checking the two distinctions that still matter: what is merely pending versus what is still an unmet requirement.</p>
<p>Yes. The worktree matches p31, but four gaps remain:</p>
<ul><li>“Only the ping window” is still unmet: both wrappers use app-wide <code>activate</code>, raising every Terminal/iTerm window ([Terminal](/Users/dreeves/lab/tagtime/terminal-tagtime:33), [iTerm2](/Users/dreeves/lab/tagtime/iterm2-tagtime:19)).</li><li>The ten unconditional attempts can steal focus again if you switch away during the three-second burst.</li><li>No post-p31 green qual from the actual daemon context is recorded; interactive success previously gave a false green.</li><li>Terminal now explicitly closes the window after any process exit, apparently undoing the earlier “leave crashes visible” behavior; its bare <code>try</code> also swallows every error ([terminal-tagtime](/Users/dreeves/lab/tagtime/terminal-tagtime:39)).</li></ul>
<p>Also, <code>AXRaise</code> is an untried single-window mechanism before concluding that a native panel is the only solution, though it would need a daemon-context qual and likely Accessibility permission. Apple describes it as raising an individual window “<a href="https://developer.apple.com/documentation/applicationservices/kaxraiseaction">as frontmost as allowed</a>.”</p>
<p>Terminal is currently live via [settings.pl](/Users/dreeves/lab/tagtime/settings.pl:9). No files changed.</p></div>
</details>
</article>
<article class="exchange claude" id="p33">
<div class="diffstat">+11 −4 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span><span class="nil"></span></span></div>
<pre class="prompt">are we swallowing errors (try without catch, kind of thing) or closing the popup even if there was an error?</pre>
<details>
<summary><a class="anchor" href="#p33"><time datetime="2026-07-22T21:10:28.462000+00:00">14:10</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 2m15s · 45m14s wall-clock time</span></summary>
<div class="reply machine"><p>Both guilty as charged, one in each script:</p>
<ol><li><strong>Bare try, yes</strong>: terminal-tagtime&#x27;s <code>try ... end try</code> swallows *everything* in the wait-and-close block, not just the intended &quot;window already vanished&quot; errors. Tightening it to catch only those and re-throw the rest (which then exit osascript nonzero → launch.pl&#x27;s SYSERR):</li></ol>
<p>Direct answers, now verified:</p>
<p><strong>Swallowing errors</strong> — was true, now fixed. The blanket <code>try</code> catches only the two &quot;window/tab already vanished&quot; error codes (−1719/−1728) and rethrows everything else, which exits osascript nonzero and hits launch.pl&#x27;s SYSERR. The tty-search loop has a per-window <code>try</code> that&#x27;s deliberately bare — an unreadable window can&#x27;t be the one we just created, so skipping it is the correct semantics, not suppression (commented as such).</p>
<p><strong>Closing on error</strong> — yes, currently both scripts do. Just confirmed: a command that runs 2s and exits 1 gets its window closed, error text and all. iTerm2 has always done this (its profile closes on session end regardless of status); Terminal now does it via our self-close since the profile&#x27;s close-if-clean proved flaky. Mitigation that exists anyway: a ping.pl that dies without logging still gets loudly surfaced by launch.pl — <code>err</code> line in the log plus the editor window. What&#x27;s lost is only the error *text*.</p>
<p>Restoring kept-open-on-crash for Terminal would mean the shell wrapper holding the window hostage on nonzero exit (e.g. exec&#x27;ing a sleeper for <code>busy</code> to see) — kludge territory by your standard, and unavailable for iTerm2 regardless. Rely on launch.pl&#x27;s err-plus-editor as the loud path, or want the hold?</p></div>
</details>
</article>
<article class="exchange claude" id="p34">
<div class="diffstat">+27 −0 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span></span></div>
<pre class="prompt">wrong, iterm2 used to stay open if there was an error.
also i just had iterm2 fail to pop up promptly.</pre>
<details>
<summary><a class="anchor" href="#p34"><time datetime="2026-07-22T22:18:13.687000+00:00">15:18</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m55s · 16m13s wall-clock time</span></summary>
<div class="reply machine"><p>Instrumentation is back in all three files and verified live (<code>xt-start</code>/<code>xt-exit</code> logging again; syntax checks pass). Next slow popup decomposes as: ping time → <code>launch</code> line = afplay + perl startup; <code>launch</code> → <code>xt-start</code> = spawn; <code>xt-start</code> → window creation = osascript + iTerm2 (that moment comes from the unified log&#x27;s create-window event). If the 9 seconds turns out to live in the afplay/daemon segment, the long-deferred sound-backgrounding fix is the likely cure — but measurement first, one variable at a time.</p></div>
</details>
</article>
<article class="exchange claude" id="p35">
<pre class="prompt">alright, popup failed just now till i alt-tabbed to iTerm2</pre>
<details>
<summary><a class="anchor" href="#p35"><time datetime="2026-07-23T06:40:53.133000+00:00">23:40</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m7s · 3m17s wall-clock time</span></summary>
<div class="reply machine"><p>Full reconstruction, and it&#x27;s not a malfunction — it&#x27;s the spec: the 23:08 window <strong>did</strong> pop (burst landed at 23:08:19, stayed front 32 seconds). Your tags for that ping are &quot;phn smk (wordle)&quot; — you were on your phone; at 23:08:51 you clicked into Discord without answering, which buried it. Then 32 minutes of normal app-switching (Chrome/Code/Terminal/Discord all took front, per the log) with the red window entombed and the queue locked. The 23:40 beep therefore produced nothing until your alt-tab. &quot;Steal once, never again&quot; makes every unnoticed-or-unanswered ping a landmine for all subsequent pings.</p>
