Chromium Code Reviews (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out

Issue 460102: Recognize EMSGSIZE as non-fatal on OS X. (Closed)

11 years ago by jeremy
9 years, 6 months ago
TVL, agl
CC:, John Grabowski, Mark Mentovai


Recognize EMSGSIZE as non-fatal on OS X. BUG=29225 TEST=PageCyclerTest.Intl2File test should succeed. Committed:

Patch Set 1 #

Unified diffs Side-by-side diffs Delta from patch set Stats (+15 lines, -9 lines) Patch
M chrome/test/page_cycler/ View 1 chunk +2 lines, -1 line 0 comments Download
M ipc/ View 3 chunks +13 lines, -8 lines 0 comments Download


Total messages: 5 (0 generated)
* Also a small fix to page_cycler_tests. I couldn't figure out how to write a ...
11 years ago (2009-12-06 13:31:39 UTC) #1
Does this mean we spin like a top until the other side reads enough to ...
11 years ago (2009-12-07 00:45:08 UTC) #2
The underlying issue is that EMSGSIZE can be returned if the sockets output buffer has ...
11 years ago (2009-12-07 09:01:08 UTC) #3
My concern is even if we don't spin a tight loop here, are we gonna ...
11 years ago (2009-12-07 14:58:39 UTC) #4
11 years ago (2009-12-07 17:42:27 UTC) #5
This will spin the message loop while it remains the case that there is some
space in the send buffer, but not enough. However, the renderer process handles
messages on a different thread from rendering so hopefully this situation
doesn't persist for long. (At least it did last time I checked.)

Given that I don't have a good solution and that this situation is rare then
it's LGTM.

(Although you should probably add a comment to this effect somewhere in the code
before submitting.)

Powered by Google App Engine
This is Rietveld 408576698