From: Stefan Seefeld (seefeld_at_[hidden])
Date: 2007-04-01 21:18:59
David Abrahams wrote:
> on Sun Apr 01 2007, Stefan Seefeld <seefeld-AT-sympatico.ca> wrote:
>> David Abrahams wrote:
>>> on Sun Apr 01 2007, "MÃ¥rten RÃ¥nge" <marten.range-AT-gmail.com> wrote:
>> Whether boost is 'monolithic' or a 'library collection' seems to me
>> a rather academic question, at least as far as boost users are
>> concerned, as they can't download and install individual libraries.
> They can only download the whole thing at once, but as for
> installation, if they want a subset of headers, what's wrong with bcp?
> If they want a subset of binaries, what's wrong with our
> bjam/configure options?
How many boost users are actually building boost themselves, compared
to those who use prepackaged 'system' libraries ? (And from those who
do build themselves, how many are sufficiently familiar with the infrastructure
to know how to use these fine tools well (compared to those who are very happy
to just be able to get the whole thing up and running so they don't need to
touch it ever again) ?
>> As far as the release process is concerned, it would certainly help if
>> the regression testing would be at least somewhat modular.
>> Having test runs involve not only all libraries, but also all build variants
>> and all (available) toolchains is a very significant slowdown.
> Unfortunately, I don't believe all build variants _are_ tested. But
> anyway, I don't understand what you mean about modularity and
> slowdowns. The tests are distributed across many machines. If you
> dedicate a single testing machine to one toolchain and build variant,
> you can't do this much faster. Many testers do incremental testing,
> so only the changed stuff gets rebuilt.
You are right, the total execution time is still the same. However,
a more modular system is more easily parallelizable. More testers
could contribute cycles as the resource requirements wouldn't be quite as
high. This, then, makes the cycles from checkin to report containing
associated test results smaller, helping to get fixes in quicker. Etc. etc.
> Anyway, I hope after 1.34 ships we can get to work on a system for
> testing much more effectively. The current one is definitely showing
> its age.
That would be good indeed.
-- ...ich hab' noch einen Koffer in Berlin...
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk