[clisp-list] please help preprare new releases of CLISP

[Hoehle, Joerg-C. <Joe...@t-...>]

Don Cohen writes:
>Now maybe I can try to build a current CVS clisp.
>It would be useful to mark states that are thought to be relatively
>stable and buildable.
Do you mean "mark releases" or snapshots of CVS?

As far as releases, 2.39 seems like a perfect candidate for "stable": When you look at the ChangeLog or NEWS, you see that there have been only bugfixes, except from the MS-windows I/O speed improvements and some SAVEINITMEM change.
http://clisp.cvs.sourceforge.net/*checkout*/clisp/clisp/src/NEWS

Pro release:
 o MacOS non-blocking socket fix - enough reason on its own for a release?
 o sigsegv fixes

However, time is not ripe for a release IMHO.  YMMV:
1. There are several portabiliy bugs open.

2. I don't have access to a compile farm (firewall).  The minimal thing
   to do prior to release is to test CVS whether it builds an passes tests
   on a lot more platforms than my MS-VC6 and Linux/i386 builds (and
   Yaroslav's mingw, and Dave's Fedora Core etc.).

I'd really appreciate if users with access to some compile farm (either sourceforge or whatever is under the Sun) would build and test CLISP now and then on a lot of different architectures and report errors.
Who volunteers?
BTW, providing patches is even better than reporting errors :-)

3. Beside bug-fixes, it would be nice to stabilize recent features.  The one I think about in this context is the January discussion about the SAVEINITMEM executable file command line arguments.  I haven't followed that closely, but it seems to me there were open suggestions about how to improve the current state (if needed?).

Thanks for reporting the broken links, BTW.  Although we could check links ourselves, there are so many things we could do ourselves (given a lot of time) that I really appreciate every little report and especially patches.
Sam Steingold went ahead and did changes to the home page. Please report if there's still something broken or whatever flows to your mind regarding CLISP.

Regards,
	Jorg Hohle.