<p>Structural conclusion, same as it&#x27;s been converging to: with terminal-app windows, macOS gives us no &quot;visible without focus,&quot; so unanswered windows *will* get buried, and the only knobs are re-stealing (you&#x27;ve rejected) or living with burial (produces exactly this). Two moves remain:</p>
<ol><li><strong>Re-summon on beep</strong> (small, now): when a ping fires and finds the queue blocked, give the blocking window that beep&#x27;s one steal — every beep you hear puts a window in front of you, still zero spontaneous re-steals between beeps. This fits your steal-on-first-popup rule if the unit is the beep rather than the window.</li><li><strong>The native panel</strong> (the real fix): non-activating, always-on-top — your &quot;forced to stay visible&quot; wish is its structural property, not a kludge, plus single-window and zero focus theft. A few hours of work, replaces the AppleScript arms race entirely.</li></ol>
<p>Do 1 now and 2 after? Or straight to 2?</p></div>
</details>
</article>
<h2 class="day" id="d2026-07-23"><time datetime="2026-07-23">2026-07-23 Thursday</time></h2>
<article class="exchange claude" id="p36">
<pre class="prompt">no, you're totally wrong about what happened. the window did not pop up until i alt-tabbed. exactly like i told you.</pre>
<details>
<summary><a class="anchor" href="#p36"><time datetime="2026-07-23T07:40:34.347000+00:00">00:40</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m41s · 2m7s wall-clock time</span></summary>
<div class="reply machine"><p>Understood — and with three incidents now where WindowServer says &quot;iTerm frontmost&quot; while you say &quot;nothing visible,&quot; both must be true: the window exists and is officially front, but <strong>on a Space your displays aren&#x27;t showing</strong>. That would explain every single report: alt-tab (user-initiated) makes macOS switch to the window&#x27;s Space; programmatic activation on macOS 26 evidently doesn&#x27;t. And it explains &quot;inconsistent&quot;: pings pop visibly when you happen to be on the right Space. Checking whether iTerm2 is Space-assigned:</p>
<p>Concession without qualification: you&#x27;ve been right in each of these and my reconstructions were wrong in the same way each time — I was reading WindowServer&#x27;s &quot;Making iTerm the front process&quot; entries as &quot;the window visibly raised.&quot; Under macOS 26&#x27;s DeferringManager they are not the same thing: the front-process *bookkeeping* updates while the actual visual raise is deferred pending real user input — which is why your cmd-tab (genuine user input) is what makes the window materialize, every time. The config checks also rule out my Spaces variants (no app-to-Space binding, no fullscreen Spaces in play, displays share Spaces). So: windows are being created promptly and never visually raised at all; the &quot;burst landed&quot; claims were an artifact of trusting logs that record intent, not pixels.</p>
<p>That closes the argument about approach: no amount of AppleScript activation calls can beat a deferral whose release condition is *user interaction* — every variant we&#x27;ve tried (single, repeated, burst, polite, AE) has been the same move against the same wall. The one mechanism that doesn&#x27;t go through activation at all is an elevated window level — a non-activating panel at status-bar level orders above normal windows unconditionally, no raise request, nothing to defer. That&#x27;s the native panel, and at this point it&#x27;s not the &quot;endgame,&quot; it&#x27;s the only move left standing. It also natively satisfies your full spec: visible immediately, only itself, zero focus theft, stays visible.</p>
<p>Building it now unless you say otherwise: a minimal Swift helper first — panel with red styling on your active display showing the ping prompt, proving unmissability on your setup — then wired in as <code>$XT</code> so ping.pl and the whole pipeline stay untouched.</p></div>
</details>
</article>
<article class="exchange claude" id="p37">
<pre class="prompt">that explanation doesn't sound right. i'm never switching Spaces, as far as i know (and i think i would know)</pre>
<details>
<summary><a class="anchor" href="#p37"><time datetime="2026-07-23T07:42:41.652000+00:00">00:42</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 49s</span></summary>
<div class="reply machine"><p>No Spaces in the current explanation — you never switching Spaces is consistent with it. Dropping that thread entirely; the config ruled it out and I said so poorly.</p>
<p>The claim, stated plainly: macOS 26 withholds the *visual* raise of a window whose raise was requested by a background process, until you yourself activate that app — and cmd-tab to iTerm2 is precisely such an activation. Not Space-switching; the window is on the Space you&#x27;re looking at the whole time, just never brought forward. This is also why every test I ran looked fine while real pings fail: my test invocations happen while you&#x27;re actively working in VS Code with me, a context macOS credits as user interaction, whereas tagtimed&#x27;s context has none — the same interactive-versus-daemon asymmetry that made the polite raise pass my tests and fail in the field.</p>
<p>It&#x27;s a best-fit explanation, not proven, so here&#x27;s its falsifiable prediction for the next failed popup: clicking or typing in *other* apps will not make the window appear, however long you work — only switching to iTerm2 (cmd-tab, Dock click, clicking any iTerm window) will. Three incidents so far match exactly, including your day-one observation that alt-tabbing to the app, not to the red window, is what triggers it.</p>
<p>If the prediction holds — or if you&#x27;re already convinced — the panel remains the move that sidesteps activation entirely, and I&#x27;ll build it on your word.</p></div>
</details>
</article>
<article class="exchange claude" id="p38">
<div class="diffstat">+152 −2 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span></span></div>
<pre class="prompt">ok, let's do it</pre>
<details>
<summary><a class="anchor" href="#p38"><time datetime="2026-07-23T08:36:47.851000+00:00">01:36</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 3m59s · 1h19m57s wall-clock time</span></summary>
