Package Details: openbrf-git 2024.11.03-1

Git Clone URL: https://aur.archlinux.org/openbrf-git.git (read-only, click to copy)
Package Base: openbrf-git
Description: Mount&Blade resource editor by Marco Tarini.
Upstream URL: https://forums.taleworlds.com/index.php?topic=72279.0
Licenses: GPL
Conflicts: openbrf
Replaces: openbrf
Submitter: swyter
Maintainer: swyter
Last Packager: swyter
Votes: 3
Popularity: 0.33
First Submitted: 2024-11-06 17:30 (UTC)
Last Updated: 2024-11-06 17:30 (UTC)

Latest Comments

swyter commented on 2024-12-22 05:37 (UTC)

@hanker These warnings should be fixed after the latest few changes, give it another go when you can and let me know if that helps. :)

swyter commented on 2024-12-18 03:21 (UTC) (edited on 2024-12-18 03:27 (UTC) by swyter)

@hanker Hmm, looks like your default compiler options are a bit more picky, treating warnings as errors, I rebuilt it a couple of weeks ago and it did build fine for me. It may be a GCC VS Clang thing. While it's true that an user-controlled format string can be used for crashing the program or exploits, in my opinion this is a bit overzealous for something like OpenBRF, but it may be the new defaults for Arch, Qt6, GCC or whatever part that is overriding this. You can try appending -Wno-error=format-security to the compiler arguments from the PKGBUILD, or something.

https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html#index-Werror_003d

I'll eventually get to fix this in the actual source code.

hanker commented on 2024-12-16 10:04 (UTC)

There are build errors regarding sprintf:

An example:

brfData.cpp: In member function ‘const char* BrfData::GetAllObjectNamesAsSceneProp(int*, int*) const’:
brfData.cpp:72:42: error: format not a string literal and no format arguments [-Werror=format-security]
   72 |                 sprintf(lastName,mesh[i].baseName);

swyter commented on 2017-02-02 17:37 (UTC) (edited on 2017-02-02 17:38 (UTC) by swyter)

I don't have my Arch install with me right now. I will add a ternary max(nproc, 1) when I can. But you can replace the make -j $[`nproc` - 1] with a: > make -j $[ (n = `nproc` - 1) < 1 ? 1 : n ] ...or, if you want to be fancy... > make -j $[ (`nproc` - 1) | 1 ] ...and it should work fine. Have fun modding M&B!

Caesim404 commented on 2017-02-02 17:08 (UTC)

Just tested and make does fail with -j 0

swyter commented on 2017-02-02 16:55 (UTC)

@Caesim404 Probably nothing if 'make' sanitizes input as it should. It benefits enormously from the parallelization, so it's worth putting it as default for the majority of people that won't bother tweaking their local config. You can always edit it to your liking before running it. At least I haven't hardcoded any value.

Caesim404 commented on 2017-02-02 16:13 (UTC) (edited on 2017-02-02 16:16 (UTC) by Caesim404)

I suggest not setting the job count for make in the PKGBUILD. One usually sets it in their /etc/makepkg.conf. Also what would happen with only 1 core?

swyter commented on 2016-12-03 18:12 (UTC) (edited on 2016-12-03 18:13 (UTC) by swyter)

VCG lib has migrated from Sourceforge/SVN to GitHub/Git. I have just fixed the sources, now it builds. I also improved the .pro file replacements and used the actual number of CPUs minus one at compilation time to make it fast while staying responsive. Let me know if it builds for you. Post a comment when the build breaks.