on 2/12/06 1:49 PM, Joe Koski at email@hidden wrote:
After reading your e-mail, I decided that the make file should work
and without changes, so I exported an ORGANIZATION= (just in case)
make, which completed without errors, and did "make check" again.
Now all my
errors are in one test routine called "socket" with 32 errors out
tests. All of several thousand other "make check" tests are OK.
like I need to ask the clisp folks about what's happening. I'll
only if it is Mac OS relevant.
The clisp-2.38 folks were aware of the "socket" problem, which they
because "Older BSDs (specifically, Mac OS X 10.4.3) require that
filled with 0 before use (i.e., unused fields must be 0)". They have
in case anyone wants to try it with OS X 10.4.x. Their patch works
OS X 10.3.9. My maxima-5.9.2, which is why I needed clisp, also
FWIW, the fields are *NOT* unused in case of simultaneous initiation
of a TCP connection, per RFC 793, page 31, figure 8. Some TCP state
machines are insufficiently complex to handle this case correctly. I
guess it would be possible to modify the MacOS X state machine to work
in all cases but that particular one, with uninitialized field
contents, but it's probably better to zero the structures.
Do not post admin requests to the list. They will be ignored.
Unix-porting mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden