git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10982 0039d316-1c4b-4281-b951-d872f2087c98

The problem seems to be an error when overwriting file which is in use
(that's how EvictFileFromSystemCache works on Windows). So I used
a standard workaround for such Windowsic problems and created a wrapper
which retries 10 times. Similar code was there before my porting patch.

TBR=agl

Review URL: http://codereview.chromium.org/40160

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10981 0039d316-1c4b-4281-b951-d872f2087c98

- Methods of RenderView are called only on render thread in DevToolsAgent(earlier routing_id() was called on IO thread)

BUG=8378

Review URL: http://codereview.chromium.org/39182

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10980 0039d316-1c4b-4281-b951-d872f2087c98

This test failed for a period of time on Linux because of a bad V8
revision.
Review URL: http://codereview.chromium.org/39181

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10979 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: http://codereview.chromium.org/27240

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10978 0039d316-1c4b-4281-b951-d872f2087c98

terminate because it contains exponential regexps.  It is safe to
disable because we run the same tests (sans the nonterminating ones)
as part of the v8 tests.


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10977 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: http://codereview.chromium.org/40155

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10976 0039d316-1c4b-4281-b951-d872f2087c98

Implemented ResourceDispatcher::OnDownloadProgress to handle
ViewMsg_DownloadProgress IPC message, implementation is a clone of
ResourceDispatcher::OnUploadProgress because they have identical behavior.

Review URL: http://codereview.chromium.org/39104

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10975 0039d316-1c4b-4281-b951-d872f2087c98

performance impact.

TBR=dglazkov

Review URL: http://codereview.chromium.org/39177

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10974 0039d316-1c4b-4281-b951-d872f2087c98

1. Added entry to ResourceResponseHead so that it contains
   either a base::PlatformFile (OS_WIN) or
   base::FileDescriptor (OS_POSIX) for passing the file
   handle from browser to renderer process.
2. Also added IPC messages for reporting download progress
   and ACK message for it. ResourceLoaderBridge::Peer::OnDownloadProgress
   is added so that the peer is notified of the download
   progress in the renderer process.
3. Load flag to kick start the resource loading for media
   files. LOAD_MEDIA_RESOURCE is added so that
   ResourceDispatcherHost knows how to use a different
   ResourceHandler for handling media resource request.

Review URL: http://codereview.chromium.org/27168

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10972 0039d316-1c4b-4281-b951-d872f2087c98

This is the same change as "http://codereview.chromium.org/20219/show", which I reverted it because it caused build breaks on sandbox. To investigate this build break, it seems this build break is somehow caused by "base/string_util.cc" that includes the TrimWhiteSpaceUTF8() function. To fix this build break, I moved the TrimWhiteSpaceUTF8() function to "base/string_util_icu.cc".

BUG=7377
Review URL: http://codereview.chromium.org/28310

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10970 0039d316-1c4b-4281-b951-d872f2087c98

Also set the background color.

Also extend NineBox a little bit.

Review URL: http://codereview.chromium.org/40139

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10969 0039d316-1c4b-4281-b951-d872f2087c98

further review.

BUG=none
TEST=covered by layout tests
Review URL: http://codereview.chromium.org/40146

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10968 0039d316-1c4b-4281-b951-d872f2087c98

Promptly after writing the message for the last commit, I forgot to
git add the actual files! So here they are...


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10967 0039d316-1c4b-4281-b951-d872f2087c98

We had a bunch of layout tests where we never rebaselined the images
from Windows (because we were pixel exact). However, now that we
turned on anti-aliasing, we need to rebaseline these. Because they
were new baselined, I missed them in the last mega-batch of
rebaselines.


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10966 0039d316-1c4b-4281-b951-d872f2087c98

Now that we don't render with GTK, we don't need this wrapper for GDK
to draw widgets via Skia.


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10965 0039d316-1c4b-4281-b951-d872f2087c98

Setting the focus after the WebWidget handles the input event
instantly closes the popup menu created by the select control.

