Call ChangeInfo.Load() with fail_on_not_found=True to get a good error message in this case.

BUG=none
TEST=manual

Review URL: http://codereview.chromium.org/8692007

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@112018 0039d316-1c4b-4281-b951-d872f2087c98

Use it in subprocess2 and trial_dir to remove manually bookeeping.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8688011

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@111941 0039d316-1c4b-4281-b951-d872f2087c98

No code change beside tests. This is mainly a regression test to make sure I
don't regress that code path.

TBR=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8695003

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@111535 0039d316-1c4b-4281-b951-d872f2087c98

This will be necessary to override member functions eventually. It also better
replicates what subprocess.Popen is.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8570005

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@111530 0039d316-1c4b-4281-b951-d872f2087c98

Introduced git+http(s) fake URL schema to support git urls which don't have "filename.git" portion (http://foo/bar vs http://foo/bar/filename.git).

Review URL: http://codereview.chromium.org/8590024

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@110937 0039d316-1c4b-4281-b951-d872f2087c98

Change return value from list to tuple for communicate(timeout != None).

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8539015

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@110094 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: http://codereview.chromium.org/8536043

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109750 0039d316-1c4b-4281-b951-d872f2087c98

the current canonical upstream.  The Chromium mirror is out of date.
Review URL: http://codereview.chromium.org/8540026

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109742 0039d316-1c4b-4281-b951-d872f2087c98

Queries omahaproxy.appspot.com for related branch number.

Review URL: http://codereview.chromium.org/8498038

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109718 0039d316-1c4b-4281-b951-d872f2087c98

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8536018

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109697 0039d316-1c4b-4281-b951-d872f2087c98

Add relevant new tests.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8539016

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109692 0039d316-1c4b-4281-b951-d872f2087c98

It will simplify importing utility modules from other projects. Otherwise I was getting name conflicts with 'test'.

Reenable W0403 that was disabled in the previous CL.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8508015

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109636 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109635 0039d316-1c4b-4281-b951-d872f2087c98

Disable temporarily W0403, will be reenabled on the next CL

R=dpranke@chromium.org
BUG=
TEST=

Review URL: http://codereview.chromium.org/8508017

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109435 0039d316-1c4b-4281-b951-d872f2087c98

The main fix is 'Accept new empty files, like __init__.py'

TBR=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8519005

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109431 0039d316-1c4b-4281-b951-d872f2087c98

"Add callback support for stdout and stderr."
"Add tests for both universal_newlines=True and False."
"Implement accelerated tee support for POSIX."

The reason is that the semantic of
subprocess2.check_call(cmd, timeout=60, stderr=subprocess2.STDOUT) changed,
stderr wasn't redirected to VOID anymore and for an unknown reason, it was
hanging until the timeout happened. Will reapply r109239 with a new regression
test to detect that bug.

TBR=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8505046

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109369 0039d316-1c4b-4281-b951-d872f2087c98

subprocess2.check_output() raises an OSError instead of a CalledProcessError
when the executable is not found.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8511025

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109325 0039d316-1c4b-4281-b951-d872f2087c98

This removes all the need of threading, which removes the contention on the GIL
lock.

Taking S2Test.test_check_output_tee_large as baseline, numbers are
real/user/sys in seconds:
Ubuntu workstation: ~25x  2.4/1.9/1.5 -> 0.10/0.70/0.02
OSX 10.6 laptop:    ~40x  6.4/5.3/3.9 -> 0.15/0.80/0.07
Cygwin on win7:      ~4x  2.8/2.2/1.3 -> 0.60/0.16/0.30

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8462008

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109283 0039d316-1c4b-4281-b951-d872f2087c98

Fix stdout+stderr on Windows to binary mode in the child test process to make
sure testing is done correctly.

R=dpranke@chromium.org
BUG=
TEST=manually tested on OSX, Windows, cygwin and ubuntu


Review URL: http://codereview.chromium.org/8472002

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109282 0039d316-1c4b-4281-b951-d872f2087c98

It's currently an inefficient thread implementation. Interestingly
enough, callback support is significantly faster on cygwin than on
native python.

Writing an efficient implementation is punted for a later change,
one per implementation.

Stops using a temporary file since it's not necessary anymore.

The goal is to reduce the number of places where a similar paradigm
is used by having a canonical generic implementation.

