Keith Packard writes:
Around 15 o'clock on Feb 19, Kurt J. Lidl wrote:
You ought to make the xf86jmp_buf larger than 200 bytes.
I have access to a machine where the jmp_buf storage takes 232
bytes already!
I made it 400 bytes -- Linux x86 uses 156 bytes, and 256 seemed likely to
On Thu, 20 Feb 2003, Egbert Eich wrote:
Guys, please bear with me: why do we need this typedef for
xf86jmp_buf at all? This buffer is normally allocated with
something like:
buf = malloc(sizeof(jmp_buf));
When setjmp.h is included this size is known at compile time.
We'd just have to
On Wed, Feb 19, 2003 at 10:30:53PM -0800, Keith Packard wrote:
Around 23 o'clock on Feb 19, David Dawes wrote:
Yes, I think that should work (providing we don't get too much other
stuff included as a result of including setjmp.h directly).
Is there any interest in having this tested before a
Dr Andrew C Aitchison writes:
On Thu, 20 Feb 2003, Egbert Eich wrote:
I may not know what I'm talking about.
I thought that the problem was that modules are supposed to work
across operating system and compiler, so the compile-time size
might not be big enough when the module is
Around 18 o'clock on Feb 20, Egbert Eich wrote:
However this doesn't work too well in freetype. jmp_bufs are
part of larger structs. Converting this not totally trivial.
I think it's more practical to simply make the jmp_buf big enough;
forcing the library to restructure itself to permit
On Wed, 19 Feb 2003, Keith Packard wrote:
Yes, they are rather system specific, but FreeType2 uses setjmp and longjmp
extensively for error recovery. Disallowing setjmp and longjmp would
make using FreeType2 rather difficult.
I believe the approach taken will at least work in the majority
Around 15 o'clock on Feb 19, Kurt J. Lidl wrote:
You ought to make the xf86jmp_buf larger than 200 bytes.
I have access to a machine where the jmp_buf storage takes 232
bytes already!
I made it 400 bytes -- Linux x86 uses 156 bytes, and 256 seemed likely to
be problematic on some machines.
Around 15 o'clock on Feb 19, Stuart Anderson wrote:
It seems like the wrapped versions could be extended so that they would
work correctly.
That's not possible. The fundemental problem is that setjmp captures the
execution context (more or less the continuation) from its caller,
including the
Keith Packard wrote (in a message from Wednesday 19)
Around 15 o'clock on Feb 19, Stuart Anderson wrote:
This approach strikes me as being inherently non-portable wrt the module ABI.
setjmp/longjmp are too system specific to be used in modules.
Yes, they are rather system specific,
On Wed, Feb 19, 2003 at 12:20:04PM -0800, Keith Packard wrote:
Around 15 o'clock on Feb 19, Stuart Anderson wrote:
This approach strikes me as being inherently non-portable wrt the module ABI.
setjmp/longjmp are too system specific to be used in modules.
Yes, they are rather system specific,
Mike has reported to me a crash with the new FreeType backend and the
Caslon fonts included with SuSE. (Mike, could you be so kind as to
follow up with a stable URL for the fonts?)
I'm stuck. The problem doesn't happen with xfs, and it doesn't happen
with a monolithic server. So it must be in
MF Caslon fonts (same problem):
Hold on... I'm seeing a crash with the Caslon fonts. Are you seeing
something else?
Juliusz
___
Devel mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/devel
Around 14 o'clock on Feb 18, Mike FABIAN wrote:
On my machine there is no crash neither with the Caslon fonts nor with
the Omegaserif fonts, only an error message when trying to open the
font:
The version of TrueType in XFree86 is catching an error inside the Unicode
cmap tables when
13 matches
Mail list logo
