Re: cons development

[Dmitry Sagalovskiy]

Thanks for the reply. I should clarify that I am also a grateful user of
cons. I work at a small financial company, where we've been using cons
under FreeBSD ever since I discovered it while looking on Google for
"alternative to make". And we've been very happy with it overall, except
for an occasional missing feature. I've written a module with new commands
that we happen to need, but I am talking about features that are really
part of the core rather than extensions. For example:

  The -n flag (or something similar to show how much stuff is about
  to get rebuilt), which I am sure there's been plenty of discussion about,
  and I know it's not easy because I've thought about it too.

  Better control over caching. We use cache a lot, and every once in a blue
  moon a file would get built incorrectly (or a hard-linked version of it
  would get edited by someone inadvertently), and then it's very hard to
  get the right version into the cache without erasing the whole cache
  (which we hate doing).

Then there are other smaller and bigger things that we haven't given much
thought to, but which would be immensely useful. E.g. parallelizing (I
know there's been other work on it, but I don't know whether there is
anything stable or well-supported).

We don't have anyone who has time to spend looking at cons internals when
it seems like something is not working right, or when we want something
which seems possible but doesn't seem to exist. So we want it to be
actively supported :) You know, with a regularly updated webpage, bug
fixes, improvements and what-not. There might not be a whole lot of
changes, because it's damn good as it is, but there are certainly some
things that would happen if there was active support. (Even little things
like making those Cons::Plus and ConsExtensions modules available from the
main cons page, if they are useful to many cons users.)

In fact, my company is so interested in having active support, that we'd
be willing to pay for it, except that no one here has any idea how it can
be made to work, with free software like this.

If cons is not being actively worked on because SCons is about to become
strictly better anyway, then maybe it's fine to abandon it (although I
know Perl already and don't feel like learning Python just to re-do our
build system). But if people are going to keep on using cons, then it
would be a shame to let a good thing just get left behind and forgotten.

Any thoughts?

Dmitry


On Fri, 20 Dec 2002 address@hidden wrote:

> In the Windows/DevStudio environment, we are using Cons 2.3.0, with
> Cons::Plus and a rewrite of ConsExtensions that we call Cons::DSProject
> with great success.  We have not worked out all the funny little
> dependency issues that we are still finding that go along with Windows
> development, but our builds have never been faster or more reliable.  Real
> incremental building on a nightly basis.  For the first time it is now
> practical for every developer to do complete builds on a regular basis.
>
> I think I can safely say for all of us at Philips Medical, Bothell, Wa,
> thanks to creators of Cons, Cons::Plus, and ConsExtensions.
>
> Rick Croote
> Software Engineer
> Philips Medical Systems - Ultrasound
> address@hidden
>
>
>
>
>
>
>
>
>
>
>
> Dmitry Sagalovskiy <address@hidden>
> Sent by:
> address@hidden
> 12/20/2002 10:09 AM
> Please respond to Dmitry Sagalovskiy
>
>
>         To:     address@hidden
>         cc:     (bcc: Rick Croote/ATL-BTL/MS/PHILIPS)
>         Subject:        cons development
>         Classification:
>
>
>
> What's the status of cons development? Is anyone working on it at all?
> Would anyone be willing to add or improve certain features in cons if
> there is interest among cons users? What kind of financial support would
> the cons project need to move forward?
>
> Dmitry
>
>
>
> _______________________________________________
> address@hidden
> http://mail.gnu.org/mailman/listinfo/cons-discuss
> Cons URL: http://www.dsmit.com/cons/
>
>
>