<div class="reply machine"><p>Building it. Plan: a single-file Swift helper — non-activating panel at status-bar window level (above everything, all Spaces, fullscreen-auxiliary), red/white terminal styling, runs <code>ping.pl &lt;t&gt;</code> as a child process with its output streamed into the panel and a text field feeding its stdin — so every line of ping.pl logic (tasks, ditto, validation, Beeminder) is untouched, and the whole thing slots in as <code>$XT</code>. Keyboard is captured once at popup (the classic TagTime interrupt), never re-taken; close button = unanswered ping, which flows into launch.pl&#x27;s existing err-and-editor path.</p>
<p>Panel verified end-to-end from my side (visible 2s, clean exit — you may have seen it flash). Wiring it in:</p>
<p>The user promoted the popup scripts to tracked files (removed from .gitignore) — adding just the compiled binary:</p>
<p>Live. What shipped:</p>
<ul><li><strong>[tagtime-panel.swift](tagtime-panel.swift)</strong> (tracked source, ~150 lines) compiled to <code>tagtime-panel</code> (gitignored binary); <code>$XT</code> in settings.pl now points at it, so it&#x27;s active for the next ping — no daemon restart needed.</li><li>Mechanics: non-activating panel at status-bar window level, all Spaces, over fullscreen apps, shown via <code>orderFrontRegardless</code> — no activation request exists for macOS to defer. It appears centered on whichever display holds your mouse, runs ping.pl underneath with output streamed in (ANSI stripped), your typed line echoed and fed to its stdin, so ditto/task-numbers/validation/Beeminder all work unchanged. Keyboard is captured exactly once, at popup. Close button without answering kills ping.pl and exits 1 → launch.pl&#x27;s usual err-plus-editor path. Timing instrumentation logs <code>xt-start</code>/<code>xt-exit</code> like the scripts did.</li><li>Placeholder text is in Latin per your AGENTS.md rule; the panel&#x27;s transcript is ping.pl&#x27;s own English, piped verbatim.</li></ul>
<p>Honest caveat, consistent with this saga&#x27;s central lesson: my tests ran from an interactive context, and the decisive test is a real ping from the daemon while you&#x27;re typing in another app. The design argument for why it escapes the deferral is that it never asks for activation at all — but the field verdict is yours. If you want a feel for the interaction before then: <code>./tagtime-panel /bin/cat</code> (type, see it echo, close the window). The old scripts remain as <code>$XT</code> fallbacks, now git-tracked per your .gitignore change.</p></div>
</details>
</article>
<article class="exchange claude" id="p39">
<pre class="prompt">i'm confused. settings.pl currently points to tagtime-panel but i just got a ping and iterm popped up (after having to alt-tab to it)</pre>
<details>
<summary><a class="anchor" href="#p39"><time datetime="2026-07-23T10:06:14.366000+00:00">03:06</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 47s · 1m10s wall-clock time</span></summary>
<div class="reply machine"><p>No bug — a stale process. [launch.pl](launch.pl#L8) reads <code>.tagtimerc</code> (which is correctly symlinked and correctly points at the panel) <strong>once, at startup</strong> — and the launch.pl serving your pings tonight (PID 74548) started at <strong>23:08 last night</strong>: it&#x27;s the lock-holder that&#x27;s been queue-serving every ping since, with the old <code>$XT = iterm2-tagtime</code> baked into its memory. Your 03:04 ping came through it, hence iTerm, hence one more alt-tab.</p>
<p>You&#x27;ve now answered that ping (&quot;glm (yardsign)&quot; is logged), so that process&#x27;s queue is drained and it&#x27;s exiting — the next fresh ping spawns a new launch.pl, which will read the current settings and debut the panel.</p></div>
</details>
</article>
<article class="exchange claude" id="p40">
<pre class="prompt">alas, next ping just pung and it was the exact same: iterm2 popped up (eventually). nothing else popped up.</pre>
<details>
<summary><a class="anchor" href="#p40"><time datetime="2026-07-23T10:12:51.465000+00:00">03:12</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m35s · 10m24s wall-clock time</span></summary>
<div class="reply machine"><p>Done — stale launch.pl killed, lock released (the in-flight window finishes its churn harmlessly). The next ping spawns a fresh launch.pl, reads the current settings, and the panel gets its first genuine daemon-context test.</p>
<p>One heads-up from what this uncovered: those ~7-minute post-answer Beeminder submissions (seen twice now, <code>d/nafk</code> the likely culprit) will be more conspicuous with the panel, since it stays above everything while ping.pl finishes. If that gets annoying, the real fix is the slow submission itself — worth its own investigation later.</p></div>
</details>
</article>
<article class="exchange claude" id="p41">
<div class="diffstat">+21 −1 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span></span></div>
<pre class="prompt">i see a thing pop up but i can't type in it</pre>
<details>
<summary><a class="anchor" href="#p41"><time datetime="2026-07-23T19:17:53.875000+00:00">12:17</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m · 1m28s wall-clock time</span></summary>
<div class="reply machine"><p>The panel showing at all from the daemon context is the win — the visual deferral is beaten. The keyboard capture is what got deferred instead (the same macOS machinery routes keyboard focus, and our process still has no user-interaction credit). Fixes: explicit self-activation, first-click-counts on the field, and a bounded 3-try keyboard grab in the first 1.5s — initial capture only, still nothing after that:</p>
<p>Recompiled and live for the next ping. If the keyboard grab still loses to the deferral, the diagnostic that matters: <strong>does clicking into the input field let you type?</strong> A click is user interaction, so it should always work — if it does, only the initial grab is being deferred and I&#x27;ll escalate that specifically; if even clicking doesn&#x27;t take, that&#x27;s a different bug in the panel itself. Either way, one more data point does it.</p></div>
