From: Rob Stewart (stewart_at_[hidden])
Date: 2004-10-01 13:00:31
From: "Peter Dimov" <pdimov_at_[hidden]>
> Aleksey Gurtovoy wrote:
> > David Abrahams writes:
> >> Aleksey Gurtovoy <agurtovoy_at_[hidden]> writes:
> >>> What should we do about the following Boost.Build toolsets which
> >>> filename's length exceeds 31 characters:
> >>> tools/build/v1/como-win32-4.3.3-vc7.1-tools.jam
> >>> tools/build/v1/intel-win32-7.1-vc6-stlport-4.5.3-tools.jam
> >> Sorry to state the obvious, but... shorten the names?
> > How exactly? Every part of the name carries essential information that
> > cannot be omitted without disabling other possible configurations. For
> > example, the only way we could omit "vc6" in
> > "intel-win32-7.1-vc6-stlport-4.5.3-tools" is if we explicitly decide
> > that we don't care about, let's say,
> > "intel-win32-7.1-vc7-stlport-4.5.3-tools", and nobody ever will.
> > ^^^
> > And so on.
No matter how you compress things, won't there be combinations
that exceed the character limit? We can deal with that when it
happens, of course, but I was thinking that there can be a
dictionary file that maps from the long, complex set of
discriminators to a short name. Then, even if you don't
recognize a particular toolset name, you can look it up in the
dictionary file to learn what it means.
Creating new entries is simply a matter of encoding all of the
relevant details and picking a short name that is unique and,
whenever possible, has some mnemonic correlation to the details.
Duplicate names are possible if folks aren't careful; the
scanning tools could look for and complain about that condition.
-- Rob Stewart stewart_at_[hidden] Software Engineer http://www.sig.com Susquehanna International Group, LLP using std::disclaimer;
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk