crash while testing imread.m

[John W. Eaton]

On 25-Feb-2010, Ben Abbott wrote:

| I encountered a new failure building today (its been about 2-3 weeks since my 
last build).
| 
|   scripts/image/imread.m .................................Assertion failed: 
(semaphore_info != (SemaphoreInfo *) NULL), function LockSemaphoreInfo, file 
magick/semaphore.c, line 525.
| 
| $ hg tip
| changeset:   10353:7ed1f2e831ba
| tag:         tip
| user:        John W. Eaton <address@hidden>
| date:        Wed Feb 24 16:53:46 2010 -0500
| summary:     qr.cc: increase tolerance for test
| 
| >From gdb ...
| 
| Assertion failed: (semaphore_info != (SemaphoreInfo *) NULL), function 
LockSemaphoreInfo, file magick/semaphore.c, line 525.
| 
| Program received signal SIGABRT, Aborted.
| 0x94227732 in __kill ()
| (gdb) bt
| #0  0x94227732 in __kill ()
| #1  0x94227724 in kill$UNIX2003 ()
| #2  0x942ba98d in raise ()
| #3  0x942d0a44 in abort ()
| #4  0x942bd5ec in __assert_rtn ()
| #5  0x1678a36c in LockSemaphoreInfo ()
| #6  0x1675ace7 in GetMagickInfo ()
| #7  0x16757a7a in SetImageInfo ()
| #8  0x1672fd83 in ReadImage ()
| #9  0x1660e115 in Magick::readImages<std::vector<Magick::Image, 
std::allocator<Magick::Image> > > (sequence_=0xbffe7f6c, address@hidden) at 
STL.h:2301

It might be related to the problem discussed in this thread:

  
https://www-old.cae.wisc.edu/pipermail/octave-maintainers/2010-January/014891.html

It is still not clear to me whether it is safe to call the init
function from inside the dynamically linked __magick_read__ function.
I would prefer to keep the init function in the dynamically linked
function if possible so that the GraphicsMagick libraries would not
have to be linked with Octave unless they are actually needed.

jwe