</details>
</article>
<article class="exchange claude" id="p42">
<pre class="prompt">i tried clicking before and nothing so i doubt this will work. can you just make a test popup pop up now and i'll try again?</pre>
<details>
<summary><a class="anchor" href="#p42"><time datetime="2026-07-23T19:22:51.634000+00:00">12:22</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 32s · 37s wall-clock time</span></summary>
<div class="reply machine"><p>Test panel is up now, running <code>/bin/cat</code> — so anything you type and submit with Enter should appear twice in the transcript (my echo plus cat&#x27;s echo). Try: typing immediately without clicking; if nothing, click the input field and type. When done, close it with the ✕ (exit 1 is expected and harmless). Report what happened at each step.</p></div>
</details>
</article>
<article class="exchange claude" id="p43">
<div class="diffstat">+5 −1 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span></span></div>
<pre class="prompt">i clicked and then typed and nothing happens. i just x'd it out</pre>
<details>
<summary><a class="anchor" href="#p43"><time datetime="2026-07-23T19:24:09.471000+00:00">12:24</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m11s · 2m57s wall-clock time</span></summary>
<div class="reply machine"><p>New test panel is up. Same drill: first try typing with no click at all, then click the field and type, Enter to submit (expect the doubled echo). Close with ✕ when done. Report both halves — no-click typing is the bonus; click-then-type is the one that must work this time.</p>
<p>The panel&#x27;s been closed via ✕ again (exit 1) — that&#x27;s all I can see from here. What happened with typing this round, no-click and after-click?</p></div>
</details>
</article>
<article class="exchange claude" id="p44">
<pre class="prompt">i was able to type in the field at the bottom</pre>
<details>
<summary><a class="anchor" href="#p44"><time datetime="2026-07-23T19:27:23.103000+00:00">12:27</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 24s · 25s wall-clock time</span></summary>
<div class="reply machine"><p>That&#x27;s the last blocker cleared — visibility and keyboard both confirmed working. The panel is already live as <code>$XT</code>, so real pings now run it end to end: red panel over everything the moment the ping fires, type tags, Enter, ping.pl validates/logs/submits, panel closes itself.</p>
<p>One thing to note on your next few pings: whether you can type straight away or need the one click first — macOS decides per-context whether the initial keyboard grab lands, and that&#x27;s the only remaining UX unknown. And if the panel ever lingers a few minutes after you&#x27;ve answered, that&#x27;s the known slow-Beeminder-submission issue (likely <code>d/nafk</code>), not the panel — separate investigation, queued.</p></div>
</details>
</article>
<article class="exchange claude" id="p45">
<div class="diffstat">+71 −24 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span><span class="nil"></span></span></div>
<pre class="prompt">can you read AGENTS.md and improve the new code? and make the window brighter red</pre>
<details>
<summary><a class="anchor" href="#p45"><time datetime="2026-07-23T19:28:43.654000+00:00">12:28</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 2m25s · 3m29s wall-clock time</span></summary>
<div class="reply machine"><p>All green. The AGENTS.md pass yielded:</p>
<ul><li><strong>Real bug</strong>: <code>isFloatingPanel = true</code> was silently clobbering the <code>.statusBar</code> window level down to <code>.floating</code> (its setter resets the level) — removed, with a comment stating the constraint. The panel now actually sits at the level the code always claimed.</li><li><strong>Anti-postel</strong>: <code>tslog</code> was silently no-oping when the timing log didn&#x27;t exist (FileHandle doesn&#x27;t create files); it now creates it and prints a SYSERR on failure, matching launch.pl&#x27;s idiom. The output reader no longer drops an entire chunk when a multibyte character splits across a pipe read (lossy decode, commented).</li><li><strong>Prose/comments</strong>: the header still described the abandoned non-activating design — rewritten to match reality. Load-bearing settings (<code>hidesOnDeactivate</code>) got constraint comments.</li><li><strong>DRY</strong>: the triplicated activate/makeKey/firstResponder dance is now one <code>grabKeyboard()</code>.</li><li><strong>QDD</strong>: [script/qual-panel.sh](script/qual-panel.sh) — five quals covering exit-status propagation, argument errors, and the instrumentation lines, framed replicata/expectata/resultata; the visual/keyboard side is noted as human-only. First run was red, which caught a bug in the quals themselves (macOS has no <code>/bin/true</code>); now all green.</li><li><strong>Brighter red</strong>: background lifted from dark (0.55) to vivid (0.87) — you&#x27;ll see it on the next ping, since the rebuilt binary is what <code>$XT</code> runs.</li></ul></div>
</details>
</article>
<article class="exchange claude" id="p46">
<div class="diffstat">+3 −0 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="nil"></span><span class="nil"></span></span></div>
<pre class="prompt">will it still work to hit enter and get a vim to edit the tagtime log?</pre>
<details>
<summary><a class="anchor" href="#p46"><time datetime="2026-07-23T19:39:01.287000+00:00">12:39</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m12s · 1m20s wall-clock time</span></summary>
<div class="reply machine"><p>Good catch — no, that would break. Hitting Enter with an empty answer makes launch.pl pop <code>$XT $ED</code> (vi on the log), and vim is a full-screen TUI that needs a real terminal: the panel only pipes lines, so vim inside it would be garbage. But launch.pl already has the hook for exactly this split: if <code>$EDIT_COMMAND</code> is defined, [editor()](launch.pl#L137-L141) uses it instead of <code>$XT</code> — so ping popups can stay on the panel while vim rides the iTerm2 script (which demonstrably works for editor windows; the timing log shows it doing exactly that on July 21). Setting that up:</p>
