[build failure] package dmraid #119
Labels
No Label
blocked upstream
bug
build-failure
duplicate
enhancement
help wanted
informational
invalid
invalid/corrupt package
packaging issue
priority: high
question
support
wontfix
No Milestone
No project
No Assignees
2 Participants
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: ALHP/ALHP.GO#119
Loading…
Reference in New Issue
No description provided.
Delete Branch "%!s(<nil>)"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
The dmraid package fails because of the flag "-Werror=format-security". While it is set by default, it doesn't build with it set.
Log part of the issue:
Is this a thing we can work around or do we ask the maintainer/the project to fix it?
Also the GCC calls look extremly messy with -O2 set 4 times while we set -O3 one time. Not sure we're building really with -O3 in the end.
So this is kind of a general request for guidance. How can I help to improve this? Do we need to report this to the project? Is it normal to just "work around" it and remove the flag for this one package? I mean I don't think it can be build atm without removing the flag.
Some projects force a certain level of optimization, which we should not touch. About
-Werror=format-security
: That should be handled by upstream Arch. If we begin to modify PKGBUILDS for something like that we'll never see the end.I feel like that might just be historic, like forcing a higher optimizasion but GCC instroduced a more stable -O3 so it could in theory be build with a higher optimization now.
If nobody tries, nobody will know that for sure :D
No, I didn't want to ask if we should introduce a patchset locally. It's more a: What did I miss?
With a standard arch installation, shouldn't then all PKGBUILDs of Arch "just build" without tinkering around?
So yes, this seems to be handled by the maintainer, but isn't documented in the PKGBUILD of Arch.
And the question was: Should we create bugreports for the maintainers to fix this documentation issue - or go to the project itself and fix the cause of this issue (or at least ask for a fix)?
Certainly, but I see no way to do this besides patching makefiles (or whatever the respective project is using). First we had to detect the buildsystem in use, then we would have to (in a brute-force variant) replace all
-On
? Well, certainly not very elegant. If you can came up with something more sophisticated I'm all ear. Even detecting the buildsystem would take some work, on the other hand this could play into #124.In an ideal world, yes. Some PKGBUILDs do not get touched for some time, and things change, new compiler versions come in, etc.
The correct way would be to use the bugtracker, at least if it's fixable by the Arch maintainer. If this needs to be fixed in the upstream project itself, then the upstream bugtracker it is. You can certainly open a bug in both if you want (linking to upstream issue).
Yeah the shim would see all flags beeing send to the compiler and could modify all of them. So we could filter out all "-O2"s and set a single -O3 for example. There's no need to know or understand the build system, as it would just modify the calls at the end of the line going to the compiler.
Here's an example how this looks like (the code is from the yaourt project). :)
https://github.com/RubenKelevra/pacman.store/blob/master/pacman%20wrapper/for-bashrc.sh
Ah okay. I thought they would get a new pkgrel if one of the dependencies/build tools is upgraded.
Okay, good to know :)
I think this can be closed. The original request is answered (as in: we can't do anything about it), and the shim is tracked in #124.