feat: v1.1.0 #10
feat: v1.1.0 #10
Conversation
remove dependency on pylspclient
See https://iterm2.com/feature-reporting/Hyperlinks_in_Terminal_Emulators.html for information about terminal hyperlinks. This is supported by multiple terminals and makes it easier to find out more information about a failing check. This is the equivalent of PR #9 rebased on branch 1.1 (closes #9) --------- Co-authored-by: Lingjie Li <[email protected]>
Resolve #14 --------- Co-authored-by: Lingjie Li <[email protected]>
…12) * Adds a --line-filter argument support similar to clang tidy. This allows filtering of diagnostics on per file and per line range basis * Add clangd-tidy-diff a CLI command similar to clang-tidy-diff, which runs clangd-tidy on all files changed in a unified diff and filters the diagnostics to only contain those from changed lines. This allows easier onboarding for large existing codebases. Inspired by clang-tidy-diff script from the llvm project (https://clang.llvm.org/extra/doxygen/clang-tidy-diff_8py_source.html). This allows adding a CI check of only the changes lines This PR is the equivalent of PR #8 rebased on branch 1.1 closes #8 --------- Co-authored-by: Lingjie Li <[email protected]>
clangd_tidy/line_filter.py
Outdated
| def allows(self, start: int, end: int) -> bool: | ||
| return any( | ||
| self._interval_intersect(line_range, LineRange(start, end)) | ||
| for line_range in self.line_ranges | ||
| ) |
There was a problem hiding this comment.
I think you simplified my code really nicely in a few places such as the introduction of the LineRange class, but my original intention in this function was to be as backwards compatible with the original clang-tidy format
It is my understanding the logic is a bit more complex, such as it supports partial file paths and mentioning files without a specific file range
There was a problem hiding this comment.
Oh, I didn't realize this was an original clang-tidy feature. I'll work on aligning them.
There was a problem hiding this comment.
Hey @jmpfar I have fixed the compatiblity issue. Would you be willing to help with a final test?
There was a problem hiding this comment.
Sure, tell me what you want me to test
One thing that might be an issue:
clangd-tidy/clangd_tidy/line_filter.py
Lines 93 to 99 in 5e8ce45
I guess you had to do that because of the new clang format diagnostics (?) but this causes issues with the unused-includes/missing-includes diagnostics.
I'd like to have these filtered out as well, because the current situation makes these show up even when they're not part of the diff.
In my very quick tests, I also seen this behavior happen with the ovl_no_viable_oper diagnostic, which appears to be a non-clang-tidy analytic, though I am not sure where it is coming from (clangd?)
There was a problem hiding this comment.
though I am not sure where it is coming from (clangd?)
Adding -c flag will show the source.
unused-includes/missing-includesdiagnostics.
They come from clangd source.
ovl_no_viable_operdiagnostic
I didn't filter out clang diagnostics to align with clang-tidy, but after reconsidering, I now think it's better to keep only clang-format diagnostics unfiltered.
Please check my latest update.
There was a problem hiding this comment.
Sure, tell me what you want me to test
Thank you! Just check if it works in your cases.
For example, I recently discovered a bug with the exit code caused by invisible diagnostics (real error filtered out by line filter, and support info filtered out by formatter).
There was a problem hiding this comment.
I did a few more simple tests with my workflow and it looks like it's working well, thanks for your work!
lljbash
force-pushed
the
v1.1.x
branch
2 times, most recently
from
b76ef5f to
5e8ce45
Compare
|
closes #17 |
Changelog
Features
--query-driver(Add support for--query-driver#14).--line-filtersupport andclangd-tidy-diff(feat: add ability to run clangd-tidy over source control diffs [1.1] #12).Refactor
clangdintegration with an async LSP client, removingpylspclient.Misc