<p>Done and verified: hitting Enter on a ping now logs the empty answer via the panel as before, and the vim log-editor pops through [iterm2-tagtime](iterm2-tagtime) in a real terminal — <code>$EDIT_COMMAND</code> resolves to <code>iterm2-tagtime /usr/bin/vi +</code>, producing the same window the old flow did. (Edit is in settings.pl, which is gitignored — flagging per protocol.)</p>
<p>One honest caveat: those vim windows inherit the iTerm2 raise-burst behavior, so if macOS defers the raise mid-typing you may occasionally still owe an alt-tab for the *editor* window specifically — though since it only appears right after you&#x27;ve pressed Enter in the panel, you&#x27;re at the keyboard and not mid-flow elsewhere, which is the burst&#x27;s best case.</p></div>
</details>
</article>
<article class="exchange claude" id="p47">
<div class="diffstat">+29 −1 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span></span></div>
<pre class="prompt">seems to be working. what about the double-quote (ditto mark) shortcut? i haven't tested that yet. also i don't think it should let you type in the input box while the ping is being submitted</pre>
<details>
<summary><a class="anchor" href="#p47"><time datetime="2026-07-23T21:52:50.657000+00:00">14:52</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 2m10s · 2m40s wall-clock time</span></summary>
<div class="reply machine"><p>Two things, both handled:</p>
<p><strong>Ditto</strong>: the <code>&quot;</code> passes through the pipe verbatim to ping.pl&#x27;s <code>/^&quot;\s*$/</code> check, so it works — *except* that macOS text fields can apply smart-quote substitution, which would turn <code>&quot;</code> into <code>“</code> and silently break both ditto and any comment text. Disabling all auto-substitutions on the field (that&#x27;s the load-bearing fix; without it ditto works only until the OS decides to be helpful).</p>
<p><strong>Input lock during submission</strong>: locking on your ask, but not on submit — on *acceptance*. Rationale: a rejected answer (your <code>enforcenonon</code> check) makes ping.pl silently wait for a retype, so locking at submit would deadlock that case. Instead the panel locks (grays out, per the anti-magic gray-don&#x27;t-suppress principle) when ping.pl echoes the accepted answer — detectable as a line starting with the ping timestamp we ourselves passed in, so it&#x27;s matching our own data, not prose — and re-enables only for ping.pl&#x27;s &quot;press enter to dismiss&quot; error prompt (one greppable coupling to ping.pl&#x27;s own string):</p>
<p>Green. Live for the next ping:</p>
<ul><li><strong>Ditto</strong>: traced clean — your <code>&quot;</code> reaches ping.pl byte-for-byte, and the field editor now has quote/dash/text-replacement/spell substitution explicitly off, so macOS can&#x27;t curl it into <code>“</code> behind your back. Worth a real-ping test since I can&#x27;t type into the panel from here.</li><li><strong>Input lock</strong>: grays out the moment ping.pl echoes the accepted line, re-enables only for its &quot;press enter to dismiss&quot; error prompt. Deliberately *not* locked between submit and acceptance: a response failing your <code>enforcenonon</code> check makes ping.pl silently wait for a retype, and locking there would deadlock the panel.</li></ul>
<p>Related wart surfaced by that analysis, pre-existing and not panel-specific: ping.pl&#x27;s rejection is entirely silent (in the terminal flow too — the cursor just sits there). A one-line rejection message in ping.pl would make both flows honest; it&#x27;s your fifteen-year-old core script, so saying it rather than doing it.</p></div>
</details>
</article>
<article class="exchange claude" id="p48">
<div class="diffstat">+55 −3 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span></span></div>
<pre class="prompt">go ahead; i'll review what you do</pre>
<details>
<summary><a class="anchor" href="#p48"><time datetime="2026-07-23T22:23:37.414000+00:00">15:23</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m28s · 2m59s wall-clock time</span></summary>
<div class="reply machine"><p>QDD order: qual first, red, then the fix. The qual runs ping.pl against a fixture <code>$HOME</code> (scratch log, empty <code>%beeminder</code>, <code>enforcenonon=1</code>) so nothing touches your real log or Beeminder:</p>
<p>Green. For your review:</p>
<ul><li><strong>[script/qual-reject.sh](script/qual-reject.sh)</strong> (new): runs ping.pl under a fixture <code>$HOME</code> — scratch log, empty <code>%beeminder</code>, <code>enforcenonon</code> on — pipes it a <code>non</code>-tagged response then a clean one, and asserts: rejection notice printed, rejected response not logged, clean response logged. This fixture-HOME harness is a reusable pattern for future ping.pl quals. Red before the fix (silent vanishing confirmed), green after.</li><li><strong>[ping.pl](ping.pl#L136-L146)</strong>: the rejection condition is hoisted character-for-character into <code>$rejected</code> (with <code>$enforcenums</code>/<code>$enforcenonon</code> added to the existing <code>our</code> since the expression now lives inside the <code>use strict</code> block), and rejection prints a Latin notice with the TODO recap per your rules. <code>perl -c</code> passes.</li></ul>
<p>In the panel this means a rejected answer now produces visible transcript output with the field still enabled for the retype — the case that motivated not locking at submit.</p></div>
</details>
</article>
<h2 class="day" id="d2026-07-24"><time datetime="2026-07-24">2026-07-24 Friday</time></h2>
<article class="exchange claude" id="p49">
<pre class="prompt">can you double check all the recent changes? will this work for anyone on a mac? any permissions that need to be set on the new tagtime-panel?</pre>
<details>
<summary><a class="anchor" href="#p49"><time datetime="2026-07-24T22:29:10.248000+00:00">15:29</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 4m45s · 14m28s wall-clock time</span></summary>
<div class="reply machine"><p>I&#x27;ll review the uncommitted changes and the new files. Let me start by looking at the diff and the new scripts.</p>
