Just a quick message.  After compiling LyX on a Solaris 2.7 machine
(2.6?  At least, I think it's 2.7 - rats, I never know):
        SunOS 5.7 Generic sun4u sparc SUNW,Ultra-5_10
"Something" crashes whenever a postscript is included into a
document.  Note that the program works "swimmingly" if it's a document
without images, and the old builds I have (1.0.4) work fine (including
images). Now, this may have nothing to do with LyX, but here's the error
output one gets whenever a .ps file is included:

BadAlloc (insufficient resources for operation)
Unrecoverable error: limitcheck in .putdeviceprops

Unexpected interpreter error -13.
Error object: .putdeviceprops(1488)op_array(513)0x169cd0:
Execution stack at 0x15dcac:
0x166b80: 0x0f oper --F---e-- 0x0000 0x00059cf4 = %interp_exit
0x166b88: 0x03 file --G-rxe-- 0x0001 0x0016a378
0x166b90: 0x0e null --F---e-- 0x0000 0x0005b7c4
0x166b98: 0x0b int  --F------ 0x0002 0x00000006 = 6
0x166ba0: 0x0b int  --F------ 0x0000 0x00000005 = 5
0x166ba8: 0x0f oper --F---e-- 0x0000 0x0005b7ac = %oparray_pop
0x166bb0: 0x05 mpry --G-rxe-- 0x0002 0x0016afc4
0x166bb8: 0x0e null --F---e-- 0x0000 0x0005b7c4
0x166bc0: 0x0b int  --F------ 0x0001 0x00000006 = 6
0x166bc8: 0x0b int  --F------ 0x0001 0x00000005 = 5
0x166bd0: 0x0f oper --F---e-- 0x0000 0x0005b7ac = %oparray_pop
0x166bd8: 0x05 mpry --G-rxe-- 0x0039 0x001a07f2
Dictionary stack at 0x15d6e8:
0x1673b0: 0x02 dict --Gwrx--- 0x0000 0x00167520
0x1673b8: 0x02 dict --Gwrx--- 0x0004 0x001b0750
0x1673c0: 0x02 dict --Lwrx--- 0x3620 0x0016e510
0x1673c8: 0x02 dict --Gwrx--- 0x0000 0x00167520
0x1673d0: 0x02 dict --Gwrx--- 0xda2c 0x001e8be8

Actually, most of the above stuff looks like it's from ghostscript
(but - who knows, maybe some of it's from LyX?) - either way, LyX goes
down because of it.
Before blaming anybody, this is also very possibly due to a bad build...

VERSIONS
GNU Ghostscript 5.10 (1998-12-17)
gcc/g++ 2.95.1
libXpm 3.71
xforms 0.88.1


BAD BUILD?
Possibly.  Probably?  Welp, here's what I did - for better or worse...
(Worse, I imagine... hehe)
I had to change a couple of things around a wee bit.  Firstly, the
./configure program didn't see that DIR is defined in dirent.h (so I added
"# define HAVE_DIRENT_H 1" to the config.h file), and typecasted the
second argument of select incorrectly as a pointer to an integer (so I
removed SELECT_TYPE_ARG234 from approximately line 370 of 
src/spellchecker.C).
(Actually, ./configure also missed the location of ar in libtool,
which on this bloody machine is supposed to be /usr/ccs/bin/ar -
gack).
Right.  Well, the next bit of changing is rather embarassing:
including the line "#define _MATH_H 1" at the beginning of the
insets/figinset.C, FontInfo.C and FontLoader.C files so that math.h
isn't included, because the definition of "exception" there clashes
with some other definition of "exception" in some (standard system)
C++ include file (sorry, can't remember which one off the top of my
head).  Sorry.  I didn't look closely at this whatsoever.  I was
hungry and just wanted a quick fix.  :)
Probably the "next worst part" of the build was to include
-I/usr/local/include -I/usr/include to makefiles in the compiling -
sorry again, but if I didn't, the thing just wouldn't compile
(./configure put in -I/usr/openwin/include, which just didn't "cut
it", as it were).

Lastly, keep up the great work!  Looking great, and looking forward to the
new updates!

Reply via email to