On Windows, don't try to create relative paths across different drives

On Windows, don't try to create relative paths across different drives

This is problematic when the generator output is set to a different drive
from the gyp files and sources.

[borenet, 2013-07-11 12:34:06]

Patch set 2: use lower()

[borenet, 2013-07-12 20:12:53]

Friendly ping

[Sam Clegg, 2013-07-12 22:03:32]

lgtm

[borenet, 2013-07-15 14:37:13]

On 2013/07/12 22:03:32, Sam Clegg wrote:
> lgtm

Mind committing this for me?  The commit queue doesn't seem to be doing
anything.  Thanks!

[Nico, 2013-07-15 17:06:58]

Is it possible to write a test for this? Every gyp change should have a test
that explains why the change was made.

[Sam Clegg, 2013-07-15 17:30:34]

On 2013/07/15 17:06:58, Nico wrote:
> Is it possible to write a test for this? Every gyp change should have a test
> that explains why the change was made.

Does gyp have unittests as well as system tests?  If not, perhaps its time to
create some.
This seems like a good case for a unit test.

[borenet, 2013-07-16 15:16:39]

Not sure if my test is in the right place, it's a little ugly, and I haven't
been able to run it yet (see comment).  Seems like a unit test would be better
in this case.

https://codereview.chromium.org/18991011/diff/12001/test/win/generator-output...
File
test/win/generator-output-different-drive/gyptest-generator-output-different-drive.py
(right):

https://codereview.chromium.org/18991011/diff/12001/test/win/generator-output...
test/win/generator-output-different-drive/gyptest-generator-output-different-drive.py:1:
#!/usr/bin/env python
This is a rough sketch of what a test might look like.  It's probably fragile,
since we need to do the drive mapping.  Unfortunately, I haven't been able to
run this, since all of the tests are failing for me on Windows with:

Error: could not find devenv
Exception AttributeError: "'TestGypNinja' object has no attribute '_dirlist'" in
<bound method TestGypNinja.__del__ of 1f9c0f0> ignored

Any ideas?

[borenet, 2013-07-19 14:26:43]

ping

[scottmg, 2013-07-19 15:01:55]

On 2013/07/16 15:16:39, borenet wrote:
> Not sure if my test is in the right place, it's a little ugly, and I haven't
> been able to run it yet (see comment).  Seems like a unit test would be better
> in this case.
> 
>
https://codereview.chromium.org/18991011/diff/12001/test/win/generator-output...
> File
>
test/win/generator-output-different-drive/gyptest-generator-output-different-drive.py
> (right):
> 
>
https://codereview.chromium.org/18991011/diff/12001/test/win/generator-output...
>
test/win/generator-output-different-drive/gyptest-generator-output-different-drive.py:1:
> #!/usr/bin/env python
> This is a rough sketch of what a test might look like.  It's probably fragile,
> since we need to do the drive mapping.  Unfortunately, I haven't been able to
> run this, since all of the tests are failing for me on Windows with:
> 
> Error: could not find devenv
> Exception AttributeError: "'TestGypNinja' object has no attribute '_dirlist'"
in
> <bound method TestGypNinja.__del__ of 1f9c0f0> ignored

What version/edition of Visual Studio do you have installed? It seems like it's
not getting found for whatever reason.

[borenet, 2013-07-19 15:27:06]

On 2013/07/19 15:01:55, scottmg wrote:
> On 2013/07/16 15:16:39, borenet wrote:
> > Not sure if my test is in the right place, it's a little ugly, and I haven't
> > been able to run it yet (see comment).  Seems like a unit test would be
better
> > in this case.
> > 
> >
>
https://codereview.chromium.org/18991011/diff/12001/test/win/generator-output...
> > File
> >
>
test/win/generator-output-different-drive/gyptest-generator-output-different-drive.py
> > (right):
> > 
> >
>
https://codereview.chromium.org/18991011/diff/12001/test/win/generator-output...
> >
>
test/win/generator-output-different-drive/gyptest-generator-output-different-drive.py:1:
> > #!/usr/bin/env python
> > This is a rough sketch of what a test might look like.  It's probably
fragile,
> > since we need to do the drive mapping.  Unfortunately, I haven't been able
to
> > run this, since all of the tests are failing for me on Windows with:
> > 
> > Error: could not find devenv
> > Exception AttributeError: "'TestGypNinja' object has no attribute
'_dirlist'"
> in
> > <bound method TestGypNinja.__del__ of 1f9c0f0> ignored
> 
> What version/edition of Visual Studio do you have installed? It seems like
it's
> not getting found for whatever reason.