<p>The git-guard hook only allows <code>git diff</code> and <code>git log</code>, so I&#x27;ll get the changes that way.</p>
<p>One last correctness check — the panel&#x27;s accepted-answer detection depends on ping.pl flushing output promptly to a pipe, so let me confirm ping.pl autoflushes.</p>
<p>Done — I reviewed the full working-tree diff, both new qual scripts, the Swift panel, and ran the quals. Here&#x27;s what I found.</p>
<p><strong>TLDR:</strong> The changes are sound — 9 of the 10 modified files are comment-only rewording, and the one behavioral change (ping.pl&#x27;s rejection loop) is logically equivalent to the old code plus a helpful error message. But <strong>qual-reject.sh is currently red</strong> because it greps for a message (&quot;reiectum&quot;) that ping.pl doesn&#x27;t print. And no, this won&#x27;t work for anyone on a Mac out of the box — the panel is completely undocumented and unwired in the repo itself. Good news on permissions: the panel needs <strong>none</strong>.</p>
<h3>The one red qual</h3>
<p>[qual-reject.sh:32](script/qual-reject.sh#L32) asserts the rejection notice matches <code>reiectum</code>, but [ping.pl:140](ping.pl#L140) prints &quot;Error! See enforcenums / enforcenonon in settings. Try again:&quot;. I ran it: that assertion fails; the two substantive assertions pass (the rejected response stays out of the log, the clean one gets logged). So the behavior is right — either the qual&#x27;s grep or ping.pl&#x27;s message needs to change to match the other. Relatedly, the comment at [tagtime-panel.swift:142-144](tagtime-panel.swift#L142-L144) claiming a rejected answer &quot;echoes nothing and silently awaits a retype&quot; is now stale — the panel&#x27;s logic is still correct (the error line doesn&#x27;t start with the ping timestamp, so the input field correctly stays enabled), just the comment&#x27;s premise changed.</p>
<p>Everything else checks out: the <code>$rejected</code> refactor preserves the original operator precedence exactly, the added <code>our ($enforcenums, $enforcenonon)</code> is needed and correct for the <code>use strict</code> block, and <strong>qual-panel.sh passes all green</strong> (I ran it — you may have seen a few red flashes). The compiled binary (Jul 23 14:55) is newer than the source (14:54), so it&#x27;s fresh.</p>
<h3>Will it work for anyone on a Mac?</h3>
<p>Not yet — for someone cloning the repo, the panel effectively doesn&#x27;t exist:</p>
<ol><li><strong>No wiring or docs in the repo.</strong> The only thing connecting the panel to TagTime is <code>$XT = &quot;${path}tagtime-panel&quot;</code> in your personal <code>~/.tagtimerc</code>. README.md still describes the X11/xterm era, and settings.pl.template doesn&#x27;t mention the panel. A new Mac user would need the build command and the <code>$XT</code> line documented somewhere.</li><li><strong>Build requires Xcode Command Line Tools</strong> (<code>swiftc</code> isn&#x27;t present on a stock Mac; they&#x27;d need <code>xcode-select --install</code> first). Since the binary is gitignored (good call in the .gitignore change), everyone builds locally.</li><li><strong>Don&#x27;t distribute your binary</strong> — it&#x27;s arm64-only (won&#x27;t run on Intel at all), and a copied/downloaded binary picks up the quarantine attribute, at which point Gatekeeper blocks its ad-hoc signature until right-click→Open. Built-from-source has neither problem.</li><li><strong>The <code>$EDIT_COMMAND</code> trap:</strong> [launch.pl:140](launch.pl#L140) falls back to <code>$XT $ED $f</code> for task editing. You&#x27;re safe because your .tagtimerc sets <code>$EDIT_COMMAND = &quot;${path}terminal-tagtime $ED&quot;</code>, but a new user who sets <code>$XT</code> to the panel without also setting <code>$EDIT_COMMAND</code> would get vim piped into the panel&#x27;s line-based input — thoroughly broken. Worth a warning next to whatever docs you add.</li><li><strong><code>tmp/</code> is gitignored</strong>, so on a fresh clone the timing-log instrumentation fails: the panel prints SYSERR to stderr every run (non-fatal — and I verified launch.pl&#x27;s equivalent also survives, since <code>printf</code> to a failed filehandle doesn&#x27;t die there), and qual-panel.sh&#x27;s <code>wc -l &lt; tmp/popup-timing.log</code> errors. This all evaporates when the &quot;remove when diagnosed&quot; instrumentation gets cleaned up, so probably not worth fixing separately.</li></ol>
<p>Minor good news: the panel&#x27;s accepted-answer detection assumes ping.pl&#x27;s output arrives promptly line-by-line, and it does — util.pl sets <code>$| = 1</code> and ping.pl requires it.</p>
<h3>Permissions on tagtime-panel</h3>
<p>None needed — this is the quiet win of the panel approach:</p>
<ul><li><strong>No TCC prompts at all</strong>: it doesn&#x27;t synthesize input (no Accessibility), doesn&#x27;t watch keys (no Input Monitoring), and doesn&#x27;t drive other apps via AppleEvents — the old osascript-based Terminal/iTerm popups needed Automation consent; this needs nothing.</li><li><strong>Signing is already handled</strong>: I checked with <code>codesign</code> — it&#x27;s ad-hoc linker-signed automatically (required and sufficient on Apple Silicon), no quarantine attribute, and swiftc sets the executable bit.</li><li>The only environmental requirement: tagtimed must run inside the GUI login session (a terminal window or a LaunchAgent). Launched over SSH, the panel couldn&#x27;t connect to the WindowServer.</li></ul>
<p>If you want, I can fix the qual-reject.sh/ping.pl message mismatch and the stale Swift comment — say the word and I&#x27;ll pick a side (my suggestion: make the qual grep for &quot;enforcenums&quot;, which survives message rewording).</p></div>
</details>
</article>
<article class="exchange claude" id="p50">
<div class="diffstat">+1 −1 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="del"></span><span class="nil"></span><span class="nil"></span><span class="nil"></span></span></div>
<pre class="prompt">i rewrote ping.pl's error; can you update the qual?

