maven-doxia-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Lukas Theussl <>
Subject Re: Solving links and anchors
Date Tue, 14 Aug 2007 09:49:30 GMT
I also don't like disallowing dots in ids. Hhyphens ("-"), underscores 
("_"), colons (":"), and periods (".") are perfectly legal in ids 
according to the html specs, I don't see any reason why doxia should 
replace them. Underscores where used in the m1 xdoc plugin to replace 
any illegal characters, which I think is better than just stripping 
them, for readability reasons (think section titles). However, the use 
of automatically constructed ids from section titles was discouraged in 
the last version of the xdoc plugin [1] as it only led to trouble (which 
  btw very much reminds me of our current ones, see MPXDOC-158 and 
MPPDF-40 for related discussions).

The root of the problem are the conventions used by aptconvert, and 
apart from being undocumented, I don't really like them, so the question 
is: how compatible do we want/have to stay with that?

If I am not mistaken, the issue should be solved simply by requiring 
that links to other documents have to start with './' or '../', as I 
indicated at DOXIA-47. The only problem then is backward compatibility, 
since currently this is rarely the case (eg the project reports page in 
the site plugin [2] uses report.getOutputName() + ".html" );

where report.getOutputName() doesn't start with './').

However, in principle I think this is the way it should work (and always 
should have in the first place), so maybe we try to clear that up and 
document it for beta-1?



Vincent Siveton wrote:
> 2007/8/11, Dennis Lundberg <>:
>>As you probably saw on dev@maven we had some problems yesterday with
>>links being rendered in the wrong way by doxia. I have made a temporary
>>fix for this, but I feel that we need to do more to get a stable solution.
>>The issues which are related to these problems are
>>My temporary fix reverses the logic. Instead of trying to see if a text
>>is an "external link", I try to determine if a text is *not* an
>>"internal link". With a few minor adjustments (see my comment on
>>DOXIA-131) I think that this would be a more stable solution.
>>What do you think?
> Sounds goods.
> IMHO, the ideal solution will be to not change the util encodeId()
> (which respects HTML spec) to disallow dot.
> Cheers,
> Vincent
>>Dennis Lundberg

View raw message