Arch Linux User Repository

Please amend the git describe to remove the --tags option, as I have replaced the r1050 tag with an identical tag, this time with an annotation, so it will be picked up by a default describe action. Sorry about that. I also inadvertently ended up pushing an (unannotated) tag of a release binary to the Github repository. That was confusing the --tags option, so I've removed that everywhere it was used. Gosh, I hope this doesn't terribly break everyone who keeps full copies of the repository around...

I gave a quick reply to kode54 on IRC but for posterity:

1. This refers to https://github.com/bnnm/vgmstream-g7221 which is covered by a relatively restrictive license from Polycom, Inc. and as such is not something I'm willing to provide.
2. This refers to https://github.com/kode54/libg719_decode for which the same story applies.
3. My bad, fixed.
4. Yeah, it's a fair point. I actually considered renaming this to vgmstream-git already previously, but I saw that the SourceForge one is still up and http://hcs64.com/vgmstream.html refers to this as "kode54's repo" so the kode54 qualifier still sort of makes sense to me. Honestly, I'd probably have renamed this already if it wasn't such a hassle. My mistake for not zealously going for the vgmstream-git name in the first place.

Excellent. Looks like it does everything it really needs to, except for some minor quibbles:

1. g7221_decode should be using the reference decoder, as libsiren has some decoding problems that render incorrect output on some files.
2. g719_decode would be nice to have, too.
3. Really minor, but if it used the --tags option when describing the repository state, it would produce a r1050-* revision string instead of r1020. r1050 was the last actual SVN revision to be converted by git-svn.
4. Which brings me to point out that my repository is actually now the officially maintained version of VGMStream, the SourceForge repository is no longer maintained, as far as I know.

Edit: Has been fixed in upstream.

Got a segfault after printing metadata (vgmstream-cli). Fixed by:

--- cli/vgmstream_cli.c
+++ cli/vgmstream_cli.c
@@ -434,7 +434,6 @@
     /* prints done */
     if (cfg.print_metaonly) {
         if (!cfg.play_sdtout)
-            fclose(outfile);
         close_vgmstream(vgmstream);
         return EXIT_SUCCESS;
     }

This fails to build

==> Making package: vgmstream-kode54-git r1020.r1169.gc3d3aaee-2 (Wed 21 Nov 2018 04:44:36 PM EST)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving sources...
  -> Cloning vgmstream-kode54-git git repo...
Cloning into bare repository '/home/user/AUR-shit/vgmstream-kode54-git/vgmstream-kode54-git'...
remote: Enumerating objects: 6, done.
remote: Counting objects: 100% (6/6), done.
remote: Compressing objects: 100% (6/6), done.
remote: Total 21860 (delta 1), reused 2 (delta 0), pack-reused 21854
Receiving objects: 100% (21860/21860), 20.19 MiB | 1.79 MiB/s, done.
Resolving deltas: 100% (18119/18119), done.
  -> Found install-headers.patch
  -> Found no-link-audacious.patch
==> Validating source files with sha256sums...
    vgmstream-kode54-git ... Skipped
    install-headers.patch ... Passed
    no-link-audacious.patch ... Passed
==> Extracting sources...
  -> Creating working copy of vgmstream-kode54-git git repo...
Cloning into 'vgmstream-kode54-git'...
done.
==> Starting prepare()...
patching file src/Makefile.autotools.am
patching file src/Makefile.autotools.am
Hunk #1 FAILED at 11.
1 out of 1 hunk FAILED -- saving rejects to file src/Makefile.autotools.am.rej
==> ERROR: A failure occurred in prepare().
    Aborting...

Getting an error compiling a while(true) loop. Fixed by doing:

diff --git a/src/meta/ea_eaac.c b/src/meta/ea_eaac.c
index d7ab6823..45673753 100644
--- a/src/meta/ea_eaac.c
+++ b/src/meta/ea_eaac.c
@@ -2,6 +2,8 @@
#include "../layout/layout.h"
#include "../coding/coding.h"
#include "ea_eaac_streamfile.h"
+#define true  1
+#define false 0

/* EAAudioCore formats, EA's current audio middleware */