Package Details: vcmi 1.5.2-1

Git Clone URL: https://aur.archlinux.org/vcmi.git (read-only, click to copy)
Package Base: vcmi
Description: Open-source engine for Heroes of Might and Magic III
Upstream URL: http://vcmi.eu
Keywords: homm3 vcmi
Licenses: GPL-2.0-or-later AND CC-BY-SA-4.0
Conflicts: vcmi
Provides: vcmi
Submitter: Raziel23
Maintainer: Gigas002
Last Packager: Gigas002
Votes: 47
Popularity: 0.194501
First Submitted: 2013-03-12 14:11 (UTC)
Last Updated: 2024-05-31 11:58 (UTC)

Pinned Comments

Gigas002 commented on 2024-05-29 15:08 (UTC)

PKGBUILD tweaks

If you want to build with clang, add this part to cmake flags:

    -DCMAKE_CXX_COMPILER='clang++' \
    -DCMAKE_C_COMPILER='clang' \

If you want to use bundled fuzzylite:

    -DFORCE_BUNDLED_FL=ON \

If you want to use innoextract (system-wide installation won't work since vcmi has a temporary? fork of innoextract for now):

    -DENABLE_INNOEXTRACT=ON \

Gigas002 commented on 2024-05-13 03:20 (UTC) (edited on 2024-05-13 13:04 (UTC) by Gigas002)

I think I was able to reproduce the issue in a clean docker container after some experiments. It seems to be caused by qt6 compilation for some reason

qt5-tools builds translation files into ${srcdir}/${pkgname}-${pkgver}/build/launcher/translation/*.qm

But qt6-tools builds translation files into ${srcdir}/${pkgname}-${pkgver}/build/launcher/*.qm which causes make install to fail

I don't know what enforces qt6 build on your machine, but I'm pretty sure it is the reason

I'll report this issue upstream

UPD

You can track this issue up here: https://github.com/vcmi/vcmi/issues/3522. It's said to be merged around the end of may. I think I'll change the dependencies to qt6 with the next release.

Latest Comments

« First ‹ Previous 1 2 3 4 5 6 7 Next › Last »

Raziel23 commented on 2014-11-02 10:47 (UTC)

[Bump to version 0.97-1]. New upstream release. The changelog is as follows: http://forum.vcmi.eu/viewtopic.php?p=13381#13381 I switched from SDL 1 to SDL 2, added minizip dependency and updated the installation instructions.

Raziel23 commented on 2014-10-11 06:17 (UTC)

@salviati: Why you flag this package out of date? There isn't newer stable version of vcmi than 0.96. Version 0.96b is a development version (not stable one). If you want use it please use vcmi-git package.

Raziel23 commented on 2014-08-24 14:22 (UTC)

[Update without bump the pkgver]. Set MAKEFLAGS="-j1" in the PKGBUILD in order to avoid running out of memory during compilation, because some people have problems with exceeding their memory during compilation with more than one job. Better safe than sorry. This change doesn't affect the package (only compilation) so I didn't bump the pkgver to avoid recompilation for other people.

Raziel23 commented on 2014-08-24 13:51 (UTC)

@bladedglove: Good catch. Thank you for the tip. With my 2 cores (2 threads) memory usage doesn't exceed my 8 GB. I think that instead of changing /etc/makepkg.conf I will set the variable MAKEFLAGS="-j1" in the PKGBUILD to avoid such situations. The compilation time will take longer (for me with -j1 it takes about 40 minutes), but at least more people will be able build the vcmi successfully.

bladedglove commented on 2014-08-24 08:15 (UTC)

To avoid running out of memory and freezing your system on multicore computers, set a lower value for MAKEFLAGS="-jX" in /etc/makepkg.conf before running makepkg -s. Compiling with "-j4" (forcing the compiler to use all 4 cores of my i5 CPU) resulted in running out of memory (8GB) and failing at 56%, but the "-j1" value forced the compiler to use only 1 core and the top memory usage was at around ~3-4GB. Of course, using a huge swap partition/file is also a solution, but that's another story.

Raziel23 commented on 2014-08-02 12:56 (UTC)

@KlavKalashj: I've experienced freezing the system during compilation vcmi version 0.95. Even with the frozen system I was able successfully compile vcmi 0.95. For me with vcmi version 0.96 and with vcmi-git this issue not exist. I can only speak for myself, because the problem doesn't exist for me. I presume that this issue may that exist during compilation registerTypes. This part of compilation takes most memory and time compilation. I'm not vcmi developer, I'm only the package maintainer and I don't know how solve this problem. You can write about this issue in the vcmi forum. Here it is the related topic about this problem: http://forum.vcmi.eu/viewtopic.php?p=11612

« First ‹ Previous 1 2 3 4 5 6 7 Next › Last »