headers
SXEmacs Issues | 5 Jun 2007 23:57

[Bug 27] New: define-abbrev-table triggers a malformed-property-list error

http://issues.sxemacs.org/show_bug.cgi?id=27

           Summary: define-abbrev-table triggers a malformed-property-list
                    error
           Product: SXEmacs
           Version: 22.1.7
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core Lisp
        AssignedTo: hroptatyr <at> sxemacs.org
        ReportedBy: steve <at> sxemacs.org
         QAContact: sxemacs-devel <at> sxemacs.org

Created an attachment (id=36)
 --> (http://issues.sxemacs.org/attachment.cgi?id=36)
lisp backtrace

I'm seeing some breakage in my abbrevs.  This bug can easily be reproduced in a
-vanilla instance...

(define-abbrev-table 'my-abbrev
  '(("foo" "expanded foo" nil 0)
    ("bar" "expanded bar" nil 0)))

And eval it _twice_.  On the 2nd and subsequent evals, the error happens. 
Backtrace attached.
(Continue reading)

Permalink | Reply | 5 comments |
headers
SXEmacs Issues | 7 Jun 2007 11:10

[Bug 25] xetla-commit causes SIGABRT (more gc madness?)

http://issues.sxemacs.org/show_bug.cgi?id=25

Sebastian Freundt <hroptatyr <at> sxemacs.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED

--

-- 
Configure bugmail: http://issues.sxemacs.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
Permalink | Reply | 0 comments |
headers
SXEmacs Issues | 7 Jun 2007 11:11

[Bug 27] define-abbrev-table triggers a malformed-property-list error

http://issues.sxemacs.org/show_bug.cgi?id=27

Sebastian Freundt <hroptatyr <at> sxemacs.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED

--

-- 
Configure bugmail: http://issues.sxemacs.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
Permalink | Reply | 0 comments |
headers
Steve Youngs | 7 Jun 2007 17:04
X-Face
Face
Gravatar

Re: Come take a ride in the De Lorean -- 22.1.7 is here!

* Sebastian Freundt <hroptatyr <at> sxemacs.org> writes:

  > Steve Youngs <steve <at> sxemacs.org> writes:
  > Again the entire list:

The De Lorean is almost outta fuel, time to revisit the todo...

  >> o Hugely better colour support on a tty. (Nelson is already
  >> well on the way to having this done)

Nelson has this pretty much licked now, and in fact, its inclusion into
mainline will herald 22.1.7's feature freeze.

  >> o Boehm-gc. (another of Nelson's babies)

It's too late in the game for this now, deferred until 22.1.8.

  >> o etags.el is getting an overhaul (me)
  >> o sppm.texi needs updating (me)
  >> o Weak tests (me)

My system rebuild pretty much meant that I got to none of these.  All
deferred until 22.1.8.

  >> o custom.el in XE 21.5 has changed quite a bit over what we
  >> have.  We should sync up. (your name here)

No volunteers, ok I will put it on my todo, but it won't happen until
22.1.8.
(Continue reading)

Permalink | Reply | 0 comments |
headers
SXEmacs Issues | 7 Jun 2007 17:10

[Bug 23] Skiplist Tests fail with signal 10 on UltraSPARC platform

http://issues.sxemacs.org/show_bug.cgi?id=23

Steve Youngs <steve <at> sxemacs.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #8 from Steve Youngs <steve <at> sxemacs.org>  2007-06-08 01:09:56 ---
Horst tells me that this is now fixed.

--

-- 
Configure bugmail: http://issues.sxemacs.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
Permalink | Reply | 0 comments |
headers
SXEmacs Issues | 7 Jun 2007 17:18

[Bug 23] Skiplist Tests fail with signal 10 on UltraSPARC platform

http://issues.sxemacs.org/show_bug.cgi?id=23

Steve Youngs <steve <at> sxemacs.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #9 from Steve Youngs <steve <at> sxemacs.org>  2007-06-08 01:17:54 ---
And the build-report to confirm it is...

  Message-ID: <Pine.LNX.4.64.0706080038440.23868 <at> horst-tla.homeunix.net>

--

-- 
Configure bugmail: http://issues.sxemacs.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
Permalink | Reply | 0 comments |
headers
Steve Youngs | 9 Jun 2007 04:34
X-Face
Face
Gravatar

Re: [Tests fail] steve <at> sxemacs.org--2007/sxemacs--main--22.1.7--patch-81 on x86_64-unknown-linux-gnu

* Horst Günther Burkhardt <horst <at> sxemacs.org> writes:

  > ../configure  '--with-ase'

[...]

