forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From David Crossley <cross...@apache.org>
Subject Re: svn commit: r432954 - in /forrest/site: live-sites.html live-sites.pdf
Date Wed, 23 Aug 2006 07:50:45 GMT
Gav.... wrote:
> > From: David Crossley
> > Gav.... wrote:
> > > Incidentally, I changed my site.xml file to read :-
> > >
> > >         <fixcrlf srcdir="${project.site-dir}" eol="unix">
> > >                 <include name="**/*.html"/>
> > >                 <include name="**/*.css"/>
> > >         </fixcrlf>
> > >       </then>
> > >       <else>
> > >         <fixcrlf srcdir="${project.site-dir}">
> > >                 <include name="**/*.html"/>
> > >                 <include name="**/*.css"/>
> > >         </fixcrlf>
> > >
> > > So in effect adding *.css to the FOR-492 workaround, shall I commit this
> > do
> > > you think or will it need further investigation?
> > >
> > >    As with FOR-492 and problems mentioned there, my builds kept putting
> > ^M
> > > (CR-LF) on every line of every CSS file causing my build & deploy to
> > fail.
> > 
> > Are you sure that this is FOR-492. It seems like something else.
> > FOR-492 adds mixed line-edings only for processed xml comments.
> > 
> > If it is different, please add a new Issue. Then yes
> > apply the workaround, and note the new FOR-#
> 
> Ok, slightly different problem, same solution, I'll create a new issue.

Good, thanks.

> > I keep asking for you to show us the diffs for these
> > processed CSS files. Is it every line changed? Which files
> > are changed.
> 
> I did post some diffs, and at other times just thought that explanations
> Such as 'every line has a ^M on the end of every CSS file' would suffice,
> I'll undo my fix and send them though if you think you still would like to
> see them.

No, don't bother. Sending a diff to the issue tracker is usually
much easier than needing to explain in words. Also remember
that this topic has gone through many email Subject changes,
so hard to keep track.

> Expanding on that though just in case, all CSS files in the source /common/
> and /pelt/ styles sheets are and were always fine, just when a forrest -f
> publish.xnl build was run, resultant .css files in the build/skin/ directory
> all had ^M line endings. Deploying from these of course would copy these
> over into the /work/skin directory with ^M in them also.
> 
> > > Adding the above workaround seems to have cured this and the build
> > directory
> > > now has proper line endings for the CSS files (In turn therefore the
> > work
> > > directory also now has correct line endings).
> > 
> > Yes, because this is a brute-force approach. The workaround
> > will work for all kinds of line-endings problems.
> 
> Are there any negatives to applying this approach ?

It will work. The problem is that it is just covering
up the real issue. Lets move on, but please add
the new issue number as as a fixme note next to the
hack as was done for the other workarounds.

> > > No other files 'seem' to be affected at the moment, but will keep an eye
> > out
> > > on my Windows system.
> > 
> > Are you using Cygwin?
> 
> No, It didn't like my computer and so I had to get rid of it.

That might point to other problems.

> There are plenty of Windows devs in the ASF and they can't all be pretending
> to use linux.

In a previous recent thread about publishing our
website, Ross said that he did it sucessfully with Windows.

> I Have a linux box with FreeBSD on it, I'll get round to
> configuring that at some stage I guess, but in the meantime, don't want to
> mess with my production laptop too much.

Forrest should be able to work with any operating system.

Perhaps the type of problems that you have with your system
are the cause of some of the complaints we have heard about
Forrest making unnecessary output changes. Grr.

-David

Mime
View raw message