Article View: gnu.misc.discuss
Article #75117Software quality, regression tests, etc.
From: Joe Buck
Date: Tue, 06 Nov 1990 19:47
Date: Tue, 06 Nov 1990 19:47
43 lines
2992 bytes
2992 bytes
In article <connolly.657825420@convex.convex.com>, connolly@convex.com (Dan Connolly) writes:
> The GNU development effort is different from typical commercial vendors in
> that it's largely "user contributed" and comes "with no warranty, expressed
> or implied." In short, no company rests its reputation on the quality of GNU
> software.
No longer entirely true. On NeXT and some Data General machines, the standard
compiler IS gcc, and the folks at NeXT and Data General are feeding back their
changes and bug fixes to FSF. I understand that GCC 2.0 will compile C, C++,
and Objective-C and that the Objective-C part is NeXT's contribution; all of
NeXT's work is built on this compiler, so their reputation rests on its
quality.
Similarly, one of the industry-standard SPEC benchmarks is the time to compile
(a part of) GCC. Gnu code is worming its way into industry and more and more
companies are becoming dependent on it. So at least in the case of gcc, more
and more companies have a stake in its quality.
> The result is that the software is strong on features, but sort of beta-test
> quality. The maintainers typically respond outstandingly to bug reports and
> feature requests, but folks who download GNU code still have to count on a
> certain amount of mucking about before everything installs and runs correctly.
I must agree with you that GNU software installation procedures suck; they
assume too much ability on the part of the average system administrator.
Hopefully they will improve over time.
I generally agree with your suggestions for adding regression tests. But
please note that at least some tests (though far from the type of coverage
that is really needed) are included with current distributions. For example,
for gcc, a bootstrap installation procedure is used, as follows:
1) Compile gcc with your native C compiler.
2) Re-compile gcc with the compiler produced in stage 1.
3) Re-compile gcc again with the compiler produced in stage 2.
4) Test that the outputs of stages 2 and 3 are identical.
For this to work, gcc must compile a very big chunk of code (i.e. gcc)
correctly.
--
Joe Buck
jbuck@galileo.berkeley.edu {uunet,ucbvax}!galileo.berkeley.edu!jbuck
Technical Details
Message-ID:
<39411@ucbvax.BERKELEY.EDU>Path:
arc1-ca.pugleaf.net!gmdzi!unido!mcsun!uunet!shelby!agate!ucbvax!galileo.berkeley.edu!jbuckReferences:
<9011040255.AA23968@wubios.wustl.edu> <connolly.657825420@convex.convex.com>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 1025b4,eaf0aab13e254cd4 X-Google-Attributes: gid1025b4,public X-Google-ArrivalTime: 1990-11-06 11:47:54 PST Path: arc1-ca.pugleaf.net!gmdzi!unido!mcsun!uunet!shelby!agate!ucbvax!galileo.berkeley.edu!jbuck From: jbuck@galileo.berkeley.edu (Joe Buck) Newsgroups: gnu.misc.discuss Subject: Software quality, regression tests, etc. Summary: was Re: >Gnu *nix: The time for it is now! (fwd) Message-ID: <39411@ucbvax.BERKELEY.EDU> Date: 6 Nov 90 19:47:54 GMT References: <9011040255.AA23968@wubios.wustl.edu> <connolly.657825420@convex.convex.com> Sender: usenet@ucbvax.BERKELEY.EDU Reply-To: jbuck@galileo.berkeley.edu (Joe Buck) Distribution: gnu Lines: 43 Posted: Tue Nov 6 20:47:54 1990