forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ross Gardler <>
Subject Re: avoiding unnecessary changes in output formatting
Date Thu, 22 Dec 2005 10:13:35 GMT
David Crossley wrote:
> Johannes Schaefer wrote:
>>Tim Williams schrieb:
>>>Johannes Schaefer wrote:
>>>>>Author: rgardler
>>>>>Date: Wed Dec  7 11:58:13 2005
>>>>>New Revision: 354838
>>>>>check for existence of @class to prevent an extra space being added when
not needed
>>>>Is this necessary? The extra space does not do any harm.
>>>>If "Yes", I'll update my recent comitts accordingly.
>>>>Sorry for not noticing/reviewing earlier.
>>>I think this came up because it caused unexpected changes in the
>>>forrestbot build to our docs.  Someone will let us know if I'm wrong.
>>wow! small change big effect!
>>Will do my best to avoid extra spaces, then.
>>Thanks for the link.
> Yeah, when i saw our huge commit message for the generated
> docs i had visions of our users. Many users have the
> generated docs in version control. So it is important to
> keep the diffs minimal.
> Complaints are already heard on other lists about the
> verbose html output from Forrest.
> Part of their problem is committers not configuring their
> svn clients properly, so they see diffs to PDFs too.
> We have improved a lot lately with our CSS-based design.
> We had a discussion recently on dev about using a
> final transformer step at the end to tidy the
> html output into consistent format. Or perhaps
> that is the job of the html serializer. Dunno.
> Anyway should we do that? Users would get one big
> change after they upgraded tp 0.8 and thereafter
> the changes would be minimal and output would be
> consistent. A side-benefit is that it is human-readable.

How about we do this as part of the Dispatcher work. There would be no 
"big change" for those not moving from Dispatcher and it can be another 
of the "selling" points for the new publication.

Of course this means there would be a reasonably long wait for it to 
happen. Whereas the transformer/serializer approach could be done 


View raw message