Stephen Schaub wrote:
> >Stephen Schaub wrote:
> >>
> >> I would love to see better error reporting. Can we get a line number in
> the
> >> error messages when forrest encounters an error in a source file in the
> >> xdocs folder?
> >
> [snip]
>
> Dave Brondsema wrote:
> >Since the wiki files are parsed with the Chaperon parser, this is an
> >issue you should discuss on one of their mailing lists. Unless of
> >course chaperon does provide better error messages but forrest hides it.
>
> But this is _not_ just a wiki issue. Forrest doesn't provide line numbers
> for errors in .xml files either.
I do 'forrest run' in one window and use 'forrest validate'
in another window. Actually i often use 'xmllint' to validate
just the xdoc that i am working on.
> Perhaps that is Cocoon's fault. I've had
> some exposure to Cocoon in the past, and seem to remember no line numbers
> there, either. That really surprised me, because Cocoon is a highly regarded
> product, but its error reporting (at least, in the version I was using) was
> abysmal.
Cocoon has always said that it expects you to feed it
proper xml files. It is not Cocoon's job to validate the
xml for you.
> Good error reporting is vital. Developers put up with poor error reporting
> from development tools because we've come to expect it. But at least those
> cryptic compiler error messages typically provide a line number to help us
> locate the problem. A product like Forrest that is designed for non
> programmers should have at least an equivalent, if not a higher, standard
> for its error reporting.
Have you also looked at the logfiles in build/webapp/WEB-INF/logs/
> I'll try to dig into the wiki side of things to see if Chaperon is at fault
> here. But I don't feel this discussion should be dismissed on the basis of
> "the problem is in someone else's code."
The Wiki files are where the issue is. If someone knows
about a separate Wiki validator, then we could plug that
into our build system so that the wiki files are included
in 'forrest validate'.
--
David Crossley
|