This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Bitsets
On Tue, 10 Jun 2003, Zack Weinberg wrote:
> This gives us a 'data structure' library directory which can be used
> for future generic data structures; it also makes the file names quite
> a bit less cryptic. Please do not put a Makefile in libdata/bitset;
> use nonrecursive make techniques.
Please also work out (in consultation with the FSF) the copyright,
licensing and assignment situation for this data structure library from
the start, to avoid the mess with libiberty; in particular, the licensing
(presumably GPL >= 2 for code and GFDL >= 1.2 with no Invariant Sections
or Cover Texts for the manual - in any case, be consistent), the precise
form of the licence notices, and exactly what assignments will suffice for
contributions to this library.
> It would be nice if the documentation were complete, but I don't think
> that should block you from checking this stuff in. Please do,
> however, split up the existing libbitset.texi so that when new data
> structures are added they can just drop another .texi file alongside.
On the manual I will add that it suffers the problems with the libiberty
manual that I previously pointed out
<http://gcc.gnu.org/ml/gcc/2003-06/msg00292.html>. To fix the obvious
problems:
* Use @copying and @insertcopying for the copying information.
* Remember to include a copy of the GFDL in the manual.
* Remove the version numbers (unless this library will have independent
releases) and dates from the manual; they are a pain to keep up to date,
and not particularly useful. (If you wish to bind the library tightly to
GCC, including a version number is easy: gcc-common.texi. If not, it adds
another place to update for releases and branches, which must be
documented, and in any case is a bad idea. Dates tend not to get updated
but are less of a problem if you keep a careful watch on them and make
sure they are updated for releases. But you should put in comments
explaining the basis of the numbers and dates given. I suppose the choice
in the CPP manual to give a major version only, and April 2001 as a date,
is deliberate, but there is no explanation of what is a major enough
change to bump the date.)
--
Joseph S. Myers
jsm28@cam.ac.uk