wait, how did that use to work, before we had $rejected? it would just reprompt you with no explanation or something?</pre>
<details>
<summary><a class="anchor" href="#p50"><time datetime="2026-07-24T23:00:39.717000+00:00">16:00</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m22s · 1m36s wall-clock time</span></summary>
<div class="reply machine"><p>Let me look at your rewritten error message, then update the qual to match.</p>
<p>Done — the qual is updated and green. I changed the grep from <code>reiectum</code> to <code>enforce</code>, which matches your current message (&quot;Error! See enforcenums / enforcenonon in settings. Try again:&quot;) and will keep matching through most rewordings, since any version of the notice presumably names the enforcement settings. All three assertions in [qual-reject.sh](script/qual-reject.sh) now pass.</p>
<p>As for how it used to work: even worse than reprompting with no explanation — it reprompted with *nothing at all*. The old code had the enforcement check directly in the <code>do...while</code> condition, so a rejected answer just looped straight back to <code>$resp = &lt;STDIN&gt;</code>. The &quot;What are you doing RIGHT NOW?&quot; prompt prints once, *before* the loop, so on rejection you got no error, no repeated prompt, not even a blank line — just a bare cursor silently waiting for you to type again. The only clue anything happened was the absence of the echoed log line, and if you didn&#x27;t know about <code>enforcenums</code>/<code>enforcenonon</code> you&#x27;d probably conclude ping.pl had hung. It&#x27;d been that way essentially forever: <code>enforcenums</code> predates the git history (it came over in the 2011 svn import) and <code>enforcenonon</code> was added in 2013. Your <code>$rejected</code> refactor is the first time in ~15 years that a rejected answer actually tells you why.</p>
<p>One loose end from my earlier review: the comment at [tagtime-panel.swift:142-144](tagtime-panel.swift#L142-L144) still says a rejected answer &quot;echoes nothing and silently awaits a retype&quot; — no longer true, though the code&#x27;s behavior is still correct (the error line doesn&#x27;t start with the ping timestamp, so the input field correctly stays unlocked). Happy to reword that comment if you want.</p></div>
</details>
</article>
<article class="exchange claude" id="p51">
<div class="diffstat">+24 −9 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span><span class="nil"></span></span></div>
<pre class="prompt">can you make the tagtime-panel less ugly? like no scrollbar, and the place to type more obvious.</pre>
<details>
<summary><a class="anchor" href="#p51"><time datetime="2026-07-24T23:11:52.091000+00:00">16:11</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 3m37s · 4m1s wall-clock time</span></summary>
<div class="reply machine"><p>I&#x27;ll restyle the panel: drop the scrollbar, make the input field bigger and unmistakably white-on-red with padding, and let the red bleed into the title bar so it looks intentional rather than like a stock window. Then rebuild, re-run the qual, and launch it to eyeball the result.</p>
<p>Build is clean and quals pass. Now let me actually look at it — loading the run skill to launch the panel and screenshot it.</p>
<p>No project skill, so I&#x27;ll drive it directly: fake ping script → launch panel → screenshot → kill.</p></div>
</details>
</article>
<article class="exchange claude" id="p52">
<pre class="prompt">apparently i need to restart vscode here before you'll be able to do those screencaptures. want to give me a command to run and i'll pass you a screenshot?</pre>
<details>
<summary><a class="anchor" href="#p52"><time datetime="2026-07-24T23:17:42.417000+00:00">16:17</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 15s</span></summary>
<div class="reply machine"><p>Sure — run this and the restyled panel will pop up with a fake ping transcript (the fake-ping script is already set up in my scratchpad):</p>
<pre class="code"><code data-language="bash">cd ~/lab/tagtime &amp;&amp; ./tagtime-panel /private/tmp/claude-502/-Users-dreeves-lab-tagtime/fca823f8-a6e2-410b-83f1-b6528205bad5/scratchpad/fakeping.sh 1784000001</code></pre>
<p>It&#x27;ll just sit there waiting for input like a real ping (typing into it is harmless — nothing gets logged); close the window when you&#x27;ve got the screenshot and paste it here. What you should be seeing: red everywhere including the title bar, no scrollbar, and a single white input strip along the bottom with the placeholder in 16pt.</p></div>
</details>
</article>
<article class="exchange claude" id="p53">
<pre class="prompt">here it is</pre>
<div class="attachments"><img class="attachment" src="data:image/png;base64,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" alt=""></div>
<details>
<summary><a class="anchor" href="#p53"><time datetime="2026-07-24T23:19:51.371000+00:00">16:19</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 22s</span></summary>
<div class="reply machine"><p>That&#x27;s exactly what I was going for — everything landed: red runs edge-to-edge including the title bar, the scrollbar is gone, and the input is now a single rounded white strip with the 16pt placeholder and a focus ring around it. Hard to miss where to type.</p>
<p>The deployed binary is already the new one (I rebuilt with <code>swiftc -O</code> and qual-panel passed before you launched it), and since launch.pl spawns <code>tagtime-panel</code> fresh for each ping, your next real ping gets the new look with no restart of tagtimed needed.</p>
<p>One small thing to sanity-check on your next real ping: the focus ring in your screenshot means the field had keyboard focus immediately — the grab-retry logic still works with the new styling, but that was launched from your terminal (already-active app), which is the easy case. The real test is a ping arriving while you&#x27;re in another app.</p></div>
</details>
</article>
<article class="exchange claude" id="p54">
<div class="diffstat">+6 −1 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span></span></div>
<pre class="prompt">it's pretty weird how... actually i'm not sure. can you just stare at it and make sure it's really what you think looks best?</pre>
<details>
<summary><a class="anchor" href="#p54"><time datetime="2026-07-24T23:24:30.707000+00:00">16:24</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 1m46s · 2m10s wall-clock time</span></summary>
<div class="reply machine"><p>Fair challenge. Staring at it properly, two things bug me — and one of them is probably the weirdness you couldn&#x27;t name:</p>
<ol><li><strong>The dead red void.</strong> The transcript is two lines at the top, the input is at the very bottom, and between them sits ~250px of empty red. Your eye reads the prompt up top, then has to travel across nothing to find where to type. A terminal always has blank rows below the cursor, so some emptiness is idiomatic — but at 700×420 with three lines of content, the proportions read as &quot;broken,&quot; not &quot;spacious.&quot; The window predates the restyle at that size; with the input now visually anchored at the bottom, the height should shrink to match the typical transcript.</li></ol>
<ol><li><strong>The zombie traffic lights.</strong> Minimize and zoom are disabled (the window is only <code>.titled, .closable</code>), so they render as two dead reddish-gray dots next to the one functional close button. Muddy. They should be hidden entirely.</li></ol>
