Closed
Bug 927001
Opened 13 years ago
Closed 13 years ago
stackwalker not properly escaping with json output
Categories
(Toolkit :: Crash Reporting, defect)
Tracking
()
RESOLVED
INVALID
People
(Reporter: lars, Assigned: ted)
References
Details
Attachments
(1 file)
|
151.01 KB,
text/plain
|
Details |
I've traced a problem down to the stackwalker program emitting invalid json.
ValueError: Invalid \escape: line 1 column 27627 (char 27627)
| Reporter | ||
Comment 1•13 years ago
|
||
a crash that demonstrates this issue is de338c41-8ada-4112-895c-e871a2131015 on staging. I do not have an example from prod. If you cannot get this crash dump for some reason, I've got it already downloaded and I can send it to you via SD card tied to a carrier pigeon.
Blocks: 636868
| Assignee | ||
Updated•13 years ago
|
Attachment #817246 -
Attachment mime type: application/json → text/plain
| Assignee | ||
Comment 2•13 years ago
|
||
It appears to be barfing on:
"function":"`anonymous namespace\'::ThreadFunc(void *)"
The backslash doesn't actually seem to be necessary here. I'll see if there's a way to coerce jsoncpp to not produce it.
| Assignee | ||
Comment 3•13 years ago
|
||
I can't reproduce this with my local copy of stackwalker, it just spits out:
"function":"`anonymous namespace'::ThreadFunc(void *)"
There isn't anything mangling the output in between, right? If you run stackwalker on this dump (with the symbols present) and pipe it to a file does it produce the output with the extraneous quote?
| Reporter | ||
Comment 4•13 years ago
|
||
You are correct - I found an obscure place where output from the stackwalk subprocess is being run through 'repr' one line at a time before being given to the processor. I'm now trying to determine why it was put there before I strip it out.
Status: NEW → RESOLVED
Closed: 13 years ago
Resolution: --- → INVALID
You need to log in
before you can comment on or make changes to this bug.
Description
•