On Sat, Jun 27, 2009 at 07:05:48PM -0500, Karl Berry wrote: > > All. In short (and from memory -- let me know if you observe discrepancies): > > @setfilename foo[.info] > -> foo.info[-*] > -> foo.html/ (subdir if split HTML)
Looking at (texinfo)HTML Splitting, it is said that first the directory name without extension is tried, that is foo/, and If that directory cannot be created for any reason, foo.html/ will be used. > -> foo.html (file if unsplit HTML) > (I don't know what we're going do with chapter-split HTML) It goes in a subdir like split by node html. > -> foo.txt > -> foo.dvi > -> foo.pdf > -> foo.xml (Texinfo XML) > -> foo.dbk (Docbook) It is also foo.xml (unless I missed something). > As for your remark about the undesirability of determining the output > filename by looking at the contents of the input file, well, I agree > with you in principle. I don't think there was any overriding need for > @setfilename in the first place (instead of just using the input > filename) and don't know why rms needed it. But it's too late now. When there is no @setfilename, the texinfo input file basename is used by makeinfo for the output file basename. So, if it is possible to so this in TeX too, maybe then @setfilename could become optional? -- Pat
