RFC: Feature requests for 1.8.0 #220
Loading…
Reference in New Issue
There is no content yet.
Delete Branch "%!s(<nil>)"
Deleting a branch is permanent. Although the deleted branch may exist for a short time before cleaning up, in most cases it CANNOT be undone. Continue?
I think pkgconf is largely a "done" topic as @ddevault mentions.
So I would like to set up a 1.8.0 branch that is basically long-term maintenance until the project is no longer needed.
What kinds of things would people like to see in such a capstone branch?
I think one nice thing to have would be reworked manpages using scdoc. What do people think of that?
As the maintainer of scdoc, I will abstain from any subjective judgements on its utility to pkgconf.
Speaking only of the facts, scdoc is not a burden on the build process. It is written in standards-conformant C99 and targets a minimal subset of POSIX. It supports reproducible builds.
I'd like the library (and include) path mangling to be completed, not just filtering -L/foo/lib if /foo/lib is in LIBRARY_PATH, but also -Wl,-R/foo/lib if it is in LD_RUN_PATH. I'm in the process of hacking up a patch for that … just trying to figure out how to enable the tracing in the library help me find the spots …
Edit: Maybe I'll wait for feedback on the idea. I could also disable the source of those Wl,-R in pkg-config files from pkgsrc builds, since they are not that usual, anyway.
Seems like a useful feature, but not sure how to do it in a clean way.
You mean because of the variants (
-Wl -W-R/foo/lib
,-Wl,-rpath=/foo/lib
, etc.)? That's for sure inconvenient but I guess the possibilities can be enumerated. The rest seems identical to the handling of-L
…Internally, the -Wl,... flags are stored as
{'W', 'l,...'}
is the problem. So filtering-L
is easy, but rpath and stuff like that not so much. I do agree we should do it though.