Closed Bug 546651 Opened 14 years ago Closed 14 years ago

"Failed to post thread message" while debugging

Categories

(Core :: IPC, defect)

x86
Windows NT
defect
Not set
normal

Tracking

()

RESOLVED INVALID

People

(Reporter: benjamin, Assigned: benjamin)

References

Details

Attachments

(1 obsolete file)

While debugging firefox.exe I break into a function. Soon after I get "WARNING: Failed to post thread message!", probably from the plugin process. I did a little printf logging and GetLastError is ERROR_NOT_ENOUGH_QUOTA, which apparently means that that message limit of 10,000 messages is being overrun.

Ideas? Are we posting thread messages in a really tight loop or something?
This is a crappy solution: it just keeps looping if PostThreadMessage fails. I'm not sure whether we should do that only if the message queue is full, or for every type of failure. Also not sure if this is only likely to happen when a debugger is present and the Sleep should be much longer, as much as a second.
Assignee: nobody → benjamin
Attachment #427355 - Flags: review?(jmathies)
Attachment #427355 - Flags: review?(bent.mozilla)
OS: Linux → Windows NT
Comment on attachment 427355 [details] [diff] [review]
Keep looping if PostThreadMessage fails, rev. 1

>+#include "nsXULAppAPI.h"

I think this is unnecessary?

>+    ::Sleep(20);

I think 'SwitchToThread' might be more appropriate here.
Attachment #427355 - Flags: review?(bent.mozilla) → review+
We switched to using events in bug 545338.
Status: NEW → RESOLVED
Closed: 14 years ago
Resolution: --- → INVALID
Attachment #427355 - Attachment is obsolete: true
Attachment #427355 - Flags: review?(jmathies)
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: