Re: [Bug-gnulib] MODULES.html replaces MODULES.txt

[Karl Berry]

    CVS? In my experience CVS creates only .#* files, only Emacs creates
    *~ files.

Hmm, seemed like cvs to me (getting created from batch runs), but you're
proabbly right.
    
    They are so obvious - I thought it's more important to not let the
    table become too wide.

It's partly the obviousness or non-obviousness of the names, but also
knowing when one module stops and the next one begins.

    MODULES.html's objective is that anyone can easily see what's in gnulib. 

Certainly.

    An alphabetical list, as I had
    it in the beginning, is useless. 

Not to me.  The current MODULES.html is so long that it is difficult --
for me anyway -- to get a grasp on what is in it.  I like the division
into sections, but some kind of table of contents would be very useful.

    You mean one should add hyperlinks to the real files?

No (although that might be useful too), I was imagining a toc.  Whether
it's alphabetical doesn't matter.  Perhaps a better idea would be a list
at the top that more or less recapitulates the main table:

Support for non-ANSI: assert atexit strtod strtol ...
Enhanced ANSI fns: malloc realloc strftime
Memory mgmt: alloca
Integer arithmetic: gcd
...

where each name is a link to the main definition further down in the file.
Then it would really be possible to see what's there.

If it sounds reasonable but you don't feel like doing it, I can work on it.