Re: [boost] filesystem library const usage

From: Daniel Frey (daniel.frey_at_[hidden])
Date: 2002-08-01 07:20:35

John Maddock wrote:
> 
> > and
> > John (who provided the fixes) wasn't interested in maintaining them. He
> > thinks (and I agree with him), that the changes are OK and we should
> > wait if some Borland users complain.
> 
> I didn't mean I wasn't interested, just that the fixes were very strange,
> very specific to a particular situation, and therefore there wasn't any
> action necessary until there was some real code available to test against
> that compiler, if you post your final code I'm sure that someone will test
> it...

Sorry, I didn't wanted to misinterpreted your statements. Anyway, there
is code available to test, IMHO it should be the final version. The
problem is, that it is not seen as the final version by other people as
it hasn't been tested on enough platform. As it is not the final version
for others, no one seems to bother testing it. I'm simply stuck here and
I have no clue what to do. And no, I will not start buying Windows,
VC++, Borland, ... just to satisfy the regression tests for boost.
Sometimes I am not sure what you (the group, not you as an individual)
expect people to do in order to let "fixes" and "small enhancements" go
into boost. I mean, this is not a new library or an amazing new
programming technique, it's just a tiny improvement and I never thought
it could take so long. Or is this a test for me? To see if I am willing
to take over the maintainance of it for a longer time? If so, I wonder
why the people that "maintain" the libraries don't take the advice for
the NRVO (and constness for return value) and maintain their libs...

Regards, Daniel

--
Daniel Frey
aixigo AG - financial training, research and technology
Schloß-Rahe-Straße 15, 52072 Aachen, Germany
fon: [phone], fax: [phone]
eMail: daniel.frey_at_[hidden], web: http://www.aixigo.de