>>>   - We disable the use of xmltex for now.  It just doesn't work
>>>     when building the Users' Guide.  Perhaps it can be enabled
>>>     if explicitly requested, but configure won't use it by default.

it would be nice to have a configure test for the xmltex problems,
but if we knew what the problems where, we might not have them?-)

>> Seems OK to me, but perhaps we can disable xmltex only on WinDoze
>> builds? 
>I seem to remember I got pretty much the same results as Claus on Unix
>too.  Does it work for you?

- the 'underscore in file name' problem seems to be known on unixes as well.

- as for the 'computed lengths' cause 'missing item' problem, I suspect that
    there is an extra xmltex/passivetex package for dealing with that (as there
    is for latex), but I don't know which one that may be (and whoever uses
    it in the translation, should also 'require' its existence)? 

    as I said, I have a tiny perl-script to hack around that problem (attached, 
    run in users_guide/ as './fo.pl <users_guide.fo >users-guide.fo'; then 
    use xmltex on the output).

- as for the 'missing number, treated as zero', it could simply be that 
    someone is lying: in flags.xml, all tables are declared as having 3 columns,
    but they all have 4 columns!-) (diff attached)

    the html formatter doesn't seem to care about such declarations, but
    fop seems to prefer declarations over use, and will actually drop the 
    extra column (quietly!?); xmltex at least produces error messages, even
    if they are somewhat attrocious/useless, but for the approximate linenos.

addressing these three issues allows me to run xmltex over users-guide.fo
and get a .dvi file that doesn't look too bad, so perhaps it won't be necessary 
to disable xmltex after all, just figure out how to avoid the problems while 
building, and comment that both routes do not yet produce nice output.

cheers,
claus

Attachment: flags.xml.diff
Description: Binary data

Attachment: Fo.pl
Description: Binary data

_______________________________________________
Cvs-ghc mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/cvs-ghc

Reply via email to