[Bug build/34241] New: [gdb/build] Codespell word regexp refinement

https://sourceware.org/bugzilla/show_bug.cgi?id=34241

            Bug ID: 34241
           Summary: [gdb/build] Codespell word regexp refinement
           Product: gdb
           Version: HEAD
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: build
          Assignee: unassigned at sourceware dot org
          Reporter: vries at gcc dot gnu.org
  Target Milestone: ---

I noticed a commit fixing a typo gcs_availabe -> gcs_available.

I wondered whether this was detected by codespell, and found that while this is
detected:
...
$ echo availabe | codespell -
1: availabe
        availabe ==> available
...
this isn't:
...
$ echo gcs_availabe | codespell -
$ 
...

This has to do with the --regex REGEX option, which defaults to:
...
any alphanumeric character, the underscore, the hyphen, and the apostrophe
...
which AFAICT translates to --regex="[a-zA-Z0-9_\-']+".

If we drop the underscore in there, we do get:
...
$ echo gcs_availabe | codespell --regex="[a-zA-Z0-9\-']+" -
1: gcs_availabe
        availabe ==> available
...

Trying this out in pre-commit codespell:
...
diff --git a/gdb/pyproject.toml b/gdb/pyproject.toml
index 4109dfb8f82..85d3fefa889 100644
--- a/gdb/pyproject.toml
+++ b/gdb/pyproject.toml
@@ -68,3 +68,5 @@ uri-ignore-words-list = '*'

 # How to ignore blocks of code.
 ignore-multiline-regex = 'codespell:ignore-begin.*?codespell:ignore-end'
+
+regex = "[a-zA-Z0-9\\-']+"
...
we get a fair amount of new hits:
...
$ pre-commit run codespell --all-files | grep -c '==>'
144
...

A problem with this option is that it doesn't combine with -w:
...
  -r, --regex REGEX     ... . This option cannot be specified together 
                        with --write-changes.
...   
which is a convenient way to fix codespell errors.

There's an enhancement issue open about this ( 
https://github.com/codespell-project/codespell/issues/1508 ).

Regardless, we can already use this, just not in combination with -w.

I suppose if the codespell cleanup is complete, not being able to use -w won't
be such a big issue.

-- 
You are receiving this mail because:
You are on the CC list for the bug.