Re: *.lisp in /usr/info, redux
Robert Dodier <[email protected]> Fri, 12 Jun 2026 14:54:18 -0700
| Newsgroups | gmane.comp.mathematics.maxima.general |
|---|---|
| Message-ID | <CAAsY_sR8=HdR5vVNE+qY3iRuSd0n-zdrmXk9Sg9ByN=3fo+-9Q@mail.gmail.com> |
Judah, thanks a lot for dealing with the Slackware package. About the .lisp files among the .info. Those are the index files for the info and HTML documentation. At the end of the file, there is a line of code to call a function to build lookup tables from the items in the index. One of the arguments is a pathname for the directory from which to read the files. At present that's supplied as (maxima:maxima-load-pathname-directory), which returns the pathname of the directory containing the index (and therefore the .info files at present). If (maxima:maxima-load-pathname-directory) is replaced with something else, then the files will be loaded from that place. I don't know what's easier here. I guess the Slackware package could have a patch which replaces the load pathname. Or we could change the code in Maxima (doc/info/build_index.pl) which emits the index .lisp file. Probably the latter is preferable, although I don't have a strong preference at the moment. I think this stuff came up on maxima-discuss not too long ago, although I forget how it played out. About reading compressed info, I would be interested to make that happen, if only because being able to read a compressed stream is generally useful. The simplest way to make that happen is to import a library via Quicklisp or whatever, it doesn't matter how it is loaded so long as it accommodates enough Lisp implementations. The second easiest is to copy a library into Maxima and compile it along with everything else. If the library is small and simple enough, that would be OK too. If we go down the road of reading compressed files, I am leaning towards a Quicklisp (or work-alike) solution. It only needs to work for developers as it would be baked into an image to be distributed. The major sticking point with Quicklisp is that it doesn't work with GCL at present. Although there have been rumblings to the effect that that might happen, I'm not holding my breath. Frankly I think we should get something working whether GCL supports it or not. FWIW Robert