incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "B. W. Fitzpatrick" <>
Subject Re: cvs commit: incubator-site/build/site/projects/altrmi/api/org/apache/altrmi/server/impl/socket AbstractCompleteSocketStreamServer.html AbstractPartialSocketStreamServer.html CompleteSocketCustomStreamPipedBinder.html CompleteSocketCustomStreamPipedConnection.html CompleteSocketCustomStreamServer.html
Date Wed, 19 Mar 2003 16:03:03 GMT
Rodent of Unusual Size <Ken.Coar@Golux.Com> writes:
> Nicola Ken Barozzi wrote:
> > 
> > I agree that putting generated stuff in CVS sucks big time.
> > The point is that users and developers like having these on the site.
> > 
> > For example, I might need the docs but have only the jar ATM, or I want 
> > to explain a thing on the list with a link to the javadocs.
> > 
> > Anyway, a solution for Java projects seems simple: make the project run 
> > under Gump, and these generated files will be published daily.
> > But this gives us the latest docs, not the ones of the last release.
> > 
> > So IMHO we should push these docs to the site by hand and not use the 
> > site CVS for it.
> i think you are missing one important aspect by making
> the unconscious assumption that people who want the
> documentation are *always* online when they need it,
> and can reach the apache site.  from personal experience,
> my own and that of others shared with me, i can tell you
> that such is far from always being the case.  if they
> cannot easily produce it from raw sources in such a
> situation, it must be available to them in pre-generated
> form as part of a/the download they use to get the sources.
> what are the technical (not philosophical) objections to
> keeping generated data under cms?

I think you're confusing "Version Control System" and "Content
Management System".

If people can't easily generate the docs from a checked out tree, then
make them available in a tarball somewhere.  Just because you check
your code out of CVS doesn't mean that you have to get the known
universe from CVS too.

But if you want a technical objection to keeping docs in CVS, how's
this: It eats up a massive amount of disk space (O^n for the size of
each commit) in the repository, which must be backed up
somewhere--that's a huge waste of valuable storage and backup space.


Brian W. Fitzpatrick    <>

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message