Bug 59624 - прошу добавить systemd юнит для запуска сервера
Summary: прошу добавить systemd юнит для запуска сервера
Status: RESOLVED FIXED
Alias: None
Product: Sisyphus
Classification: Development
Component: whisper-cpp (show other bugs)
Version: unstable
Hardware: x86_64 Linux
: P5 normal
Assignee: esgor@altlinux.org
QA Contact: qa-sisyphus
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2026-06-23 17:16 MSK by Alexey Shabalin
Modified: 2026-07-27 18:34 MSK (History)
3 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Alexey Shabalin 2026-06-23 17:16:06 MSK
Прошу добавить systemd юнит для запуска сервера.
Для примера можно посмотреть родственный пакет llama.cpp.
Comment 1 gorbanyoves@basealt.ru 2026-06-30 09:10:50 MSK
В документации whisper.cpp есть предостережение:

Warning
Do not run the server example with administrative privileges and ensure it's operated in a sandbox environment, especially since it involves risky operations like accepting user file uploads and using ffmpeg for format conversions. Always validate and sanitize inputs to guard against potential security threats.

https://github.com/ggml-org/whisper.cpp/tree/master/examples/server#usage
Comment 2 Anton Farygin 2026-06-30 09:28:36 MSK
так не надо запускать без конфига.

ну и hardening добавить в юнит, на всякий случай.
Comment 3 Repository Robot 2026-07-27 18:34:43 MSK
whisper.cpp-1.9.1-alt3 -> sisyphus:

* Mon Jul 27 2026 Alexey Shabalin <shaba@altlinux> 1.9.1-alt3
- Refactor spec to match llama.cpp packaging structure.
- Move gcc12-c++ and vulkan build requirements to their conditional blocks.
- Make ggml model packages optional: the main package no longer requires
  whisper.cpp-ggml-base, models can be downloaded manually instead.
- Package ggml-base.en model as separate optional noarch subpackage.
- Find ggml backends near the executable without /proc (fixes test
  segfaults in hasher, port of the llama.cpp ALT fix).
- Run the full test suite and whisper-bench in %check.
- Replace whisper-server.service with parametric whisper-server@.service
  (as in llama.cpp): per-instance environment files in /etc/whisper,
  ready-made instances (base, base.en) are shipped with the model packages.
- Build with SDL2: package whisper-stream and whisper-command as separate
  subpackages (whisper-talk-llama and whisper-lsp are not packaged).
- Use WHISPER_COMMON_FFMPEG instead of the deprecated WHISPER_FFMPEG.
- Enable GGML_CPU_ALL_VARIANTS on riscv64 (as in llama.cpp).
- Add .gear/version-up for maintenance with zoryn.
* Tue Jul 21 2026 Evgeniy Gorbanyov <esgor@altlinux> 1.9.1-alt2
- Renamed to whisper.cpp (Closes: #59623).
- Added whisper-server.service systemd unit (Closes: #59624).
- Added whisper-cpu, whisper-cuda and whisper-vulkan packages (Closes: #59625).
* Tue Jun 23 2026 Evgeniy Gorbanyov <esgor@altlinux> 1.9.1-alt1
- Updated from 1.8.6 to 1.9.1.
* Tue Jun 09 2026 Evgeniy Gorbanyov <esgor@altlinux> 1.8.6-alt1
- Updated from 1.8.4 to 1.8.6.
* Tue May 26 2026 Evgeniy Gorbanyov <esgor@altlinux> 1.8.4-alt2
- The nvcc-12.9 compiler doesn't support gcc15. gcc14 is required
  to build with CUDA support.