<p>I considered bottom-anchoring the transcript chat-style so output hugs the input — that&#x27;s arguably the &quot;correct&quot; fix for the gap — but it needs an NSTextView subclass with fiddly layout invalidation, and the terminal idiom (output top-down, input below) is what ping.pl is. Shrinking the window gets most of the benefit for none of the risk. Long output (task lists) still auto-pins to the end, so nothing essential scrolls away. Making both fixes now:</p>
<p>Rebuilt and quals are green. So my honest verdict on the previous version: the input treatment, missing scrollbar, and red title bar were right, but it wasn&#x27;t yet what I&#x27;d call best — the 420px height was inherited from the old design and made no sense with three lines of content, and the disabled traffic-light dots were clutter. Both fixed: the window is now 700×300 (still an unmissable red slab, but proportionate to a typical transcript, and long task lists still auto-scroll to the newest line), and the title bar shows just the close button.</p>
<p>Same command if you want to eyeball it again:</p>
<pre class="code"><code data-language="bash">cd ~/lab/tagtime &amp;&amp; ./tagtime-panel /private/tmp/claude-502/-Users-dreeves-lab-tagtime/fca823f8-a6e2-410b-83f1-b6528205bad5/scratchpad/fakeping.sh 1784000001</code></pre>
<p>The one thing I deliberately left alone: the tomato-ish shade of red. It renders a bit orange on a P3 display, but it&#x27;s the classic TagTime red and changing it felt like scope creep — say the word if you want it deeper.</p></div>
</details>
</article>
<article class="exchange claude" id="p55">
<div class="diffstat">+11 −4 <span class="blocks" aria-hidden="true"><span class="add"></span><span class="add"></span><span class="add"></span><span class="del"></span><span class="nil"></span></span></div>
<pre class="prompt">is there a font inconsistency? and can the cursor be more noticeable? and maybe we don't need placeholder text since that same text is on the first line? and will this scroll properly if there's a lot of lines of stuff for sending to beeminder? and what about putting the entry field after the prompt instead of the bottom of the window? (i'm not sure about that last one; do what makes sense)</pre>
<div class="attachments"><img class="attachment" src="data:image/png;base64,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" alt=""></div>
<details>
<summary><a class="anchor" href="#p55"><time datetime="2026-07-24T23:31:22.095000+00:00">16:31</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 2m34s · 30m6s wall-clock time</span></summary>
<div class="reply machine"><p>All five are good catches — here&#x27;s my take on each: yes, there&#x27;s a real font inconsistency (transcript is 13pt, input was 16pt — and worse than the placeholder mismatch, your typed text would visibly shrink when it echoed into the transcript); yes, the cursor can pop (a TagTime-red caret in the white field instead of the hairline black one); agreed on dropping the placeholder — the prompt line already says it, and it was the main source of the size clash; scrolling should pin to the newest line but I&#x27;ll harden it and prove it; and on the entry field placement — I&#x27;d keep it at the bottom, and I&#x27;ll explain why after the edits.</p>
<p>Rebuilt, quals green. Rundown of your five points:</p>
<ol><li><strong>Font inconsistency — yes, real.</strong> The input was 16pt while the transcript is 13pt, so your typed line would visibly shrink when it echoed into the transcript. Everything is now the same 13pt monospaced font. (The transcript can&#x27;t go bigger than ~13pt anyway: ping.pl&#x27;s log lines are padded to 80 columns and would wrap in a 700px window at larger sizes.)</li></ol>
<ol><li><strong>Cursor — now TagTime-red.</strong> The caret in the input field is the same red as the window instead of the default hairline black. That&#x27;s the one well-supported lever for caret visibility in AppKit (width hacks rely on drawing overrides that macOS 26&#x27;s text system doesn&#x27;t reliably call).</li></ol>
<ol><li><strong>Placeholder — gone</strong>, for exactly the reason you said. With it gone, the field&#x27;s &quot;type here&quot; signal is the white strip, the focus ring, and the red caret, which the screenshot showed is plenty.</li></ol>
<ol><li><strong>Long output — yes, it scrolls pinned to the newest line.</strong> Every chunk of child output triggers a scroll-to-end, and trackpad scrolling still works for looking back. I also added the canonical NSTextView grow-with-content settings, which the previous code left to init defaults. One behavior to know about (unchanged from the old iTerm popups): when ping.pl exits after the Beeminder update, the window closes with it, so a long Beeminder flood is only readable while it streams. If you want the window to linger after Beeminder output, that&#x27;s a small ping.pl or panel change — say so.</li></ol>
<ol><li><strong>Input field after the prompt — I say no, and here&#x27;s the reasoning.</strong> A field that sits under the last output line moves: down as the task list prints, again when retro pings or errors append. For a reflex you perform dozens of times a day, a fixed target beats proximity — it&#x27;s why every chat app and REPL-with-a-real-UI puts input in a fixed bar at the bottom rather than at the text cursor. The &quot;far away&quot; problem was really the 420px window, which the 300px height already fixed: the gap between prompt and field is now a few lines, not a void.</li></ol>
<p>To see it (and test #4), run the same command — type anything, watch the flood pin to &quot;line 30 of 30&quot;, field should gray out, then close the window:</p>
<pre class="code"><code data-language="bash">cd ~/lab/tagtime &amp;&amp; ./tagtime-panel /private/tmp/claude-502/-Users-dreeves-lab-tagtime/fca823f8-a6e2-410b-83f1-b6528205bad5/scratchpad/fakeping.sh 1784000001</code></pre></div>
</details>
</article>
<article class="exchange claude" id="p56">
<pre class="prompt">oh, we didn't lose the active tasks stuff, did we?</pre>
<details>
<summary><a class="anchor" href="#p56"><time datetime="2026-07-25T00:05:42.826000+00:00">17:05</time></a> <span class="chip"></span> <span class="agent">Claude Code</span> <span class="model">claude-fable-5</span> <span class="effort">(xhigh)</span> <span class="elapsed">thought for 47s · 57m43s wall-clock time</span></summary>
<div class="reply machine"><p>No — nothing touched that. The task list comes from ping.pl (it walks your <code>.tsk</code> file and prints active tasks before the prompt), and the panel streams *everything* ping.pl prints into the transcript. Our changes were pure presentation: fonts, colors, window size, scrollbar, placeholder. Let me just confirm the task display would actually fit nicely in the shorter window:</p>
<p>Confirmed: nothing lost. The active-tasks display lives entirely in ping.pl ([ping.pl:58-73](ping.pl#L58-L73) — it prints everything up to the <code>----</code> divider line), and the panel just streams whatever ping.pl prints into the transcript. None of the restyling touched that path — same stream handling as before, just different fonts and window geometry. The task-number→tags expansion (typing <code>1</code> to log that task&#x27;s tags) also still works, since that&#x27;s all ping.pl-side too.</p>