Bug 233423
| Summary: | devel/tig: Use textproc/gsed for configure | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | Ports & Packages | Reporter: | Kyle Evans <kevans> | ||||||
| Component: | Individual Port(s) | Assignee: | Tobias Kortkamp <tobik> | ||||||
| Status: | Closed FIXED | ||||||||
| Severity: | Affects Only Me | CC: | johans, tobik | ||||||
| Priority: | --- | ||||||||
| Version: | Latest | ||||||||
| Hardware: | Any | ||||||||
| OS: | Any | ||||||||
| Bug Depends on: | |||||||||
| Bug Blocks: | 229925 | ||||||||
| Attachments: |
|
||||||||
Created attachment 199473 [details] tig.diff I rebuilt libc with the patch from bug #229925 and was able to reproduce the problem. However I only see one \s sed invocation in tools/make-builtin-config.sh which seems trivial to patch and tig builds fine afterwards. Am I making this too easy? (In reply to Tobias Kortkamp from comment #1) Nope, looks good to me. =) With these patches, I have a tendency to err towards the side of caution and switch to gsed for two reasons: 1. Hunting down the GNU-isms can be costly, time-wise, and I have a bunch of these to sort through 2. Ultimately, I'm not the maintainer of these ports, so any patches I add are a potential maintenance burden on someone else and I know gsed will do what they are intending If maintainer wants to propose a counter-patch such as this because it's a trivial fix, that is also fine. I plan to do a pass after my work is done on anything depending on gsed and gnugrep from ports to switch back to the base versions once they're suitable. A commit references this bug: Author: tobik Date: Fri Nov 23 13:59:08 UTC 2018 New revision: 485659 URL: https://svnweb.freebsd.org/changeset/ports/485659 Log: devel/tig: Fix a GNUism in tools/make-builtin-config.sh PR: 233423 Reported by: kevans Changes: head/devel/tig/files/ head/devel/tig/files/patch-tools_make-builtin-config.sh Committed. Thanks! |
Created attachment 199467 [details] svn(1) diff against devel/tig Hi, The config script here requires some GNU-isms (\s) that regex(3) in base does not currently support. This is not a problem up until the near future: \s is currently interpreted as an escaped ordinary character, which will be forbidden in the future. Move it over to using gsed, which will also do the correct thing with the GNU-isms. BINARY_ALIAS was also used because some unsalted sed invocations are used, rather than the results discovered in the config script. Thanks, Kyle Evans