bug-1907983: fix sym file header parsing #2991
Conversation
This fixes sym file header parsing so it parses the entire sym file header rather than stopping after INFO CODE_ID. While doing this, I also updated the tests for windows, mac, and linux sym file headers with dump_syms 2.3.3 output.
| if len(parts) == 3: | ||
| _, _, code_id = parts | ||
| code_file = "" | ||
| elif len(parts) == 4: | ||
| _, _, code_id, code_file = parts |
There was a problem hiding this comment.
seems like this should have an else ValueError condition in case the file is corrupted or something
There was a problem hiding this comment.
When I initially wrote the extraction code, I thought about validation a bit, but decided that's outside of the scope. This code parses the bits it understands so it can fill in some metadata. If the header is malformed or invalid in some way, we don't have any mechanisms for doing anything with that knowledge. Given that, I wrote it such that it just ignores those cases.
Bug 1814430 covers writing a better information view for symbols files which would cover whether they're valid. Adding header validation to that would be helpful.
A while back, I wrote bin/debug-sym-file.py that downloads a file and tells you things about it including whether it parses in symbolic. I used that to debug malformed symbols files for bug 1791785. We could add header validation to that.
|
Thank you! |
willkg
changed the title
This fixes sym file header parsing so it parses the entire sym file header rather than stopping after INFO CODE_ID.
While doing this, I also updated the tests for windows, mac, and linux sym file headers with dump_syms 2.3.3 output.