August 2009 - webkit-dev - lists.webkit.org

Re: [webkit-dev] [webkit-changes] [47447] trunk
by Adam Roben 18 Aug '09

18 Aug '09

On Aug 18, 2009, at 3:53 PM, bweinstein(a)apple.com wrote: > Modified: trunk/WebKitTools/DumpRenderTree/win/UIDelegate.cpp (47446 > => 47447) > > --- trunk/WebKitTools/DumpRenderTree/win/UIDelegate.cpp 2009-08-18 > 19:53:27 UTC (rev 47446) > +++ trunk/WebKitTools/DumpRenderTree/win/UIDelegate.cpp 2009-08-18 > 19:53:31 UTC (rev 47447) > @@ -43,6 +43,8 @@ > #include <WebKit/WebKit.h> > #include <stdio.h> > > +#include <atlbase.h> > + > using std::wstring; > > class DRTUndoObject { > @@ -563,6 +565,17 @@ > /* [in] */ IWebSecurityOrigin *origin, > /* [in] */ BSTR databaseIdentifier) > { > + CComBSTR protocol; > + CComBSTR host; > + unsigned short port; > + > + origin->protocol(&protocol); > + origin->host(&host); > + origin->port(&port); > + > + if (!done && gLayoutTestController->dumpDatabaseCallbacks()) > + printf("UI DELEGATE DATABASE CALLBACK: > exceededDatabaseQuotaForSecurityOrigin:{%S, %S, %i} database:%S\n", > protocol, host, port, databaseIdentifier); > + > We avoid ATL in WebKit/DRT code because ATL is not included with VC++ Express. Presumably this patch has broken the VC++ Express build. -Adam

2 2

JSC without Dependencies
by Brian Barnes 18 Aug '09

18 Aug '09

My interest in WebKit is integrating my game engine with JSC. I'm far enough into the switch over that I've begun looking at compiling it on my own. My goal -- which might be a fools errand -- is to create a win32 VS compile without dependencies. I don't know if this is something that I could give back to the project or if it's at cross purposes with the project. The goal is to make it as easy as possible for VS users (like it is for XCode users): download, hit compile, and go. Any thoughts on this? A couple steps: 1) Use relative paths to remove the need for WEBKITLIBRARIESDIR and WEBKITOUTPUTDIR -- this can be especially problemantic if you pull the nightly builds as their top directory name changes. 2) Remove the need for external libraries, where possible. The dependent libraries are: CFLite (can be done with a switch, based on previous emails) icu*** (might not be possible, google says these are unicode tools) pthreadVC2 (based on previous email should not be needed) 3) Compile all within VS (no makes or external calls) to eliminate any additional dependencies on other tools Also, has anybody started anything like this, or is it so outside the scope it won't be useful to anybody but me? Or is it just something that will be impossible to untangle? After this I thought about possibly contributing some useful convenience functions, like JSValueIsArray, JSValueMakeArray, etc. [>] Brian

9 19

JSC Suggestion: Class property locking
by Brian Barnes 18 Aug '09

18 Aug '09

Another quick suggestion, and this is obviously just another convenience method (I'm sure I could handle this by overriding): It'd be nice to be able to lock a class (it could be a flag on the definition) to make adding properties (outside the static variable and function list) illegal. For instance, I have a problem with my users misspelling a property, then not understanding why nothing happens. In this case, they'd get an error. Basically, if you can't find a setter in my static lists, then toss an error. [>] Brian

2 2

No more build breaks.
by Eric Seidel 17 Aug '09

17 Aug '09

If you're currently using "svn commit" or "git svn dcommit" to commit your patches, please consider using "bugzilla-tool land-diff" instead. Why? 1. bugzilla-tool updates (and optionally closes) the bug for you when you're done. 2. bugzilla-tool makes a nice commit entry automatically for you (not quite as cool as commit-log-editor yet, but close). 3. MOST IMPORTANT: bugzilla-tool land-diff will automatically build and run the layout tests before landing! a. Impossible break at least one platform's bots when using bugzilla-tool! b. Impossible to inadvertently land when the tree is red! If you don't have time to build and run the tests before committing, you can always mark your patch for commit-queue+ instead and the commit queue script will commit it for you (after building and running the tests). See "bugzilla-tool --help" for more information. -eric

