Raphael Geissert <[EMAIL PROTECTED]> writes: > Russ Allbery wrote:
>> It already takes over a day for Lintian to process the entire archive, >> so I don't have any immediate plans to run it on more architectures >> unless we can find some dramatic way to speed it up or find other >> places to run it besides gluck.d.o. > What about running lintian once a day on the new/updated packages of > every day? that way the amount of time spent would be reduced and all > the lintian reports would have a delay of no more than one day. That's what we do now. It takes about a half-hour, usually, or less. However, when I upgrade lintian.d.o to a new version of lintian, I have to regenerate the entire results for the whole archive to have consistent results, and I don't want that to take a week. I suppose we could do something fancy where the archive is rebuilt in the background, but gluck doesn't have a lot of spare cycles as it is, and I'm not sure it's worth the effort at the moment to get coverage of the remaining platforms. > groff was the process taking the most part of the CPU when lintian > processed libwine-dev, is there anything that can be done to reduce > this? The only options I can think of are: 1. Figure out some way to make groff faster. 2. Parse nroff in Perl and have lintian do the same checks internally. 3. Stop checking man pages. None of the options are particularly appealing except 1, which is just really hard. :) -- Russ Allbery ([EMAIL PROTECTED]) <http://www.eyrie.org/~eagle/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
