WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED INVALID
36129
Get notified when a resource is added to or removed from cache
https://bugs.webkit.org/show_bug.cgi?id=36129
Summary
Get notified when a resource is added to or removed from cache
Yong Li
Reported
2010-03-15 12:50:57 PDT
Add CacheClient to monitor cache usage
Attachments
the patch
(2.75 KB, patch)
2010-03-15 12:57 PDT
,
Yong Li
no flags
Details
Formatted Diff
Diff
fix the build error on mac
(2.79 KB, patch)
2010-03-15 13:09 PDT
,
Yong Li
eric
: review-
Details
Formatted Diff
Diff
add more info to change log
(2.90 KB, patch)
2010-05-18 08:35 PDT
,
Yong Li
no flags
Details
Formatted Diff
Diff
fixed merge error
(2.90 KB, patch)
2010-05-18 10:07 PDT
,
Yong Li
ddkilzer
: review-
Details
Formatted Diff
Diff
Show Obsolete
(3)
View All
Add attachment
proposed patch, testcase, etc.
Yong Li
Comment 1
2010-03-15 12:57:42 PDT
Created
attachment 50726
[details]
the patch
Eric Seidel (no email)
Comment 2
2010-03-15 13:00:35 PDT
Attachment 50726
[details]
did not build on mac: Build output:
http://webkit-commit-queue.appspot.com/results/857010
Yong Li
Comment 3
2010-03-15 13:09:05 PDT
Created
attachment 50729
[details]
fix the build error on mac
Eric Seidel (no email)
Comment 4
2010-03-15 13:12:19 PDT
Comment on
attachment 50729
[details]
fix the build error on mac Why do we want this change? Do other ports already do similar things through different means? The ChangeLog needs more explanation as to why this is a good idea before I would feel comfortable reviewing this.
Yong Li
Comment 5
2010-05-18 08:35:25 PDT
Created
attachment 56378
[details]
add more info to change log
Yong Li
Comment 6
2010-05-18 10:07:22 PDT
Created
attachment 56385
[details]
fixed merge error
Eric Seidel (no email)
Comment 7
2010-05-20 01:01:03 PDT
Other ports must handle low-memory situations for the WebCore cache differently than how RIM does.
Yong Li
Comment 8
2010-05-20 08:48:30 PDT
(In reply to
comment #7
)
> Other ports must handle low-memory situations for the WebCore cache differently than how RIM does.
Then how do other ports get notified when a dead resource in added to/removed from the cache?
David Kilzer (:ddkilzer)
Comment 9
2010-05-20 09:42:16 PDT
(In reply to
comment #8
)
> (In reply to
comment #7
) > > Other ports must handle low-memory situations for the WebCore cache differently than how RIM does. > > Then how do other ports get notified when a dead resource in added to/removed from the cache?
iPhone WebKit doesn't track resources on an individual basis. The kernel generates memory warnings to which WebKit responds by pruning resources based on the warning level. Is it not sufficient to use existing methods on WebCore::Cache to determine its size? Why can't your memory allocator keep track of the memory used by WebKit? (I'm not familiar with RIM's OS, although I know it's an RTOS.)
Antti Koivisto
Comment 10
2010-05-20 11:17:24 PDT
(I addition to responding memory warnings, Darwin based WebKit ports also keeps cached resource data in purgeable memory that can be automatically reclaimed by the system.) Getting notified for each cache change seems unnecessary, at least for implementing low memory handling.
David Kilzer (:ddkilzer)
Comment 11
2010-07-29 16:34:01 PDT
Comment on
attachment 56385
[details]
fixed merge error r- since there hasn't been a response to
Comment #9
or
Comment #10
. If there really isn't any other way to do this on your platform, please provide a little more information or acknowledgement.
Yong Li
Comment 12
2010-07-30 12:04:44 PDT
(In reply to
comment #11
)
> (From update of
attachment 56385
[details]
) > r- since there hasn't been a response to
Comment #9
or
Comment #10
. > If there really isn't any other way to do this on your platform, please provide a little more information or acknowledgement.
sorry, somehow I missed the notifications. I think Antti made a good point. I can do that in another way.
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug