From: Joe K. <jko...@co...> - 2006-02-13 00:10:56
|
I just joined the list because I saw the same "UNIX ERROR 49", etc. on make check with clisp-2.38 that others are seeing on Macs. The twist is that I still run OS X 10.3.9. I read through the thread to the best of my ability, and found a patch for socket.d. Has this proven to be the solution for this problem? Where is there a fresh copy of the patch? Thanks for any help. Joe |
From: Hoehle, Joerg-C. <Joe...@t-...> - 2006-02-15 15:25:14
|
Joe Koski wrote: >I created a separate test directory and rebuilt clisp-2.38 on >my Mac OS X 10.3.9 with Xcode tools 1.5 (Apple's gcc-3.3). Thank you for your effort and data point. [...] >I hope that it helps. It seems that Sam's socket.d patch >will work on my system, and if it works on 10.4.x, then only >one patch is >required. I don't know why the stream.d patch worked, but it >got me going. If I understand you well, *either* one of a) the past-2.38 CVS patch to stream.d *or* b) Sam's FILL0 patch attachment to socket.d causes the test failure count to go down from 32 to 0!?! Quiet weird at a first glance. Thanks again for reporting this. Do other Mac users (e.g. on OS X 10.4.y) observe similar behaviour? Regards, Jorg Hohle. |