Skip to content

test: add new e2e tests - #1065

Open
vasylk-crowdin wants to merge 26 commits into
mainfrom
next_tests
Open

test: add new e2e tests#1065
vasylk-crowdin wants to merge 26 commits into
mainfrom
next_tests

Conversation

@vasylk-crowdin

Copy link
Copy Markdown
Collaborator

No description provided.

@codecov

codecov Bot commented Jul 28, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.18%. Comparing base (c0a4f40) to head (d3a4c55).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files
@@            Coverage Diff             @@
##             main    #1065      +/-   ##
==========================================
- Coverage   92.20%   92.18%   -0.01%     
==========================================
  Files         141      141              
  Lines       11485    11559      +74     
==========================================
+ Hits        10589    10655      +66     
- Misses        896      904       +8     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
  • 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

@andrii-bodnar andrii-bodnar changed the title Add new e2e tests test: add new e2e tests Jul 28, 2026
@bhdnb
bhdnb force-pushed the next_tests branch 2 times, most recently from 0372fca to 1633cad Compare July 30, 2026 11:52
@andrii-bodnar
andrii-bodnar changed the base branch from next to main August 26, 2026 13:49
@vasylk-crowdin
vasylk-crowdin marked this pull request as ready for review September 1, 2026 07:32
vasylk-crowdin and others added 21 commits September 1, 2026 19:42
b522ff7 reworked groupKey() so that content lines share one sortable
group (PLAIN) instead of each keying on itself. Bare `--output plain`
path listings therefore sort alphabetically now instead of keeping the
download-completion order, but two snapshots were not regenerated:

  - download sources ... with --output plain
  - download translations --all ... with plain output

Both diffs are a pure re-sort: same lines, same count, new order.

The same commit also landed multilingual-csv.test.ts.snap empty, dropping
all 13 of its snapshots; restore them, including the
`Directory <name> created` -> `Directory '<name>'` message change.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
… path

The suite was written 2026-07-24 and never revisited, so three assertions
outlived the behavior they described:

  - `requires --engine-id` and `warns when --auto-approve-option ...` both
    read result.stdout, but CliError and output.warning are diagnostics and
    d1dac57 (2026-08-20) routed those to stderr in every output format so
    stdout carries only the result document.

  - `uploads sources` expected `File 'sources/1_android.xml'`, the local
    path. Success lines report the project path, and the fixture sets
    `preserve_hierarchy: false`, which strips the shared `sources/` parent --
    exactly as the comment on the following test already spelled out.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
…e one

The `docxMatches.length === 1` check could not pass under any CLI
behavior: the same test also asserts stdout contains `File
'Folder/sample.docx'` and that `/sources/sample.docx` still exists and
still carries the sampleV2 SRX segmentation. Those two imply at least two
`sample.docx` files, so the count assertion contradicted them rather than
marking a regression.

Its stated justification was Java parity, but v4's
ProjectFilesUtils.fileLookup is logically identical to fileLookup here, so
a dest change duplicates on v4 too - the duplication is a pre-existing
product wart, not a port regression, and cleanup is `--delete-obsolete`.

Assert both project paths instead, which is what actually happens.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The fixture's sample.docx separates the sentences with U+00A0, not U+0020
-- Word writes a non-breaking space after an abbreviation. The expectation
was typed with plain spaces, so it failed against two strings that render
identically, which is what made the Bun diff look like it compared a value
to itself.

Written with explicit U+00A0 escapes so the difference is visible in the
source rather than only in a hex dump.

Only reachable now that the impossible `docxMatches.length === 1` check no
longer aborts the test before this line.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The command fails before producing a result document, so stdout is empty
and snapshotting it recorded only "". Point the snapshot at stderr, where
the diagnostic actually goes, so it captures the error output verbatim.

Also commits the suite's snapshot file, which had never been checked in.
The `uploads sources` entry is from a green run; the `--engine-id` entry is
left out deliberately so the next run records stderr rather than resurrecting
the empty stdout capture.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants