Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

SIGSEGV in a very specific setting #17

Open
mbollmann opened this issue Oct 21, 2017 · 2 comments
Open

SIGSEGV in a very specific setting #17

mbollmann opened this issue Oct 21, 2017 · 2 comments
Labels

Comments

@mbollmann
Copy link
Member

When running

normalize -c test.cfg -f test.txt --train

with the attached sample files, Norma (0.5.0 with ICU strings) randomly segfaults (SIGSEGV) about 80% of the time.

This...

  • happens only when the parameter files are not yet created
  • happens also when the lines in test.txt are followed by 50000+ more tokens
  • only seems to happen when the lines in test.txt are at the beginning of a file
  • is the only segfault I experienced in almost 1.2 millions tokens of training

There must be something special going on here, though I don't know what it is.

test.cfg.txt
test.txt

@mbollmann mbollmann added the bug label Oct 21, 2017
@fpetran
Copy link
Collaborator

fpetran commented Oct 21, 2017

If the segfault does not occur 100% of the times, it may be due to race conditions in the multithreading. Can you send a backtrace?

@mbollmann
Copy link
Member Author

This line appears to be the culprit:

https://github.com/comphist/norma/blob/master/src/normalizer/rulebased/rule.cpp#L224

Maybe the array index can go out of bounds?

Full backtrace:
backtrace.txt

mbollmann pushed a commit that referenced this issue Apr 14, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

2 participants