Detect invalid content settings pattern that were not detected yet.

Don't allow the following content settings patterns:
- Patterns with file scheme and non empty host
- File patterns that user a wildcard '*' symbol in their path
- "file:///"
- Patterns with an IP address and a domain wildcard.

Add tests to check that "[*.]", "http://[*.]", ... are valid patterns.

BUG=104414
TEST=ContentSettingsPattern*


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=111546

[markusheintz_, 2011-11-24 15:11:42]

Please review my CL. Thanks.

[Bernhard Bauer, 2011-11-24 15:50:32]

There are multiple changes in this CL. Can you update the description to be a
bit more accurate?

http://codereview.chromium.org/8676020/diff/6005/chrome/common/content_settin...
File chrome/common/content_settings_pattern.h (right):

http://codereview.chromium.org/8676020/diff/6005/chrome/common/content_settin...
chrome/common/content_settings_pattern.h:223: // the pattern parts are not
valid.
Canonicalize seems to use a different definition of valid than Validate below.
What's the difference?

http://codereview.chromium.org/8676020/diff/6005/chrome/common/content_settin...
File chrome/common/content_settings_pattern_parser.cc (right):

http://codereview.chromium.org/8676020/diff/6005/chrome/common/content_settin...
chrome/common/content_settings_pattern_parser.cc:143: // If the host contains a
domain wildcard but no domain then it is
This would be a pattern that's just "[*.]"? I think we should allow it (but then
stringify it to "*" so that users know what's the preferred syntax).

[markusheintz_, 2011-11-24 16:37:51]

http://codereview.chromium.org/8676020/diff/6005/chrome/common/content_settin...
File chrome/common/content_settings_pattern.h (right):

http://codereview.chromium.org/8676020/diff/6005/chrome/common/content_settin...
chrome/common/content_settings_pattern.h:223: // the pattern parts are not
valid.
On 2011/11/24 15:50:32, Bernhard Bauer wrote:
> Canonicalize seems to use a different definition of valid than Validate below.
> What's the difference?

I rephrased it.

http://codereview.chromium.org/8676020/diff/6005/chrome/common/content_settin...
File chrome/common/content_settings_pattern_parser.cc (right):

http://codereview.chromium.org/8676020/diff/6005/chrome/common/content_settin...
chrome/common/content_settings_pattern_parser.cc:143: // If the host contains a
domain wildcard but no domain then it is
On 2011/11/24 15:50:32, Bernhard Bauer wrote:
> This would be a pattern that's just "[*.]"? I think we should allow it (but
then
> stringify it to "*" so that users know what's the preferred syntax).

True. Done

[Bernhard Bauer, 2011-11-24 16:58:35]

LGTM.

[commit-bot: I haz the power, 2011-11-24 17:17:43]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/markusheintz@chromium.org/8676020/5002

[commit-bot: I haz the power, 2011-11-24 19:01:32]

Change committed as 111546