is WebKit leaking or caching memory?
Hello. I'm working on an application that uses WebKit as a "rendering farm". A running program has a shared memory and an WebKit instance using the Gtk + port. I've modified WebKit and added: - support for off-screen rendering, meaning instead of showing pages to the screen I render them onto a cairo surface using the patches from the Clutter project, - a function that simulates a click on a page with given X,Y parameters. - a new signal to the Gtk+ port that triggers whenever a visible change occurred on the page (opening html, JavaScript modifications) The HTML page has a DIV tag that is sensitive to mouse clicks. A click on that DIV tag will cause the page to modify. Now I have a client program that connects to the WebKit program using IPC. Then I load the mentioned HTML page and sends click requests every 0.25 seconds with the coordinates where the DIV tag is located. What I've notices is, that after a while, memory of the WebKit program increases. The program was running for about one hour and the process memory increased for about 7 MB. Now I'm asking if the memory increment is a result of WebKit caching subsystem or could that be a leak? Keep in mind that I click/refresh the same page over and over again. Greets, Luka
On Monday 01 December 2008 13:22:08 Luka Napotnik wrote:
Now I'm asking if the memory increment is a result of WebKit caching subsystem or could that be a leak? Keep in mind that I click/refresh the same page over and over again.
Leak, Cache, Memory fragmentation are all possible. I think by default we do not cache.... so leaking and fragmentation are more likely. Someone found a leak in the Font handling of Gtk+ [1] (first bit) which might apply to the stock Gtk+ as well. z. [1] https://bugs.webkit.org/show_bug.cgi?id=15914#c45
Hello. I've fixed the pango leak but the program still gradually leaks memory. Are there any other known leaks? Greets, Luka Dne 01.12.2008 (pon) ob 14:04 +0100 je Holger Freyther zapisal(a):
On Monday 01 December 2008 13:22:08 Luka Napotnik wrote:
Now I'm asking if the memory increment is a result of WebKit caching subsystem or could that be a leak? Keep in mind that I click/refresh the same page over and over again.
Leak, Cache, Memory fragmentation are all possible. I think by default we do not cache.... so leaking and fragmentation are more likely. Someone found a leak in the Font handling of Gtk+ [1] (first bit) which might apply to the stock Gtk+ as well.
z.
[1] https://bugs.webkit.org/show_bug.cgi?id=15914#c45 _______________________________________________ webkit-dev mailing list webkit-dev@lists.webkit.org http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
participants (2)
-
Holger Freyther
-
Luka Napotnik