I'm using VS2010 Express.  I've pasted my environment variables below.  They
seem correct to me:

ALLUSERSPROFILE=C:\ProgramData
APPDATA=C:\Users\borenet\AppData\Roaming
CommonProgramFiles=C:\Program Files\Common Files
CommonProgramFiles(x86)=C:\Program Files (x86)\Common Files
CommonProgramW6432=C:\Program Files\Common Files
COMPUTERNAME=BORENET0-W
ComSpec=C:\Windows\system32\cmd.exe
DevEnvDir=C:\Program Files (x86)\Microsoft Visual Studio 10.0\Common7\IDE\
DXSDK_DIR=C:\Program Files (x86)\Microsoft DirectX SDK (June 2010)\
EDITOR=vim
FP_NO_HOST_CHECK=NO
Framework35Version=v3.5
FrameworkDir=C:\Windows\Microsoft.NET\Framework\
FrameworkDIR32=C:\Windows\Microsoft.NET\Framework\
FrameworkVersion=v4.0.30319
FrameworkVersion32=v4.0.30319
HOMEDRIVE=C:
HOMEPATH=\Users\borenet
INCLUDE=C:\Program Files (x86)\Microsoft Visual Studio
10.0\VC\INCLUDE;C:\Program Files (x86)\Microsoft SDKs\Window
s\v7.0A\include;
LIB=C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\LIB;C:\Program Files
(x86)\Microsoft SDKs\Windows\v7.0A\
lib;
LIBPATH=C:\Windows\Microsoft.NET\Framework\v4.0.30319;C:\Windows\Microsoft.NET\Framework\v3.5;C:\Program
Files (x86
)\Microsoft Visual Studio 10.0\VC\LIB;
LOCALAPPDATA=C:\Users\borenet\AppData\Local
LOGONSERVER=\\HOT-DC-2
NUMBER_OF_PROCESSORS=12
OS=Windows_NT
Path=C:\Program Files (x86)\Microsoft Visual Studio 10.0\Common7\IDE\;C:\Program
Files (x86)\Microsoft Visual Studi

PATHEXT=.COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC
PROCESSOR_ARCHITECTURE=AMD64
PROCESSOR_IDENTIFIER=Intel64 Family 6 Model 44 Stepping 2, GenuineIntel
PROCESSOR_LEVEL=6
PROCESSOR_REVISION=2c02
ProgramData=C:\ProgramData
ProgramFiles=C:\Program Files
ProgramFiles(x86)=C:\Program Files (x86)
ProgramW6432=C:\Program Files
PROMPT=$P$G
PSModulePath=C:\Windows\system32\WindowsPowerShell\v1.0\Modules\
PUBLIC=C:\Users\Public
SESSIONNAME=Console
SKIA_OUT=O:\out
SystemDrive=C:
SystemRoot=C:\Windows
TEMP=C:\Users\borenet\AppData\Local\Temp
TMP=C:\Users\borenet\AppData\Local\Temp
UATDATA=C:\Windows\CCM\UATData\D9F8C395-CAB8-491d-B8AC-179A1FE1BE77
USERDNSDOMAIN=AD.CORP.GOOGLE.COM
USERDOMAIN=GOOGLE
USERNAME=borenet
USERPROFILE=C:\Users\borenet
VCINSTALLDIR=C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\
VS100COMNTOOLS=C:\Program Files (x86)\Microsoft Visual Studio
10.0\Common7\Tools\
VSINSTALLDIR=C:\Program Files (x86)\Microsoft Visual Studio 10.0\
windir=C:\Windows
WindowsSdkDir=C:\Program Files (x86)\Microsoft SDKs\Windows\v7.0A\
windows_tracing_flags=3
windows_tracing_logfile=C:\BVTBin\Tests\installpackage\csilogfile.log

[borenet, 2013-07-19 16:15:41]

On Fri, Jul 19, 2013 at 11:36 AM, Scott Graham <scottmg@chromium.org> wrote:
>I'm not sure, but it's possible gyptest.py doesn't work with Express. I know
the bots are running Pro anyway.
>
>You might try setting GYP_MSVS_VERSION=2010e, I don't know if that is enough to
help it.

