diff options
author | lukasza <lukasza@chromium.org> | 2016-03-15 14:58:50 -0700 |
---|---|---|
committer | Commit bot <commit-bot@chromium.org> | 2016-03-15 22:00:33 +0000 |
commit | 756abda6b173d7b698bf43a0726e43eaab8c03dd (patch) | |
tree | 0a5d5b744e536167cab9e3536e69cbf6e16e25bf /content/shell/browser | |
parent | 03ad6f33b5e9d4f3b22cfd6cda5fab9002fe52d6 (diff) | |
download | chromium_src-756abda6b173d7b698bf43a0726e43eaab8c03dd.zip chromium_src-756abda6b173d7b698bf43a0726e43eaab8c03dd.tar.gz chromium_src-756abda6b173d7b698bf43a0726e43eaab8c03dd.tar.bz2 |
Make kExposeInternalsForTesting usable again for interactive content_shell.
In https://crrev.com/1763113002 I've tried to separate
layout-test-specific cmdline switches into a separate header file and
handle some of them under
LayoutTestContentBrowserClient::AppendExtraCommandLineSwitches rather
than ShellContentBrowserClient::AppendExtraCommandLineSwitches. This
was wrong - this meant that devs cannot pass these switches when
debugging via content_shell running in interactive / non-layout-test
mode.
So - the current CL undoes part of https://crrev.com/1763113002 and
restores processing of some of the switches back to
ShellContentBrowserClient:
- kExposeInternalsForTesting - this is useful for debugging layout tests
(see https://crbug.com/594566).
For now it is okay to leave propagation of the following switches in
LayoutTestContentBrowserClient (because moving them in
https://crrev.com/1763113002 was not a regression as explained below):
- kStableReleaseMode - it only has effect if used together with
--run-layout-test switch (i.e. see how
ShellMainDelegate::BasicStartupComplete first checks
switches::kRunLayoutTest before checking switches::kStableReleaseMode
in a nested if).
- kEnableLeakDetection only affects BlinkTestController and
LayoutTestContentBrowserClient so it won't have any effect without
--run-layout-test switch.
- kEnableFontAntialiasing and kAlwaysUseComplexText only affects
LayoutTestContentBrowserClient and LayoutTestRenderProcessObserver so
it won't have any effect without --run-layout-test switch.
BUG=594566
Review URL: https://codereview.chromium.org/1799923002
Cr-Commit-Position: refs/heads/master@{#381330}
Diffstat (limited to 'content/shell/browser')
-rw-r--r-- | content/shell/browser/layout_test/layout_test_content_browser_client.cc | 4 | ||||
-rw-r--r-- | content/shell/browser/shell_content_browser_client.cc | 4 |
2 files changed, 4 insertions, 4 deletions
diff --git a/content/shell/browser/layout_test/layout_test_content_browser_client.cc b/content/shell/browser/layout_test/layout_test_content_browser_client.cc index 8626064..cc66507 100644 --- a/content/shell/browser/layout_test/layout_test_content_browser_client.cc +++ b/content/shell/browser/layout_test/layout_test_content_browser_client.cc @@ -98,10 +98,6 @@ void LayoutTestContentBrowserClient::AppendExtraCommandLineSwitches( command_line->AppendSwitch(switches::kEnableFontAntialiasing); } if (base::CommandLine::ForCurrentProcess()->HasSwitch( - switches::kExposeInternalsForTesting)) { - command_line->AppendSwitch(switches::kExposeInternalsForTesting); - } - if (base::CommandLine::ForCurrentProcess()->HasSwitch( switches::kStableReleaseMode)) { command_line->AppendSwitch(switches::kStableReleaseMode); } diff --git a/content/shell/browser/shell_content_browser_client.cc b/content/shell/browser/shell_content_browser_client.cc index 4fe9084..dd9840a 100644 --- a/content/shell/browser/shell_content_browser_client.cc +++ b/content/shell/browser/shell_content_browser_client.cc @@ -234,6 +234,10 @@ void ShellContentBrowserClient::AppendExtraCommandLineSwitches( base::CommandLine* command_line, int child_process_id) { if (base::CommandLine::ForCurrentProcess()->HasSwitch( + switches::kExposeInternalsForTesting)) { + command_line->AppendSwitch(switches::kExposeInternalsForTesting); + } + if (base::CommandLine::ForCurrentProcess()->HasSwitch( switches::kEnableCrashReporter)) { command_line->AppendSwitch(switches::kEnableCrashReporter); } |