From: bill_kempf (williamkempf_at_[hidden])
Date: 2002-07-22 09:31:23
--- In jamboost_at_y..., "David Abrahams" <david.abrahams_at_r...> wrote:
> From: "Vladimir Prus" <ghost_at_c...>
> > Further, we have a *lot* of other places where classes are used.
> > each virtual target is a class. We'd have the same performance
> > them -- finding the generator would require creating a lot of
> > targets.
> Yep, we're running up against the language limitations.
> I sometimes wonder if Python packaged with something like py2exe
> been acceptable for Boost.
Since I was one of the people arguing strongly against Python at the
time, I think I should at least comment here ;).
>From a personal stand point, I'm regretting the stance I took back
then. The capabilities present in a system like ScCons are very
appealing. Meld some of the Boost.Build stuff onto that system (for
instance, toolsets and variants) and I'd feel like I was in build
nirvana. And since we've had some changes in my corporation it's a
little easier to convince the powers that be to allow the
installation of such a toolset. :(
But, despite my regrets, I still think the arguments I made back then
are valid. Putting on my thinking hat from those days and using it
to evaluate py2exe, I'm not sure even that would have worked for me.
If the resultant executable were small enough I might have been able
to sneak it in "through the back door" by placing it under version
control and having the build process pull it out and run it locally.
If I'd been caught doing this I'd have had to justify it, which would
have been a pain, but I might personally have taken that route. I
think others might not have been able to do even this, though. (And
yes, I know this is stupid, because boot strapping Jam today is
hardly any different, but we live in the world we live in.)
What really stinks about this, though, is that a small number of
people (such as myself) seem to be holding Boost back from having a
portable, usable and robust build system (not that the Jam stuff
can't ever achieve this, or that it's far from it today, but I think
you know what I mean). :(
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk