Biggest debian source packages

Posted Sep 3, 2015 20:29 UTC (Thu) by zack (subscriber, #7062)
In reply to: Biggest debian source packages by ededu
Parent article: Debsources as a platform

> Does anyone know how can we see what projects (source packages) are the biggest in terms of SLOC in debian? For example, looking at https://sources.debian.net/src/linux/4.1.6-1 I see that linux 4.1 has almost 13 Mlines of code, and likely libreoffice 5.0.1 has about 4.5 Mlines of code. However, I would like to have a list in decreasing order of size.

There is no interface to do complex queries like the one you're asking for on the web, but your question can easily be answered using the Debsources DB.

The top-ten (not counting different versions of the same packages) is like this (format is sloc|package|version):

I've put a list that includes all packages that have at least 100ksloc up here http://upsilon.cc/~zack/stuff/debsources-pigs.20150903.txt

While we haven't yet automated the periodic publishing of Debsources database dumps, one that is ~6 month old is available from Zenodo at https://zenodo.org/record/16106, for those who might want to play with it.

Some aggregate stats extracted from the same database (but in 2014) are available in this paper: https://upsilon.cc/~zack/research/publications/debsources...

Hope this helps,
Cheers.

(Log in to post comments)

Biggest debian source packages

Posted Sep 4, 2015 9:46 UTC (Fri) by jwilk (subscriber, #63328) [Link]

No gcc-5 on the list... How come?

Biggest debian source packages

Posted Sep 4, 2015 10:00 UTC (Fri) by zack (subscriber, #7062) [Link]

Here is why: http://sources.debian.net/src/gcc-5/latest/

The GCC package in Debian still uses tarball-in-tarball, which Debsources does not expand (as mentioned in the article), so sloccount returns bogus results on that package.