From: Douglas Gregor (gregod_at_[hidden])
Date: 2002-11-09 16:27:16
On Saturday 09 November 2002 02:22 pm, Victor A. Wagner, Jr. wrote:
> Maybe it's my expectations that are out of whack.
> Are there any rules for what's "commit"ed to the CVS main branch?
There are no rules. The CVS trunk (main branch) is for Boost development. We
try to keep it stable (because many users and developers seem to work out of
CVS), but sometimes things break.
> things I've found:
> 1) the documentation main page still has ???, 2002 - Version 1.29.0
> on it. Is this really the latest? how can we tell??
The CVS trunk is the latest (unreleased, experimental) code.
> 2) regex doesn't compile on vc7
Please report any errors either to the list or to the library author directly
so that they can be fixed.
> 3) function causes errors throughout boost (and _really_ unfortunately in
> my projects for work which make use of multi_array)
This is also for VC7? If so, it's fixed now.
> So, I guess the question I really have is:
> Are there any expectations of stability for the mainline?
> If so, what are they?
The mainline is generally stable. However, if you are using a compiler that
often requires workarounds, the incidence rate of brokenness will probably
increase. (For instance, I don't have access to VC7, so I have to wait until
someone complains or I get regression test results before I can fix a
> else what expectations do the developers have that anyone will bother
> testing any of their changes?
If it's in a regression test, there are usually results available within a
day. Obviously we like to see users and other developers using/testing the
development versions of libraries so that we can get feedback sooner.
> am I wasting my time attempting to build and use the mainline CVS branch
> every day?
That depends: if you report any problems you find when trying to use Boost
from CVS, so that we can fix the problems, then you're helping Boost become a
better library and I don't believe that's a waste of time. If you'd rather
not deal with submitting bug reports, then perhaps it is a wste of time for
you to use Boost from CVS, but we appreciate the help navigating the
minefield of compiler brokenness.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk