Add signed types - #368
Conversation
…uce code duplication
Add span overloads based on static extent
|
|
@chfast the PR is ready for review. The failed tests are false negatives since Clang warns of using "delete with a message" as a C++26 extension but we are already guarding against that with the INTX_DELETE macro. Let me know if you want any changes :) |
|
@chfast sorry for the ping, have you been able to take a look at the PR? |
chfast
left a comment
chfast
left a comment
There was a problem hiding this comment.
Hey, I only did a quite overview. This PR adds too many things at the same time. We need to split these into multiple PRs (e.g. future C++ standards, basic sint type: wrapper + reuse of the compatible two's complement operators, other signed operators).
I understand, I mainly added the reusable methods to address the automated quality control checks regarding code duplication. Just so I'm clear, do you want me to remove the newer C++ standard features as well as the reusable methods for now, and just focused the PR on making signed integers available, regardless of potential code duplication? |
|
|
@chfast I have removed the unrelated changes from this PR so they can be added to a different PR at another time. This PR now focuses solely on adding a signed integer type, "wrapping" over the base unsigned type operators when logical The failing tests seem unrelated to this PR, besides the code duplication warning (can be solved in a future PR by adding back some of the refactors) |
|
@chfast let me know what you think about it now, I removed the unrelated changes. The signed type behaves as a "wrapper" using the underlying unsigned type's operators where they make sense to. |
|
@chfast let me know if you want any further changes to this PR, I'm down to further refine this until you find it satisfactory 👍 |
2 participants
See #112
This PR aims to add signed variants for custom integers as a wrapper over the unsigned types, using two's complement