Glad to see there is at least someone
still out there. Here at Philips Medical the latest Cons 2.3.0 is
and has been in heavy use now for years and has been very helpful. There
is a push to go to SCons, but the cost of the new language, Python, is
prohibitive for us in the that all of our infrastructure is in Perl. We
are converting a lot of infrastructure to C# and that is a big enough job,
so we are not willing to take on Python and add to the confusion. I
also do not see C# as being able to do what Perl can do, such as eval(),
so I see Perl sticking around for a very long time when it comes to builds.
As for Perl OO, it is very easy, just
keep it simple. I hate to say it, but Cons did it the hard way and
I would like to see a rewrite throwing out most of the functionality and
keep it just a simple dependency scan and build engine with some better
plug and play build specific features. It should be totally version
control dumb. One of these days I'm going to do it!!
I cannot say enough about how great
Cons is, even if I hate going into the source, it performs very fast, beyond
my expectations and very accurately. One big draw is the SCons multiple
process building, and pcons attempted that, I do not know how well, but
if I rewrite Cons it will have that for sure. Oh, and I should mention
that we are using Cons with Visual Studio on a Windows environment, using
ClearCase for version control. Our source is several million lines
of code using multiple brands of compilers. Definitely a better need
for plug-in modules for different types of tools, as the scanning requirements
are all different.
Environment and Tools Team
Philips Medical Systems
Pierre THIERRY <address@hidden>
2006-11-28 06:07 PM
Re: CONS specs update?
Scribit Pierre THIERRY dies 27/10/2004 hora 03:05:
> > I'd hate to see yet another effort to revive Cons drop flat.
> So do I. Sorry for the very long silence, I didn't find time to sit
> a table with a notepad, and think about how everything the discussed
> about could fit together. I think I was lacking inspiration to do
> synthesis of that...
Well, sorry for the next long silence. I must tell that I finally won't
try to make cons evolve. Mainly because I see pretty clearly in my head
how to do much of the work in an object-oriented way, and because I
never really understood Perl's object oriented mechanisms.
I still plan to develop a build management tool with the design I
sketched previously, but probably in Common Lisp. When it works, I'll
probably write the Construct module anyway, so it should be possible to
use existing Construct files with that tool, and use Perl around it's
If I have any success in my attempt, I'll post some note here.
[attachment "signature.asc" deleted by Rick Croote/ATL-BTL/MS/PHILIPS]
Cons URL: http://www.dsmit.com/cons/