David Miller justdave at
Thu Feb 5 06:27:10 UTC 2004

On 1/27/2004 9:22 PM -0500, David Miller wrote:

>    I'd like to propose completely removing the HTML, Text, and PDF versions
> of the docs from CVS.  Under my plan, the website will automatically build
> those directories upon pulling changes to the xml directory, and the
> tarball build script will be changed to build the docs after checking out
> of cvs prior to rolling the tarball.
> Reasons for doing this:
> 1) those directories are already off-limits for direct modifications in cvs
> anyway, they're required to be generated from the XML before checkin
> 2) every time they're regenerated, it "modifies" every single file, which
> results in a HUGE chunk of space in the cvs commit log on bonsai.

This is now done on the trunk.  The html, txt, and pdf directories are now
gone, and the script will create them if they're missing.  The website now automatically regenerates the html, txt, and pdf
docs any time updates to the xml directory are detected.  This means any
changes made to the xml files will go live in the html on the website as
soon as you commit them instead of having to wait until the next time
someone compiles the html docs. :)

I plan to do this on the 2.16 branch also, but haven't gotten around to it yet.
Dave Miller      Project Leader, Bugzilla Bug Tracking System   

More information about the developers mailing list