Re: Nondeterministic compilers?

"Walter" <walter@digitalmars.com>
2 Mar 2004 11:07:11 -0500

          From comp.compilers

Related articles
Nondeterministic compilers? gah@ugcs.caltech.edu (glen herrmannsfeldt) (2004-02-26)
Re: Nondeterministic compilers? christian.bau@cbau.freeserve.co.uk (Christian Bau) (2004-02-26)
Re: Nondeterministic compilers? hannah@schlund.de (2004-02-27)
Re: Nondeterministic compilers? franck.pissotte@alussinan.org (Franck Pissotte) (2004-02-27)
Re: Nondeterministic compilers? alexc@std.com (Alex Colvin) (2004-02-27)
Re: Nondeterministic compilers? joachim.durchholz@web.de (Joachim Durchholz) (2004-03-02)
Re: Nondeterministic compilers? walter@digitalmars.com (Walter) (2004-03-02)
Re: Nondeterministic compilers? nmm1@cus.cam.ac.uk (2004-03-06)
Re: Nondeterministic compilers? norm@plumpirate.com (norm rubin) (2004-03-11)
| List of all articles for this month |

From: "Walter" <walter@digitalmars.com>
Newsgroups: comp.compilers,comp.lang.pl1
Date: 2 Mar 2004 11:07:11 -0500
Organization: Comcast Online
References: 04-02-158
Keywords: practice
Posted-Date: 02 Mar 2004 11:07:11 EST

"glen herrmannsfeldt" <gah@ugcs.caltech.edu> wrote in message
> Does anyone know about any compilers that are not deterministic?


Non-deterministic compilers are something to be avoided. Having
non-repeatable results makes them hard to debug. How can one fix
problems if the customer's problem is erratic and cannot be reproduced
by the vendor? Also, when shipping a project built with a compiler,
it is good practice to archive not only the source to the project but
the tools and compiler used to build it. Good QA practice verifies
that the project can be rebuilt exactly byte for byte from the
archived version. A non-deterministic compiler would be unreliable for
this, and hence not of interest to any organization who wants to
produce quality code.


-Walter
www.digitalmars.com free C/C++/D compilers


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.