GPC with gcc-4.
frank at fjf.gnu.de
frank at fjf.gnu.de
Thu Jul 14 16:27:17 CEST 2011
Adriaan van Os wrote:
> By the way, the gpc mailing list archive seems to be changed.
> I preferred the old archive
Well, thanks. It's always nice to hear so -- especially after it was
decomissioned. ;-) While it was active, I rarely got positive
feedback, only sometimes complaints. (BTW, I think it was you who
complained that the old one displayed email addresses in clear --
the new one mangles them a bit at least.)
Anyway, the thing is, I wrote the old archive program (crystal) for
a website for a specific project several years ago which is long
done and its website taken offline. So my direct need to support
this program exists no more.
Since crystal was written in GPC, it seemed appropriate to use it
with the old GPC mailing list (majordomo) which didn't have an
archive of its own.
But software is aging and maintenance effort was starting to
increase. E.g., a major version upgrade of MySQL required some SQL
syntax changes which Benedikt Wildenhain made (but the next upgrade
may require further changes). Upgrades of the web server (apache)
and email server (exim) so far haven't required changes to crystal,
but it can't be ruled out that future upgrades will. And, of course,
being written in GPC means that we can't even build it with a
current GCC (not to depreciate Waldek's efforts, but I hope it's
obvious that I wouldn't install a program on a public web site built
with a compiler with major changes (for gcc-4.x) that haven't
received much testing apart from running the test suite AFAIK, let
alone security auditing).
Furthermore, we (at least me as well as Anja Gerwinski who maintains
the mailing lists) have never been very happy with majordomo, so
when the GPC web site was moved to a new server (hardware), we
decided to take this opportunity to move all mailing lists to
mailman. Since mailman has its own archives, it seems logical to use
them now and get rid of the maintenance of crystal. As mailman is
rather widespread, probably a number of users are used to its
> (and it breaks a lot of links).
Indeed, that's a problem, but perhaps we can solve it with
More information about the Gpc