Subject: Re: CVS commit: basesrc/distrib/notes
To: matthew green <mrg@eterna.com.au>
From: John Hawkinson <jhawk@MIT.EDU>
List: source-changes
Date: 11/26/2001 19:46:35
matthew green <mrg@eterna.com.au> wrote on Tue, 27 Nov 2001
at 11:40:47 +1100 in <22849.1006821647@splode.eterna.com.au>:

> is there a good reason for these _generated_ files to be commited?

I think so. Generation is a bit slow and potentially weird
as it may require running the generation process multiple
times, since the size of the table of contents affects the
page numbers of subsequent pages.

By doing it this way, it doesn't slow down the regular generation
of the release notes, and normal people generating the release
notes don't have to worry about making sure they have a consistent
table of contents generated.


> > also, the name "PostScript" is not consistent with the "ps" used
> in eg, INSTALL.ps.

Correct.

As noted in rev 1.12 of distrib/notes/common/macros:

 | At present, the TOCs are generated files, and can be regenerated
 |   with "make tocs" (probably has to be run twice to account for
 |   the size of the table itself).
 | TOCs are named INSTALL.{PostScript,ASCII,HTML,more}.toc instead of
 |   INSTALL.{ps,txt,html,more}.toc because that's what \*[format]
 |   expands to.

One could add a lookup table to swap them around but it did not seem
worth the bother.

--jhawk