Your message dated Fri, 16 Jan 2026 18:25:32 +0000
with message-id <[email protected]>
and subject line Bug#1113813: fixed in llama.cpp 7593+dfsg-3
has caused the Debian Bug report #1113813,
regarding llama.cpp: please make the build reproducible
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact [email protected]
immediately.)
--
1113813: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1113813
Debian Bug Tracking System
Contact [email protected] with problems
--- Begin Message ---
Source: llama.cpp
Version: 5882+dfsg-3
Severity: wishlist
Tags: patch
User: [email protected]
Usertags: cpu
X-Debbugs-Cc: [email protected]
Hi,
Whilst working on the Reproducible Builds effort [0], we noticed that
llama.cpp could not be built reproducibly.
This is because the llama-bench.1 manual page embeds the number of
CPUs:
│ │ │ ├── ./usr/share/man/man1/llama-bench.1.gz
│ │ │ │ ├── llama-bench.1
│ │ │ │ │ @@ -63,15 +63,15 @@
│ │ │ │ │ \fB\-dt\fR, \fB\-\-defrag\-thold\fR <f>
│ │ │ │ │ (default: \fB\-1\fR)
│ │ │ │ │ .TP
│ │ │ │ │ \fB\-t\fR, \fB\-\-threads\fR <n>
│ │ │ │ │ -(default: 16)
│ │ │ │ │ +(default: 12)
│ │ │ │ │ .TP
A patch is attached that replaces this nondetermistic value with a
placeholder value; the default is still calculated at runtime.
[0] https://reproducible-builds.org/
Regards,
--
,''`.
: :' : Chris Lamb
`. `'` [email protected] / chris-lamb.co.uk
`-
--- a/debian/patches/reproducible-builds.patch 1969-12-31 16:00:00.000000000
-0800
--- b/debian/patches/reproducible-builds.patch 2025-09-02 14:32:40.264552665
-0700
@@ -0,0 +1,15 @@
+Description: Make the build reproducible
+Author: Chris Lamb <[email protected]>
+Last-Update: 2025-09-02
+
+--- llama.cpp-5882+dfsg.orig/common/arg.cpp
++++ llama.cpp-5882+dfsg/common/arg.cpp
+@@ -1297,7 +1297,7 @@ common_params_context common_params_pars
+ ).set_examples({LLAMA_EXAMPLE_MAIN, LLAMA_EXAMPLE_SPECULATIVE,
LLAMA_EXAMPLE_LOOKUP}));
+ add_opt(common_arg(
+ {"-t", "--threads"}, "N",
+- string_format("number of threads to use during generation (default:
%d)", params.cpuparams.n_threads),
++ "number of threads to use during generation (default: <num_cpus>)",
+ [](common_params & params, int value) {
+ params.cpuparams.n_threads = value;
+ if (params.cpuparams.n_threads <= 0) {
--- a/debian/patches/series 2025-09-02 14:22:06.037488030 -0700
--- b/debian/patches/series 2025-09-02 14:32:38.842189218 -0700
@@ -3,3 +3,4 @@
cmake-Install-to-private-directories.patch
gguf-py-depends-on-the-requests-library.patch
Add-soversion-to-libraries.patch
+reproducible-builds.patch
--- End Message ---
--- Begin Message ---
Source: llama.cpp
Source-Version: 7593+dfsg-3
Done: Christian Kastner <[email protected]>
We believe that the bug you reported is fixed in the latest version of
llama.cpp, which is due to be installed in the Debian FTP archive.
A summary of the changes between this version and the previous one is
attached.
Thank you for reporting the bug, which will now be closed. If you
have further comments please address them to [email protected],
and the maintainer will reopen the bug report if appropriate.
Debian distribution maintenance software
pp.
Christian Kastner <[email protected]> (supplier of updated llama.cpp package)
(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing [email protected])
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
Format: 1.8
Date: Fri, 16 Jan 2026 18:30:50 +0100
Source: llama.cpp
Architecture: source
Version: 7593+dfsg-3
Distribution: unstable
Urgency: medium
Maintainer: Debian Deep Learning Team <[email protected]>
Changed-By: Christian Kastner <[email protected]>
Closes: 1113813
Changes:
llama.cpp (7593+dfsg-3) unstable; urgency=medium
.
* autopkgtest: Fix detection of skipped tests
* Also make llama-bench(1) reproducible (Closes: #1113813)
* autopkgtest: Temporarily disable models >4GB
Not all workers support these models, and the test runner needs to be
adapted to discover this at runtime
* autopkgtest: Export JSON tests results as artifacts
Checksums-Sha1:
7eb40147daf6afc931a15bd6234308de728e7853 2852 llama.cpp_7593+dfsg-3.dsc
b7b835d1642f8c8a3e3c2f128296d616d99f0b6c 18044
llama.cpp_7593+dfsg-3.debian.tar.xz
55217200b19833c8cfb4082c58e4cce0d8d8178f 6978
llama.cpp_7593+dfsg-3_source.buildinfo
Checksums-Sha256:
11304e8867e18b6910ba258c4c9283e3101fbe7e752ae588f6b0fb0ac5babc83 2852
llama.cpp_7593+dfsg-3.dsc
55b1538051dd5742d6e7922524ad78d46654ba0451b081bf47d2d33fac4d2836 18044
llama.cpp_7593+dfsg-3.debian.tar.xz
700de755a5885b340737402e734c8d25089d17f2269b8948daa6a12d7fe2db20 6978
llama.cpp_7593+dfsg-3_source.buildinfo
Files:
f99a96ba3e6cb948e5fbc88116f492e9 2852 science optional
llama.cpp_7593+dfsg-3.dsc
74e21dc12f38eff5e38071b7cce2139e 18044 science optional
llama.cpp_7593+dfsg-3.debian.tar.xz
fc36de6d29de7f3714a29a576aaeb4f3 6978 science optional
llama.cpp_7593+dfsg-3_source.buildinfo
-----BEGIN PGP SIGNATURE-----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=kNtU
-----END PGP SIGNATURE-----
pgpcsPGsQTa92.pgp
Description: PGP signature
--- End Message ---