  >> Contents of /home/burkhardth/sxemacs--22.1.7/=build/,,make-check.out
  > make  check-am
  > Loaded module ase_heap v0.0.0 (SXEmacs module: ase-heap)Fatal error:
  > assertion failed, file ../../src/alloc.c, line 298, block !=
  > (void*)0xCAFEBABEDEADBEEF
  > make[3]: *** [check] Error 2
  > make[2]: *** [check-recursive] Error 1
  > make[1]: *** [check-recursive] Error 1

Coincidence?

Horst, can you please try --without-ase and see how that goes.

Sebastian, heads up man on this one, buddy.

--

-- 
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
|       SXEmacs - The only _______ you'll ever need.       |
|         Fill in the blank, yes, it's THAT good!          |
|------------------------------------<steve <at> sxemacs.org>---|
Permalink | Reply | 0 comments |
headers
Steve Youngs | 9 Jun 2007 16:30
X-Face
Face
Gravatar

Re: [Tests fail] steve <at> sxemacs.org--2007/sxemacs--main--22.1.7--patch-84 on x86_64-unknown-linux-gnu

* Horst Günther Burkhardt <horst <at> sxemacs.org> writes:

  > ../configure  '--without-ase'

[...]

  >> Contents of /home/burkhardth/sxemacs--22.1.7/=build/,,make-check.out
  > make  check-am
  > Loaded module ase_heap v0.0.0 (SXEmacs module: ase-heap)Fatal error:
  > assertion failed, file ../../src/alloc.c, line 298, block !=
  > (void*)0xCAFEBABEDEADBEEF
  > make[3]: *** [check] Error 2
  > make[2]: *** [check-recursive] Error 1
  > make[1]: *** [check-recursive] Error 1

OK, so that didn't work.  Can you please try forcing a 32bit build and
see how that goes...

  CFLAGS='-g -O2 -m32' ../configure RET

Sebastian, are you seeing anything like this on any of your builds?

--

-- 
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
|       SXEmacs - The only _______ you'll ever need.       |
|         Fill in the blank, yes, it's THAT good!          |
|------------------------------------<steve <at> sxemacs.org>---|
Permalink | Reply | 1 comment |
headers
Sebastian Freundt | 9 Jun 2007 19:30
X-Face
Face
Gravatar

Re: [Tests fail] steve <at> sxemacs.org--2007/sxemacs--main--22.1.7--patch-84 on x86_64-unknown-linux-gnu


Steve Youngs <steve <at> sxemacs.org> writes:

> * Horst Günther Burkhardt <horst <at> sxemacs.org> writes:
>
>   > ../configure  '--without-ase'
>
> [...]
>
>   >> Contents of /home/burkhardth/sxemacs--22.1.7/=build/,,make-check.out
>   > make  check-am
>   > Loaded module ase_heap v0.0.0 (SXEmacs module: ase-heap)Fatal error:
>   > assertion failed, file ../../src/alloc.c, line 298, block !=
>   > (void*)0xCAFEBABEDEADBEEF
>   > make[3]: *** [check] Error 2
>   > make[2]: *** [check-recursive] Error 1
>   > make[1]: *** [check-recursive] Error 1
>
> OK, so that didn't work.  Can you please try forcing a 32bit build and
> see how that goes...
>
>   CFLAGS='-g -O2 -m32' ../configure RET
>
> Sebastian, are you seeing anything like this on any of your builds?

Nope. Also see my latest build reports.
I thing the problem here isn't ase itself, it's just the first of all them
tests hence you see it there.
To make sure, Horst, could you provide a --with-modules=none or
--with-modules=noase build rpt?
(Continue reading)

Permalink | Reply | 0 comments |
headers
Sebastian Freundt | 10 Jun 2007 15:12
X-Face
Face
Gravatar

Re: [Tests fail] steve <at> sxemacs.org--2007/sxemacs--main--22.1.7--patch-84 on x86_64-unknown-linux-gnu


Horst Günther Burkhardt III <horst <at> sxemacs.org> writes:

[snip]
> Testing
> /home/burkhardth/sxemacs--22.1.7/tests/automated/skiplist-tests.el...Fatal
> error: assertion failed, file ../../src/alloc.c, line 298, block !=
> (void*)0xCAFEBABEDEADBEEF
> make[2]: *** [check] Error 2
> make[1]: *** [check-recursive] Error 1

Just my suspicion! That's a general thing, more or less.
You could now build with --with-modules=none and wont see the failure in
ase nor cl, but instead in the next test which is after that.
Something's wrong with either the malloc (allocation fails already) or the
marking phase of our GC: objects are freed (and assigned that
0xCAFEBABEBLABLA) DESPITE they can be seen.

You shouldn't be able to use that SXE anyway, the same effect should occur
in a `normally' run session (can you confirm?), the first or second
garbage collection will trigger it (M-x garbage-collect).

Sebastian
Permalink | Reply | 1 comment |

Gmane