Re: [Rpm-maint] [rpm-software-management/rpm] debugedit: Distinguish files from directories in src list file (#1436)
@vt-alt pushed 1 commit. 161c89795fae9826ae63609a9a5f9f2360cb9ccd debugedit: Distinguish files from directories in src list file -- You are receiving this because you are subscribed to this thread. View it on GitHub: https://github.com/rpm-software-management/rpm/pull/1436/files/bd3ba770a0794b280576fc5c34eea9f95d7f18d1..161c89795fae9826ae63609a9a5f9f2360cb9ccd ___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] debugedit: Distinguish files from directories in src list file (#1436)
@vt-alt pushed 1 commit. bd3ba770a0794b280576fc5c34eea9f95d7f18d1 debugedit: Distinguish files from directories in src list file -- You are receiving this because you are subscribed to this thread. View it on GitHub: https://github.com/rpm-software-management/rpm/pull/1436/files/8ab1e3646607711c81860998bdf49cc2ffe81fdd..bd3ba770a0794b280576fc5c34eea9f95d7f18d1 ___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] debugedit: Distinguish files from directories in src list file (#1436)
> Might it be an idea to split find-debuginfo.sh upstream and adopt/merge your > scripts? I split `process-debuginfo` to run them via `xargs -P` in parallel to speed up debuginfo processing. In general, modular approach might be more maintainable than monolithic script. So, this might me a good idea. > BTW. I gave you bad advice, p[0] is after the last character written Ouch, I didn't notice that too. -- You are receiving this because you commented. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/1436#issuecomment-729049023___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
[Rpm-maint] [rpm-software-management/rpm] Get rid of ugly static check_fileList buffer (#1441)
We can easily generate the big filelist locally in checkList(). Its a little slower in theory but may in turn open new parallelisation opportunities, and its just a whole lot cleaner this way. No functional changes. You can view, comment on, or merge this pull request online at: https://github.com/rpm-software-management/rpm/pull/1441 -- Commit Summary -- * Get rid of ugly static check_fileList buffer -- File Changes -- M build/files.c (40) -- Patch Links -- https://github.com/rpm-software-management/rpm/pull/1441.patch https://github.com/rpm-software-management/rpm/pull/1441.diff -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/1441 ___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] RFE: Support creating new sub-packages after %install execution (#1222)
Changed title to more generic. As mentioned in the related PR, I think the suggested approach of parsing a single section of the spec fundamentally differently from others is neither the easiest or the best solution, both from user and implementation perspective. The mixed macros vs pseudo macros vs shell script mess of spec is difficult to follow as it is, we should be making things easier not harder. An external file that is parsed for inclusion after `%install` execution can be created from the spec or external scripts/tooling, be it `%spec -f ` or whatever in the spec syntax itself (something modeled after the familiar `%files -f foo`, see below). As that file is parsed with the same macro expansion as the spec itself, this doesn't prevent deep macro magic in any way. And as that file doesn't even exist initially there can be no confusion about when and where things are expanded etc. We can have a new scriptlet that *executes* after %install for creating that file, but it needs to be *parsed* along the rest of the spec. That could of course be equally achieved from end of `%install` but conceptual separation from that would probably be better (just like we should have separate slots for unpacking and configuring the sources) As for the inclusion syntax, we could add a kind of pseudo section named `%spec` which if present, marks the beginning of a spec, just like `%files` does for file section (optionally accompanied with `%end`). This might well serve some other parsing-related purposes as well, but for starters it gives us a place to add that `-f ` for the subspec to be parsed after `%install`. At the beginning of a spec it would place it quite far from the place where said file gets created so there might well be better alternatives (such as allowing multiple `%spec -f` constructs throughout, or something) this is just to outline a design that fits with the rest of rpm / spec. That sort of thing makes it possible to add new sub-packages on the fly from any old tooling, but additionally we could look at a native Lua API that could be used for adding sub-packages, running from a special Lua-specific slotf after `%install` - *not* talking about macros here. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/issues/1222#issuecomment-728882322___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] debugedit: Do not 'edit_dwarf2' when just extracting build-id (#1435)
On Sun, 2020-11-15 at 12:18 -0800, Vitalio wrote: > > Yes, but you also have to check base_dir && dest_dir aren't set. > > I will rework. You last patch looks good. Thanks, Mark ___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] debugedit: Distinguish files from directories in src list file (#1436)
Hi Vitalio, On Sun, 2020-11-15 at 12:35 -0800, Vitalio wrote: > BTW. How does this help you speed things up? > > We do just `grep -v '/$' to filter directories from source list[1], instead > of loop similar to `while read f; do test -d $f || echo $f; done` which takes > more than 15 seconds on kernel package. > > [1] > http://git.altlinux.org/people/vt/packages/?p=rpm-build.git;a=commitdiff;h=623707e72815 aha, you have a separate process-debuginfo and find-debuginfo-files script that I assume replaces the upstream find-debuginfo.sh script. Yours seem much simpler than what we have in find-debuginfo.sh, which has grown support for dwz, mini-symtab, etc. Might it be an idea to split find-debuginfo.sh upstream and adopt/merge your scripts? BTW. I gave you bad advice, p[0] is after the last character written, so will always be the zero string terminator because we now use strlen() - 1. It should have been *(p-1) != '/'. Which should always work because we filter out zero length strings in the if statement before the block, so there is always at least one (last) character. Sorry about that. Cheers, Mark ___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint