FAIL: run_io_test.sh on MacOS
Closed this issue · 6 comments
tseemann commented
clang -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -maes -mavx -O3 -pedantic -funroll-loops -Wall -std=gnu11 -fdata-sections -ffunction-sections -Wl,-dead_strip -Wno-undef -o rwaln rwaln-tldevel.o rwaln-rwalign.o rwaln-alphabet.o -lpthread -lm
/Applications/Xcode-10.2.1.app/Contents/Developer/usr/bin/make check-TESTS
PASS: alphabet
PASS: misc_test
PASS: edist
PASS: bpm_test
============================================================================
Testsuite summary for kalign 3.1
============================================================================
# TOTAL: 4
# PASS: 4
# SKIP: 0
# XFAIL: 0
# FAIL: 0
# XPASS: 0
# ERROR: 0
============================================================================
Making check in dev
/Applications/Xcode-10.2.1.app/Contents/Developer/usr/bin/make check-TESTS
Running kalign basic io test:
a2m.good.1
read/writea2m.good.1 FAILED
with ERROR 139 and Message:
[2019-10-27 20:10:28] : LOG : Start io tests.
[2019-10-27 20:10:28] : LOG : reading: ../dev/data//a2m.good.1
[2019-10-27 20:10:28] : LOG : Detected protein sequences.
[2019-10-27 20:10:28] : LOG : Done reading input sequences in 0.001876 seconds.
FAIL: run_io_test.sh
==================
1 of 1 test failed
==================
make[2]: *** [check-TESTS] Error 1
make[1]: *** [check-am] Error 2
make: *** [check-recursive] Error 1
TimoLassmann commented
Don't have a mac - while I'll try to get hold of one, could you test if kalign itself works on a mac?
% kalign dev/data/a2m.good.1
% kalign dev/data/a2m.good.1 -f msf
% kalign dev/data/a2m.good.1 -f clustal
Thank you!
tseemann commented
On mac by default one compiles with clang not gcc. I might force gcc.
Those 3 tests work fine on MacOS.
TimoLassmann commented
That's a relief - thanks for testing this!
I bet the problem has to do with the way make check finds the input files for the tests.
tseemann commented
MacOS filesystem is case-preserving but not case-sensitive. Could that be an issue?
TimoLassmann commented
This turned out to be a real bug. My code assumed that sequence files containing gaps would be aligned. The test case a2m.good.1 contains gaps but the sequences are unaligned.....
tseemann commented
Ouch, and good catch!