ParseSymFileError: firefox 38FBF249F41F7A64B3A6098BCC225F920
Categories
(Eliot :: General, task, P2)
Tracking
(Not tracked)
People
(Reporter: willkg, Unassigned)
Details
I'm seeing a handful of ParseSymFileErrors in the Mozilla Symbolication Service starting August 11th for firefox 38FBF249F41F7A64B3A6098BCC225F920.
The file record on symbols.mozilla.org is here:
https://symbols.mozilla.org/uploads/files/file/48861769
The header looks like this:
MODULE Linux x86_64 65B29E6A53804B5D86B2D3497464C71C0 firefox
INFO CODE_ID 6A9EB26580535D4B86B2D3497464C71C4EEACF30
INFO GENERATOR mozilla/dump_syms 2.2.0
65B29E6A53804B5D86B2D3497464C71C0 does not match the debug id 38FBF249F41F7A64B3A6098BCC225F920.
Did dump_syms somehow put a different debug id in the file? Is this a problem elsewhere with the build?
|
Reporter | |
Comment 1•2 years ago
|
||
I think it's this nightly build task:
https://firefox-ci-tc.services.mozilla.com/tasks/RKqxuPHUQEKQKdyF6YLNXQ
I'm not sure what else I can figure out from here. I'm going to write up an issue in the dump_syms project.
|
|
Reporter | |
Comment 2•2 years ago
|
||
There hasn't been any action on the dump_syms bug, so maybe it's fine. I'm going to unassign myself for now. If no one has done anything by 2024, we should mark it WONTFIX.
|
||
Comment 3•2 years ago
|
||
I'm slowly working my way through dump_syms' backlog, is this something that is still happening?
|
|
Reporter | |
Comment 4•2 years ago
|
||
I don't see other instances of the error, so either it's not still happening or the files that are messed up aren't being requested.
|
|
||
Comment 5•2 years ago
|
||
OK, let's keep an eye on this. I've gone over the changes we did in dump_syms since we encountered this error and there's nothing that really stands out. However, the path where we store the file is not really provided by dump_syms, but rather by symbolstore.py. We read the MODULE line to produce this path, so I don't see how that could have failed. Either way given that dump_syms now has a lot more functionality than before I might get rid of (most) of the pieces of symbolstore.py that handle the .sym file and let dump_syms do the job (including creating the proper path). This should make it harder for a mismatch to happen.
Description
•