If you unzip this and change to Retimer/UnitTests/ptest1 the call eclipse nodbgcomp (just do EXACTLY what I have been doing) ['../../java_top']. test(start). This runs for several minutes and reproducibly you get the following: (lots of stuff) Information: LP Count: 564 Information: LP Count: 565 Information: LP Count: 566 Information: LP Count: 567 Information: LP Count: 568 Information: LP Count: 569 ?140 Warning: Basis lost - recovering. Information: LP Success: 569 Information: LP Count: 570 Information: LP Success: 570 Information: At level: 48 No. of conflicts:4. After filtering: 2 Information: LP Count: 571 Information: LP Count: 572 Information: LP Count: 573 Information: LP Count: 574 Information: LP Count: 575 Information: LP Count: 576 Information: LP Count: 577 Information: LP Count: 578 Information: LP Count: 579 Segmentation violation - possible reasons are: - a faulty external C function - certain operations on circular terms - machine stack overflow - an internal error in ECLiPSe Aborting execution.... *** ECLiPSe fatal error: reset/0 called Signal 10 Aborting execution.... *** ECLiPSe fatal error: reset/0 called Signal 10 Aborting execution.... *** ECLiPSe fatal error: reset/0 called Signal 10 *** ECLiPSe fatal error: reset/0 called Signal 10 Aborting execution.... *** ECLiPSe fatal error: reset/0 called Signal 10 Aborting execution.... *** ECLiPSe fatal error: reset/0 called Signal 10 Aborting execution.... *** ECLiPSe fatal error: reset/0 called Signal 10 Aborting execution.... *** ECLiPSe fatal error: reset/0 called Signal 10 Aborting execution.... *** ECLiPSe fatal error: reset/0 called Signal 10 Aborting execution.... *** ECLiPSe fatal error: reset/0 called system interface error: Too many open files in open("/usr/local/eclipse/5.0/lib/version.pl", read, _628) ===================================== There are 2 bugs here, one is the bug that I believe to be in XPRESS the second and not very urgent is that it's probably a good idea to do as little as possible in the fatal error handler, and to read the version information at the start of eclipse There is an almost identical behavior when you run this on windows. Changing stack sizes seems to have no effect on this bug. Can we run it on the intsrumented version of eplex ? Thank you Stefano 2nd Floor The Tower Building EMAIL: Stefano.Novello@parc-technologies.com 11 York Road TEL: +44 20 72614017 SE1 7NX LONDON FAX: +44 20 72614001
This archive was generated by hypermail 2.1.8 : Wed 16 Nov 2005 06:08:01 PM GMT GMT