4 6

Ragarding Cygwin Installation
by raghavendrab.cs＠gmail.com 17 Aug '09

17 Aug '09

Dear All, I am working on webkit on windows plarform , but struck in the begining stage , Cygwin installation. I have installed the cygwin, but I am not able to update it using bash. whenever I do ./update-webkit it says... updating opensource ..(after some 2 min..) svn: OPTIONS of 'http://svn.webkit.org/repository/webkit/trunk': could not connect to server (http://svn.webkit.org) Died at ./update-webkit line 89. can anyone help me out? Regards, Raghavendra.

2 1

WebKit font size
by thouraya andolsi 17 Aug '09

17 Aug '09

Hi, I run the GTkluncher and I saw that the fonts are very small, if we do not use a css file. Does any one knows how to solve this problem? Is there some internal font setting options for font family, font size, etc to prevent Webkit to use very small font size for display? Best regards. Thouraya.

1 0

Application caching in the browser?
by Audrey Lee 16 Aug '09

16 Aug '09

Hello people, Are any of you implementing application caching in the browser? Here is the use-case I am interested in: End-user-Bob uses his browser to navigate to the_bob_end_user_crm_app.com While there he navigates to these URLs: the_bob_end_user_crm_app.com/accts/show/123 the_bob_end_user_crm_app.com/accts/show/321 the_bob_end_user_crm_app.com/accts/show/45678 Each of the above URLs displays dynamic HTML from Rails or an app-server. Next, he disconnects his browser from the web. He navigates to this URL: the_bob_end_user_crm_app.com/accts/show/87654 He receives a message: You are not connected to the net. the_bob_end_user_crm_app.com/accts/show/87654 is unavailable. He navigates to this URL: the_bob_end_user_crm_app.com/accts/show/123 He sees HTML which is cached in his browser via "application caching" Are any of you doing anything like this? I know it is possible because gmail.com exhibits this type of behavior. I'm looking for demos/tutorials/discussions which would teach me how to do this. Thanks, Audrey

2 1

Nightly build cannot render wikipedia properly
by Ryosuke Niwa 16 Aug '09

16 Aug '09

Use the newest Nightly build and visit http://en.wikipedia.org/wiki/Main_Page You see the sidebar showing up below the main pane. This is a regression in the past 12 hours. Ryosuke

1 1

Dependency tracking in the Qt build system
by Mark Rowe 14 Aug '09

14 Aug '09

On multiple occasions today we've been forced to manually do a clean build on the Qt build bot in order to work around test failures or crashes caused by bad builds. In the most recent case <http://trac.webkit.org/changeset/47165 > reordered some virtual methods in Element. This resulted in 14 failures on the Qt buildbot: <http://build.webkit.org/builders/Qt%20Linux%20Release/builds/472 >. Can someone with knowledge of the Qt build system please investigate what needs done to remove the need for this manual intervention in incremental builds so that keeping the Qt buildbot green is less painful? - Mark

5 5

GDOM patch spam
by Eric Seidel 14 Aug '09

14 Aug '09

I just closed out many of the GDOM bugs for patch spam after emailing the author of these patches. The proper process for getting code into WebKit does not involve uploading 15 unexplained patches at once. That tends to result only in annoying reviewers and getting you banned from the bug tracker. :) The GDOM binding patches have some history of trouble. I think if anyone wants to work on GDOM patches they need to talk to one of the long-standing Gtk contributers and approach this one patch at a time. http://trac.webkit.org/wiki/WebKit%20Team contains a list of some of the historical gtk contributers. -eric

5 4