Debian 宣布 Debsources，可搜索软件包源代码

Debian 项目宣布 Debsources，允许浏览和搜索所有 Debian 发行版软件包源代码的 Web 服务，现在是托管在官方Debian基础设施在https://sources.debian.org。

Debian 称，该服务之前托管在域名 sources.debian.net（已重定向到 sources.debian.org）， 他们现在将其迁移到 Debian 硬件和基础设施上，并借此机会正式宣布了这一服务。Debsources 内容可通过 HTML 界面和 JSON API 访问，会定期更新，目前是一天更新四次。

Debsources 被广泛用来进行各种 Debian 的源代码研究。

Dear developers,

We're happy to announce that Debsources, the Web application that
allows to browse and search the entire source code of all Debian
releases, is now hosted on the official Debian infrastructure and
available at https://sources.debian.org.

You may already know this service as previously hosted at
sources.debian.net . We took the move to Debian hardware as the
opportunity to officially announce it here.


Service description
===================

Debsources is a web service that exposes the content of Debian source
packages on the Web, both via an HTML user interface and a JSON API
[0]. To that end, an updater runs regularly (currently 4 times a day)
to unpack and index new packages from all the available Debian
suites. It also runs various plugins, for example to count lines of
code and measure the size of packages; this information is exposed in
the interface.

[0]: https://sources.debian.org/doc/api/

Since all extracted information is stored in a PostgreSQL database,
various data warehouse features are available, such as the list of
file duplicates for every file, or package name search.


More than source code
=====================

Worth noting are two "sub-apps" that run on the same infrastructure:

* patches [1]: for packages using 3.0 (quilt) package format, the list
  of patches is extracted and exposed, along with some metadata.

* copyright [2]: the copyright files of packages are exposed through
  this interface, and displayed in a convenient way if they are
  machine readable.

[1]: https://sources.debian.org/patches/
[2]: https://sources.debian.org/copyright/


Learn more
==========

Debsources has been described extensively and used to conduct various
studies on the Debian source code. To learn more check out the paper:

Matthieu Caneill, Daniel M. Germán, Stefano Zacchiroli. The Debsources
Dataset: Two Decades of Free and Open Source Software. In Empirical
Software Engineering, Volume 22, pp. 1405-1437, June, 2017

An open access preprint of the paper and additional information about
Debsources are available on the Debsources documentation page [3].

[3]: https://sources.debian.org/doc/about/


Credits
=======

Many credits go to everyone who submitted patches and feedback over
the years, thank you! In particular, Stefano Zacchiroli started the
project and wrote the updater infrastructure, and Orestis Ioannou
developed the patches/ and copyright/ apps.

This new instance of the service is hosted by DSA - without whom
sources.debian.org would not exist. Many thanks to the whole team for
their guidance in the d.n -> d.o migration, and in particular to
Julien Cristau and Peter Palfrader.

The previous instance of Debsources had been hosted by IRILL [4] for
more than 4 years, who has also spearheaded the initial development of
the service - thanks for the non-interrupted service!

[4]: https://www.irill.org/


As always, we're more than happy to hear your feedback; you can
contact us at qa-debsources@lists.alioth.debian.org . (See the footer
of the Debsources web page for bug report and contribution
information.)

For the Debsources team,
--
Matthieu

附件: signature.asc
Description: PGP signature