#261 FATAL error when performing tide-search with trunk

[Alice Cheng]

Terminal output:

./crux tide-search --overwrite T 3294_3300.ms2 yeast_tide
INFO: Beginning tide-search.
WARNING: The output directory 'crux-output' already exists.
Existing files will be overwritten.
WARNING: The file 'crux-output/tide-search.log.txt' already exists and will be overwritten.
INFO: CPU: tryptophan.gs.washington.edu
INFO: Wed Apr 1 12:29:48 PDT 2015
INFO: Running tide-search...
INFO: Reading index yeast_tide
WARNING: The file 'crux-output/tide-search.target.txt' already exists and will be overwritten.
WARNING: The file 'crux-output/tide-search.decoy.txt' already exists and will be overwritten.
INFO: Converting 3294_3300.ms2 to spectrumrecords format
FATAL: boost::bad_lexical_cast occured while parsing spectrum.
Does your spectra contain z-lines?

System Info:
Tryptophan
gcc --version = 4.4.7

Build information (compiled under debug):
Svn info:
URL: svn+ssh://acheng94@svn.code.sf.net/p/cruxtoolkit/code/crux/trunk
Repository Root: svn+ssh://acheng94@svn.code.sf.net/p/cruxtoolkit/code
Repository UUID: 212732f7-e82a-48c5-b261-38c1ba6edc5e
Revision: 16679

svn diff = none

I have tried downloading a fresh copy out of trunk and compiling under default build type, but I still get the same error.

I have tried ssh into tyrosine and lysine to run "make", but the binary files still give the same error.

I have used an older crux binary with the same tide index and ms2 file and it works, so my files should be valid.

I would attach the binary file, but the file is too large.

[Kaipo]

Hi Alice, I tried this on tyrosine in both debug and release mode and didn't get the error. Could you point me to the debug binary that is causing the problem?

[Kaipo]

I'm not sure exactly why this error is happening but it seems to go away with a newer version of gcc. In my testing, the problem showed up when I compiled with gcc 4.4.7 but not with gcc 4.6.2 or 4.6.3.
Alice, can you try using a newer version of gcc (I am using 4.6.2 but there are also newer modules for 4.7.0, 4.8.1, and 4.9.1) and seeing if the problem goes away? If so, I don't think this is worth tracking down. It might be some type of compiler bug or boost compatibility issue.

[Alice Cheng]

I tried a deep clean, cmake, and make with gcc 4.9.1 after loading it from module, and I'm still getting the same error.

[Kaipo]

Hmm, I am still looking into this but it's really odd. It seems that when compiled with gcc 4.4.7, 4.7.0, or 4.9.1 the error occurs. When compiled with gcc 4.6.2, 4.6.3, or 4.8 it does not.

[Kaipo]

Not sure why the different behavior between compiler versions but I committed a fix for this to proteowizard. I have a feeling it will fix the problem with issue 256, but I dont have a mac for testing.

[Kaipo]

fixed in r7364 of proteowizard project