On Mon, 23 Aug 1999, Hubert Feyrer wrote: > On Tue, 24 Aug 1999, NAKAJIMA Yoshihiro wrote: > > I found that `textproc/t1lib' and `textproc/t1libx' have the > > same PKGNAMEs. > > > > % grep PKGNAME textproc/t1lib*/Makefile > > textproc/t1lib/Makefile:PKGNAME= t1lib-0.8 > > textproc/t1libx/Makefile:PKGNAME= t1lib-0.8 > > > > Is it right? > > No. I started work on updating t1lib to 0.9.1, including the X stuff, incidently wiping out t1libx, but got distracted by other things. It looks like all that's left to do is construct a default config file. The test app, "xglyph", runs in place but not installed.
for this sort of documentation? > > > > > > It looks like some of the docs under /usr/share/doc/psd > > > use the me package. Is this the recommended one to use > > > for new documentation? > > > > No, it should be done in -mdoc (see `man 7 mdoc', and `man 7 mdoc.samples') > > for a variety of reasons. > I would say that this sort of documentations need a "table of contents" > and benefit from "decimal numbered sections", which -me and -ms provide, > but -mdoc not. Or have I missed some thing in -mdoc? mdoc does provided numbered lists, and adding a number sequence to a document using troff number registers is trivial. A table of contents is a bit more work, but not difficult. I would think that, these days, hypertext references are more important than page number references for the dead tree edition. At the moment, only mdoc provides a section cross reference directive, and this is automatically translated into hypertext links by -mdoc2html. But, point taken. A set of "NetBSD TOC macros" might be a good thing. > ... > > * very nice looking output > For man-pages, but for longer documents (see above)? > > ... > > * we have the -mdoc2html macros in-tree > A few days ago I saw an article in <news://gnu.groff.bug> about an html > back end and that someone starting up to maintain groff. (If someone is > interest in the thread, I have saved at least the interesting parts.) I believe that text conversion at the back-end level is misguided; it's an attempt to get a lot of bad html at low cost. The approach taken by -mdoc2html is to translate at the request level. Consequently, information as to the structural nature of the marked-up text is still available. This information is completely filtered out by the time a post-markup text-domain (i.e., back end) tool gets to it. These tools do things like scan for http:// in text and make them links. Such a tool can't do relative section linking, because it doesn't actually know what is a cross reference and what isn't. It can only guess... > ... > > But troff-ing Packages.txt is, in general, an excellent idea, and actually > > As would it be for INSTALL. Both are long documents, which will need a > table of contents with correct page numbers. It has already been done for INSTALL, although, without a page-numbered TOC. email@example.com