Hello. [ I've added [EMAIL PROTECTED], since this is more than a discussion about the patch now. Maybe I should have included it in the original recipient list. ]
Alexandre Duret-Lutz wrote: > > >>> "Richard" == Richard Dawe <[EMAIL PROTECTED]> writes: > > Richard> Hello. > > Richard> Below is a patch that adds HTML documentation > Richard> generation in a similar way to PDF, PostScript, etc. - > Richard> by adding 'html' targets for texinfo sources from > Richard> info_TEXINFOS. > > The problem, with html, is that nobody agree about what the > ouput should be. I'd say that if we support html, we should > use the default makeinfo output (which is to split on nodes), > and let the user say `AM_MAKEINFOFLAGS = --no-split' when wanted. > This is what we do for info files already. OK. But then it gets hard to know which files to remove. Removing *.html seems a bit dangerous to me. Perhaps some support script could produce a list of nodes, so that we know remove the right HTML files. What should it be called? texinodes? > Still some people prefer using texi2dvi, and some also want to > distribute html files. It seems hard to satisfy these. Do you mean texi2html, rather than texi2dvi? I think using makeinfo should be the default. Then how about supporting an option in AUTOMAKE_OPTIONS, to use texi2html instead: html-texi2html? Clearly texinodes (or whatever it's called) would need two modes, to cope with different output filenames: makeinfo- and texi2html-modes. > Richard> I know this is different to the way proposed in > Richard> TODO. I read TODO afterwards. ;) > > Let's ignore TODO. I don't think we need html_TEXINFOS. Not until > the GNU Coding Standards introduce a $(htmldir) variable. > > Right now we only install info files in $(infodir). Other > ps/pdf(/html) targets are there only for convenience. > > Richard> For consistency, if an html_TEXINFOS thing (primary?) > Richard> is added, it seems that there should be pdf_TEXINFOS, > Richard> etc. too. > > eww... 8) > Richard> This patch also does not update the example given > Richard> about html_DATA in the automake.texi manual. So this > Richard> patch isn't complete - it's more a conversation > Richard> starter. ;) > > For a starter the patch looks very great to me! I think it is > worth a NEWS entry. Also the test case should better make sure > files are cleaned up properly (I'd do this with `$MAKE distcheck'), > with or without --no-split (if we are going to support this). OK. I think we should support --no-split. I was trying to make an "easy" starter patch. Thanks, bye, Rich =] -- Richard Dawe [ http://www.phekda.freeserve.co.uk/richdawe/ ]
