home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.unix.questions:15904 comp.sys.hp:15262
- Newsgroups: comp.unix.questions,comp.sys.hp
- Path: sparky!uunet!ukma!darwin.sura.net!ra!atkinson
- From: atkinson@itd.nrl.navy.mil (Randall Atkinson)
- Subject: Re: supported version control systems
- Message-ID: <C1DHs8.5pB@ra.nrl.navy.mil>
- Followup-To: comp.sys.hp
- Sender: usenet@ra.nrl.navy.mil
- Organization: Naval Research Laboratory, DC
- References: <1993Jan11.195216.6253@sierra.com> <C1DAGw.LD4@hpuerca.atl.hp.com>
- Distribution: na
- Date: Sun, 24 Jan 1993 19:21:44 GMT
- Lines: 39
-
- Someone wrote:
- : Well, you may want to tell the powers that be that RCS is a public domain
- : system as well. If it makes any difference, I've had nothing but great
- : results with CVS and RCS.
-
- In article <C1DAGw.LD4@hpuerca.atl.hp.com> sailer@hpuerca.atl.hp.com (Lee Sailer) writes:
- > Yes, there are FD versions of RCS, CVS, and other software. However, the
- > HP fully supports the version of RCS that they ship, so in that
- > sense HP RCS is "better" than FD RCS.
-
- >Of course, FD RCS is "better" in other senses.
-
- For example, the RCS supplied by HP with my Softbench stuff
- occasionally dumps core in the manner I've seen with other programs
- that try to use the mmap() function [mmap() is officially not
- supported by HP and is known to be broken on the HP 7xx with HPUX 8x].
- Someone who works at HP on Softbench might want to check and see if
- they accidentally left an mmap() call in their RCS sources.
-
- The fix for this is to move to the freely distributable RCS and CVS
- sources (being quite sure to compile them with NO_MMAP configured so
- that no attempt is made to call mmap() :-). This also means that one
- gets a more capable version of RCS/CVS than HP sells. Funny how the
- free software is more dependable than the HP software (recall that HP
- keeps telling us customers that their software quality control is why
- we pay so much for their software :-).
-
- Oh, folks compiling GNU software that uses the GNU "configure"
- program should note that the "configure" program sees the mmap()
- declaration in the header file and assumes that it works (while it
- doesn't actually work under HPUX) so you normally have to manually
- edit the output of the configure program to explicitly reset the
- HAS_MMAP thingy to indicate that there is no mmap() function.
- If one does this, problems with getting the GNU software running will
- be significantly reduced (though HPUX is still remarkably time
- consuming to port software to).
-
- Ran
- atkinson@itd.nrl.navy.mil
-