R=dpranke@chromium.org
BUG=
TEST=Tested manually on Windows, cygwin, linux, OSX 10.6


Review URL: http://codereview.chromium.org/8374026

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109239 0039d316-1c4b-4281-b951-d872f2087c98

.tac files are used for twisted applications, like buildbot.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8501022

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@109238 0039d316-1c4b-4281-b951-d872f2087c98

Otherwise, silly checks like EOL whitespace triggers constantly.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8494013

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@108899 0039d316-1c4b-4281-b951-d872f2087c98

This would cause random try job failures, especially when switching around
DEPS <-> svn directory conversion.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8431021

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@108144 0039d316-1c4b-4281-b951-d872f2087c98

TEST=Run a passthru command like gcl di or gcl revert without a changelist name. Ensure that gcl exits with an error message.
BUG=none

Review URL: http://codereview.chromium.org/8413015

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@107663 0039d316-1c4b-4281-b951-d872f2087c98

BUG=
TEST=


Review URL: http://codereview.chromium.org/8353039

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@107590 0039d316-1c4b-4281-b951-d872f2087c98

Fixed slight bug in CMDpassthru where the command name was being passed to the CL Load rather than the CL name.

TEST=Try passthru commands like gcl revert and gcl stat and ensure that they perform as expected on each file in a CL.
BUG=none

Review URL: http://codereview.chromium.org/8404044

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@107576 0039d316-1c4b-4281-b951-d872f2087c98

BUG=79356
Review URL: http://codereview.chromium.org/8394016

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@107212 0039d316-1c4b-4281-b951-d872f2087c98

BUG=79356
Review URL: http://codereview.chromium.org/8369019

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@107206 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none

Review URL: http://codereview.chromium.org/8395005

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@107129 0039d316-1c4b-4281-b951-d872f2087c98

Windows users wouldn't get the same behavior on git cl vs gcl.
Improve automatic CRLF<->LF conversion, some gcl users would be \n repeated in
their description depending on the editor used.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8360007

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@107106 0039d316-1c4b-4281-b951-d872f2087c98

Instead of crashing, prints an error message.

BUG=
TEST=


Review URL: http://codereview.chromium.org/8363030

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@106867 0039d316-1c4b-4281-b951-d872f2087c98

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8366042

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@106866 0039d316-1c4b-4281-b951-d872f2087c98

isatty() wasn't defined on Windows, causing it to crash.

R=dpranke@chromium.org
BUG=
TEST=manually gclient sync on windows


Review URL: http://codereview.chromium.org/8371006

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@106865 0039d316-1c4b-4281-b951-d872f2087c98

Otherwise isatty() could not be proxied correctly.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8370004

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@106864 0039d316-1c4b-4281-b951-d872f2087c98

Crashes.

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@106709 0039d316-1c4b-4281-b951-d872f2087c98

For now only affect gclient help to figure out if there is any issue with it.

Because of the way it's done, there is no way to disable coloring, which could
be annoying when redirected. This needs to be taken in consideration.

R=dpranke@chromium.org
TEST=manually tested on linux, cygwin and windows, both run manually and
redirected to a text file.
BUG=


Review URL: http://codereview.chromium.org/8365001

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@106708 0039d316-1c4b-4281-b951-d872f2087c98

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8361006

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@106706 0039d316-1c4b-4281-b951-d872f2087c98

directory doesn't exist, on fresh checkout and --jobs >1.

It happens in the case where there is 3 dependencies:
a
a/b/c/d
a/b/c/e

'a' is processed first. Then 'a/b/c/d' and 'a/b/c/e' are fired simultaneously.
If both are not present, both svn checkout tries to mkdir b and b/c and a race
condition occurs between their call for isdir() and mkdir().

This works around the issue by creating the intermediary directories ourself
before calling out svn and managing the error ourself.

TBR=dpranke@chromium.org
BUG=
TEST=fresh checkout shouldn't be flaky


Review URL: http://codereview.chromium.org/8359018

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@106636 0039d316-1c4b-4281-b951-d872f2087c98

This eats stdout while gcl is committing, which is annoying. It needs tee-like
support first. Will revisit this change later, reverting in the meantime.

TBR=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8344085

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@106498 0039d316-1c4b-4281-b951-d872f2087c98

Reuse a common lock to reduce the number of locks created. Poor cygwin users.

R=dpranke@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/8353010

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@106494 0039d316-1c4b-4281-b951-d872f2087c98