|
From: Bruno H. <br...@cl...> - 2005-06-03 19:01:28
|
Hi,
Sam and I want to prepare a new release. For this reason, we now enter a
half code-freeze. At this point,
- No changes of visible behaviour of clisp, except to fix regressions
(vs. 2.33.2) and bugs.
The only commits that go in at this stage are:
- fixes and implementation improvements for features that have been added
since 2.33.2,
- fixes for things that are unanimously recognized as bugs,
- any kind of doc changes,
- any kind of changes to the release infrastructure.
Bruno
|
|
From: Will N. <wi...@mi...> - 2005-06-03 19:06:17
|
On Friday 03 June 2005 20:00, Bruno Haible wrote: > Hi, > > Sam and I want to prepare a new release. For this reason, we now enter a > half code-freeze. At this point, I have some fixes for the FFI on mips that I posted a while ago. Would these be a possible candidate for the next release? I need to get them synced with the current CVS, but I suspect that Debian is the only person building on mips at the moment. |
|
From: Bruno H. <br...@cl...> - 2005-06-03 20:50:17
|
Will Newton wrote:
> I have some fixes for the FFI on mips that I posted a while ago. Would
> these be a possible candidate for the next release?
Thanks for the fixes. But more work needs to be done on them to ensure that
they will not kill the support of big-endian IRIX/mips systems. I won't
apply them as-is, the risk of breaking IRIX support (or MIPSEB in general)
is just too high. The right/safe way to do this is
1. #ifdefize between MIPSEB and MIPSEL in avcall.h.in,
2. #ifdefize between MIPSEB and MIPSEL in avcall-mips.c,
3. create different avcall-mips-eb.s and avcall-mips-el.s which two
appropriately built gcc compilers.
> I need to get them synced with the current CVS, but I suspect that Debian
> is the only person building on mips at the moment.
And even in Debian, I think, ports like mips have been downgraded to
"second class citizens" a few weeks ago.
Bruno
|
|
From: Will N. <wi...@mi...> - 2005-06-03 21:13:27
|
On Friday 03 June 2005 21:49, Bruno Haible wrote: > Thanks for the fixes. But more work needs to be done on them to ensure that > they will not kill the support of big-endian IRIX/mips systems. I won't > apply them as-is, the risk of breaking IRIX support (or MIPSEB in general) > is just too high. The right/safe way to do this is > 1. #ifdefize between MIPSEB and MIPSEL in avcall.h.in, > 2. #ifdefize between MIPSEB and MIPSEL in avcall-mips.c, > 3. create different avcall-mips-eb.s and avcall-mips-el.s which two > appropriately built gcc compilers. Debian builds Linux mips(eb) and mipsel, so the test suite runs OK on both of those architectures. I don't have access to any IRIX machine to test these changes however. Are there any IRIX users on this list? I will ask on clisp-list too, although the archives don't turn up any IRIX questions. Hmm, it seems SGI may provide access to a development system. I will look into that if there are no volunteers. > > I need to get them synced with the current CVS, but I suspect that Debian > > is the only person building on mips at the moment. > > And even in Debian, I think, ports like mips have been downgraded to > "second class citizens" a few weeks ago. That proposal seems to have run into a lot of opposition from developers, mips should be around in Debian for the next few years at least. |
|
From: Sam S. <sd...@gn...> - 2005-06-06 21:34:26
|
> * Bruno Haible <oe...@py...> [2005-06-03 21:00:27 +0200]:
>
> - fixes and implementation improvements for features that have been
> added since 2.33.2,
is it possible to do something about these warnings?
;; Compiling file /cygdrive/d/gnu/GBBopen/source/gbbopen/units.lisp ...
WARNING: Redefining metaobject class #<STANDARD-CLASS STANDARD-UNIT-CLASS> has
no effect.
;; Wrote file /cygdrive/d/gnu/GBBopen/clisp-2.33.83/gbbopen/units.fas
;; Loading file /cygdrive/d/gnu/GBBopen/clisp-2.33.83/gbbopen/units.fas ...
WARNING: Redefining metaobject class #<STANDARD-CLASS STANDARD-UNIT-CLASS> has
no effect.
WARNING: Redefining metaobject class #<STANDARD-CLASS DIRECT-LINK-DEFINITION>
has no effect.
WARNING: Redefining metaobject class
#<STANDARD-CLASS GBBOPEN-EFFECTIVE-SLOT-DEFINITION>
has no effect.
WARNING: Redefining metaobject class
#<STANDARD-CLASS EFFECTIVE-NONLINK-SLOT-DEFINITION>
has no effect.
WARNING: Redefining metaobject class
#<STANDARD-CLASS EFFECTIVE-LINK-DEFINITION>
has no effect.
;; Loaded file /cygdrive/d/gnu/GBBopen/clisp-2.33.83/gbbopen/units.fas
--
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://www.jihadwatch.org/> <http://ffii.org/> <http://www.iris.org.il>
<http://www.palestinefacts.org/> <http://www.dhimmi.com/>
A year spent in artificial intelligence is enough to make one believe in God.
|