Closed npcarter closed 7 years ago
Does not reproduce; works fine for me on ody (Linux) and OS/X, with gcc or icc, with debugging flags or without. Is it reproducible for you, and can you give me more info so I can reproduce?
To run just that test without having to run the whole testsuite:
% cd testsuite
% ./i21-rewind.pl .. .. tmpfoo
The i21-rewind.pl did have an issue that I found (a typo in a commandline path), which I just checked in a fix for, but it was not causing the test to fail.
Hmm, I can’t replicate this in a clean checkout today, either. Two possibilities occur to me: 1) Your path fix corrected the issue I was seeing. Last week, the error I got did look like a path problem, as directly running i21-rewind gave errors about “unable to find hmmsearch in /hmmer/src” or something similar. (When there was an hmmsearch executable in the src directory of my build).
2) I was doing the wrong thing and building h3-develop against easel-master, which caused the problem. I ran into this in my first attempt to reproduce the problem today, since h3-develop now expects there to be a file in easel-develop that isn’t present in easel-master.
-Nick
On Apr 29, 2017, at 8:51 AM, Sean R. Eddy notifications@github.com wrote:
Does not reproduce; works fine for me on ody (Linux) and OS/X, with gcc or icc, with debugging flags or without. Is it reproducible for you, and can you give me more info so I can reproduce?
To run just that test without having to run the whole testsuite:
% cd testsuite % ./i21-rewind.pl .. .. tmpfoo The i21-rewind.pl did have an issue that I found (a typo in a commandline path), which I just checked in a fix for, but it was not causing the test to fail.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/EddyRivasLab/hmmer/issues/124#issuecomment-298167226, or mute the thread https://github.com/notifications/unsubscribe-auth/AEaQ5O6PJZTDxt3yrBOJ96_Esd0h9Fadks5r0zI6gaJpZM4NLprO.
ok, closing.
At least on the Odyssey cluster, test #255 fails when I do a clean checkout of h3-develop, followed by autoconf, configure, make check. I ran into this while testing the fix to the hmmc2 bug I reported, and decided to try it in a clean checkout because I couldn't see how a change to hmmc2 would affect this test.