This change allows the popup menus to appear after one click
(as expected) and does not trigger the issue where the main
page never regains keyboard focus.

Review URL: http://codereview.chromium.org/40147

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10964 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: http://codereview.chromium.org/40152

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10963 0039d316-1c4b-4281-b951-d872f2087c98

going through the proper window shutdown sequence when the external host crashes.

The fix is to ensure that we clean up correctly when the external tab instance goes away. Will add a unit test in a subsequent CB.

Bug=1688967

Review URL: http://codereview.chromium.org/39086

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10962 0039d316-1c4b-4281-b951-d872f2087c98

passing consistently for the past 2 weeks.  Tests are:
LayoutTests/http/tests/navigation/redirect302-subframeload.html
LayoutTests/http/tests/navigation/success200-frames-loadsame.html
LayoutTests/http/tests/navigation/timerredirect-subframeload.html
LayoutTests/http/tests/navigation/success200-subframeload.html
LayoutTests/http/tests/navigation/redirect302-frames.html
LayoutTests/http/tests/navigation/timerredirect-frames.html
LayoutTests/http/tests/navigation/metaredirect-subframeload.html
Review URL: http://codereview.chromium.org/39171

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10961 0039d316-1c4b-4281-b951-d872f2087c98

Ensure that we display a sad face for a windowed plugin when the plugin process crashes. This was a regression introduced by creating the plugin parent on the browser UI thread, which remains valid when the plugin process crashes.

The fix is to track plugin wrapper windows in the PluginProcessHost class and destroy any remaining windows in the destructor. Removed the DestroyWindowTask
which destroys the windows on the UI thread. We can achieve the same result by posting WM_CLOSE messages to these windows.

This fixes http://code.google.com/p/chromium/issues/detail?id=7673

Bug=7673

Review URL: http://codereview.chromium.org/40120

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10960 0039d316-1c4b-4281-b951-d872f2087c98

* Now that WebCore uses floating-point scroll deltas, eliminate complicated carryover code and just use simple floating-point arithmetic when calculating scroll delta.
* Now that WebCore supports scrolling by page, plumb this instead of using a hacky "10 times the normal scroll amount" constant.
* Don't pretend shift was down when it wasn't (e.g. WM_MOUSEHWHEEL).
* Use SPI_GETWHEELSCROLLCHARS for horizontal scrolling, per MSDN.
* Fix horizontal scrolling to be "scroll down to go right" as the comment said (behavior was backwards)
* Clean up code.
* Reorder Mac/Linux code to match Windows code ordering.
Review URL: http://codereview.chromium.org/40135

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10959 0039d316-1c4b-4281-b951-d872f2087c98

We don't use GTK for drawing in WebKit any more, so we don't need it.


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10958 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: http://codereview.chromium.org/40142


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10956 0039d316-1c4b-4281-b951-d872f2087c98

I made a small change to grit so we can use our resource ID as the translation ID making it a bit easier to read.

The old locale_settings*.rc files would #include the google_strings.rc or chromium_strings.rc.  Instead, I added it to the vcproj file using $(CHROMIUM_BUILD) to grab the right file.  This required changing the output names in the grd files.

TBR=deanm

Review URL: http://codereview.chromium.org/39152

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10955 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: http://codereview.chromium.org/39168

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10954 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10952 0039d316-1c4b-4281-b951-d872f2087c98

Changes spellcheck_unittest to read dictionaries directly from the src tree.  Removes the build rules to copy test dictionaries to Dictionaries/.

Review URL: http://codereview.chromium.org/40082
Patch from rohitrao.

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10951 0039d316-1c4b-4281-b951-d872f2087c98

Previously, our command line objects were created /after/ gtk_init had
processed |argc| and |argv|. With this patch, it's created before hand
and main's |argc| and |argv| are now kept pristine. Thus our command
line will now reflect GTK arguments that previously it wouldn't have
seen. This should be harmless.

