On 3/30/12 8:07 PM, Faré wrote:
On Fri, Mar 30, 2012 at 22:21, Raymond Toy <toy.raymond@gmail.com> wrote:
On 3/30/12 4:24 PM, Faré wrote:
Same here on Linux x84. Everything works great with CMUCL 20b, but 20c fails with these low-level failures. I suspect a bug in CMUCL itself. How did you run this test? I did make lisp=cmucl CMUCL=<20c> test-lisp.
No problems with 20c or latest snapshot on OSX. (Unicode builds.)
But I see you ran this on a linux box. I'll to check that out too....
BTW, the gmane list for cmucl devel has been broken for quite some time. It receives messages, but you can't send from there. I don't know why.
Yes, you indeed test with make lisp=cmucl CMUCL=$cmuclbinary test-lisp
Can't reproduce this with the 2012-03 snapshot on my Linux (openSuSE 64bit) box. Do you have any stale fasls that are getting loaded into cmucl on startup?
Using the 20c release tarball (since I fail to compile the latest from git),
I don't recall anything special being needed to compile the latest git sources with 20c. If you can remember, please let me know. Ray