Summary: | Baloo Crash - EPub Extractor | ||
---|---|---|---|
Product: | [Frameworks and Libraries] frameworks-kfilemetadata | Reporter: | Jacek Pawlyta <cunio> |
Component: | general | Assignee: | Pinak Ahuja <pinak.ahuja> |
Status: | RESOLVED DUPLICATE | ||
Severity: | crash | CC: | aspotashev, christoph, igor.poboiko |
Priority: | NOR | Keywords: | drkonqi |
Version: | unspecified | ||
Target Milestone: | --- | ||
Platform: | unspecified | ||
OS: | Linux | ||
Latest Commit: | Version Fixed In: | ||
Sentry Crash Report: |
Description
Jacek Pawlyta
2015-09-05 19:52:29 UTC
Can you please locate and provide the .epub file it crashed on? @Igor, how I can do it? I have quite a couple of hundred of epub files. (In reply to Jacek Pawlyta from comment #2) > @Igor, how I can do it? I have quite a couple of hundred of epub files. Well, there is a hard way. First of all, is it always reproducible? If so, you can first of all clean Baloo index by running in shell # balooctl disable # balooctl enable Next step is to get PID of baloo_file_extractor process: # ps ax | grep baloo And then you can attach a "strace" tool to that process # sudo strace -e write=1 -p PID What it will do is following: baloo_file process performs indexing of contents by calling baloo_file_extractor and feeding it with filenames using pipelines. This will intercept all "write()" system calls, printing all data that passed through that pipeline. WARNING: it will produce a lot of output. You can redirect it to some file by adding ">file 2>&1" to the end of last command. File "file" can grow up to several hundreds of megabytes though, be careful. Hopefully the last filename that it prints before crash will be the one. (In reply to Jacek Pawlyta from comment #2) > @Igor, how I can do it? I have quite a couple of hundred of epub files. Another way is to use bisecting. You can put all your .epub-files into some directory Baloo won't index (you can add it using SystemSettings), and then move half of it to place where it will index. Then wait for baloo to index them. Repeat with half of files left, and so on. Do it until you find the one :) And, BTW, what Linux distributive do you use? Any chance it is source-based, so you can build a baloo from source code with some patches? :) The epub extractor checks not enough (and double frees), see other bug. *** This bug has been marked as a duplicate of bug 361727 *** |