Review URL: http://codereview.chromium.org/39172


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10950 0039d316-1c4b-4281-b951-d872f2087c98

LayoutTests/fast/replaced/width100percent-textarea.html . The old
output was wrong because it had our old fonts.

BUG=none
TEST=none

Review URL: http://codereview.chromium.org/40140

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10946 0039d316-1c4b-4281-b951-d872f2087c98

fixed a couple layout tests.
Review URL: http://codereview.chromium.org/39153

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10945 0039d316-1c4b-4281-b951-d872f2087c98

when not results are found for a search string. Using your halo
methods does the trick.

BUG=8367
TEST=see bug

Review URL: http://codereview.chromium.org/39165

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10944 0039d316-1c4b-4281-b951-d872f2087c98

debug mode only.

Adding them back to tests_fixable in their original spot.

Review URL: http://codereview.chromium.org/37009

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10943 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: http://codereview.chromium.org/39166

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10942 0039d316-1c4b-4281-b951-d872f2087c98

instead integrated in the BrowserURLHandler for special schemes. This solves
a number of problems and cleans things up nicely.

Most of the functions were not necessary to have in the header file of the
browser about handler, so I made them local to the .cc file. I moved everything
around, but there was no change to any of the About...() functions.

This improves the about:memory page to not include the memory of the new tab
page it replaced. The entry for itself also has the proper title. This works
by using a meta refresh to the actual page, the the process transition no longer
happens at the same time as the about:memory page computation.

This also fixes problems with the about:network and about:ipc dialogs opening
blank pages and also re-opening the dialog when you close the browser.
Review URL: http://codereview.chromium.org/27238

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10941 0039d316-1c4b-4281-b951-d872f2087c98

Changed several references from "char" to "uint8" which is the appropriate definition for byte data buffers.

This change is small, but modifies the data source interface Read() method, and the VideoSurface data structure.  Mocks and
tests all have very small changes.

There are some changes in the mock_media_filters.h file that go beyond the scope of just changing char to uint8.  These
changes are required for future tests, and simply fill in parts of mocks that previously were NOTIMPLEMENTED().  The MockVideoFrame
was modified to allow other tests to create a video frame directly without using the MockFilterConfiguration data structure.  Also,
it is now possible to construct a mock pipeline that has no audio.   The currently checked-in pipeline will NOT work if you don't
have audio.  In a 2nd changelist, I will submit a new Pipeline_Impl and unit test that DOES support video without audio.  
Review URL: http://codereview.chromium.org/40059

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10940 0039d316-1c4b-4281-b951-d872f2087c98

We're opting to use pre-built 32-bit libs (from distro packages) rather than building locally to satisfy 64-bit dependencies.

Review URL: http://codereview.chromium.org/40064


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10939 0039d316-1c4b-4281-b951-d872f2087c98

- Fix a small leak (solved by re-selecting the initial bitmap into the dc at dtor time)
- Remove one set of GetDC(NULL) + GetDeviceCaps() + ReleaseDC()
- Remove some dead code

The removal of GetDC(NULL) might speed up Vista drawing a little bit, as is rummored that
this operation is expensive when DWM + Aero is enabled.

The integer overflow check is no longer possible. I'll file a bug so it is not lost.


TEST= existing test suffice

Review URL: http://codereview.chromium.org/21516

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10938 0039d316-1c4b-4281-b951-d872f2087c98

The only change from before is moving a method from a windows #ifdef
section to outside it.

TBR=evan

Review URL: http://codereview.chromium.org/39158


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10937 0039d316-1c4b-4281-b951-d872f2087c98

Minor changes to valgrind test scripts to add running on additional unit
tests.

LGTM-ed here: http://codereview.chromium.org/27306/show
(I didn't have my @chromium.org account then)



git-svn-id: svn://svn.chromium.org/chrome/trunk/src@10936 0039d316-1c4b-4281-b951-d872f2087c98