Setting GYP_MSVS_VERSION=2010e didn't help.  In the meantime, thoughts on the
test itself?  It seems a little heavy-handed to me.

[Nico, 2013-07-19 16:25:21]

The test looks good to me.

Not sure If checking os.name is the best way to check for windows (I think we
check sys.platform in most other places in gyp?), but I'm not terribly familiar
with Windows. lgtm with some conscious decision about this.

(You can probably send tryjobs to run your new test)

[scottmg, 2013-07-19 16:51:17]

https://codereview.chromium.org/18991011/diff/12001/pylib/gyp/common.py
File pylib/gyp/common.py (right):

https://codereview.chromium.org/18991011/diff/12001/pylib/gyp/common.py#newco...
pylib/gyp/common.py:136: if os.name == 'nt':
i'm not sure whether this will work on cygwin... isn't there some /cygdrive/
crud it uses instead? does splitdrive work on that? so maybe sys.platform ==
'win32' makes more sense. (other places use sys.platform, so if you think it
works on cygwin, then sys.platform in ('win32', 'cygwin') for consistency)

[borenet, 2013-07-19 17:58:42]

Patch set 4: use sys.platform instead of os.name.  I guess I'll keep iterating
and running trybots until the test works correctly.

https://codereview.chromium.org/18991011/diff/12001/pylib/gyp/common.py
File pylib/gyp/common.py (right):

https://codereview.chromium.org/18991011/diff/12001/pylib/gyp/common.py#newco...
pylib/gyp/common.py:136: if os.name == 'nt':
On 2013/07/19 16:51:17, scottmg wrote:
> i'm not sure whether this will work on cygwin... isn't there some /cygdrive/
> crud it uses instead? does splitdrive work on that? so maybe sys.platform ==
> 'win32' makes more sense. (other places use sys.platform, so if you think it
> works on cygwin, then sys.platform in ('win32', 'cygwin') for consistency)

Changed to use sys.platform == "win32".  I just verified that this isn't
necessary on cygwin, since relative paths do work between drives, because of the
/cygdrive prefix and Unix-style paths.

[borenet, 2013-07-22 16:49:31]

The trybot is succeeding as of patch set 16:
http://build.chromium.org/p/tryserver.nacl/builders/gyp-win32/builds/1115

PTAL

[scottmg, 2013-07-22 16:56:36]

lgtm, but try a gyp-win64 too as win32 doesn't do a VS build.

[borenet, 2013-07-22 18:47:59]

On 2013/07/22 16:56:36, scottmg wrote:
> lgtm, but try a gyp-win64 too as win32 doesn't do a VS build.

gyp-win64 succeeds as of patch set 19.

[borenet, 2013-07-22 19:48:15]

Could someone submit this for me?  I suspect the linux trybot failure is because
I'm in an SVN checkout...

[scottmg, 2013-07-22 20:21:07]

On 2013/07/22 19:48:15, borenet wrote:
> Could someone submit this for me?  I suspect the linux trybot failure is
because
> I'm in an SVN checkout...

(Sorry, I submitted that linux_clang try job accidentally on the wrong CL)

I tried to patch it in, but it's complaining that 

error: test/win/generator-output-different-drive/prog.c: No such file or
directory
error: test/win/generator-output-different-drive/prog.gyp: No such file or
directory

as if it's trying to diff from that file that doesn't exist. Was this in a svn
or git checkout?

[scottmg, 2013-07-22 20:30:57]

Bleh, I dunno. I recreated the files since they were small. Landed at r1683.

[borenet, 2013-07-22 20:33:08]

On 2013/07/22 20:30:57, scottmg wrote:
> Bleh, I dunno. I recreated the files since they were small. Landed at r1683.

Thanks!