Re: OT Verifying code across gcc versions [Re: Commercial support on GCC/G++ open source versions]

[Paul Pluzhnikov]

Maurizio Loreti <mlo@foobar.it> writes:

> > > So what gives you confidence to "jump to new version as soon as it
> > > is released"?
> 
> Because we have exetensive test cases ready; and, for example, we are
> testing future compiler releases with their cvs snapshots (see my
> comments about gcc release 4) and preparing bug reports for the gcc
> working group if any fails. 

Ok, so here is the summary:

OP states that he does not want to move to newer version of g++
because that would require him to re-verify that all of his stuff
still works and all of the 3rd party stuff still works as well.

You ignore OP's statement and state that he should upgrade
nonetheless. The reason for your confidence is that *you* have test
suites that pass cleanly with the newer version(s).

However we don't know whether you test on the same platforms, with
the same optimizations (don't tell me you test all permutations of
the various -f... flags, I would not believe you :), and whether
your code looks anything like his.

Further you state that you don't have any external dependencies
(this turns out not to be entirely true in the end) while he does
have them.

I don't think your argument is very convincing. Do you?

OTOH, I do see that *you* are in good position to move to whichever
version you desire (provided you keep your test suite updated and
covering all the new code you wrote yesterday).

Cheers,
-- 
In order to understand recursion you must first understand recursion.
Remove /-nsp/ for email.