Add clang-format binaries and scripts for all platforms

Add clang-format binaries and scripts for all platforms.

The binaries are hosted in Google storage, and pulled in
via a runhook.

Each included binary is roughly 1.1MB. The binaries referenced here were built
from clang at revision 198831.

On the wiki, I've started instructions for how to update the
binaries in the future:
https://code.google.com/p/chromium/wiki/UpdatingClangFormatBinaries

Note that clang_format/scripts are copied here only as an
expedience; once we have a proper mirror set up, we'll pull these via DEPS

BUG=240309
TBR=brettw,phajdan.jr

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=244372

[Nico, 2013-12-12 22:51:21]

\o/

Needs a third_party/clang_format/LICENSE file, other than that lgtm.

How large are the binaries?

Do you have instructions on how to update them somewhere (something like
https://code.google.com/p/chromium/wiki/UpdatingClang except simpler)?

Mention in the CL description which revision the binaries were built at. (As far
as I know, `clang-format --version` doesn't say)

I wonder if we should pull the scripts via DEPS to make clear that we have no
downstream deps for them. (We already have an llvm svn mirror)

[Nico, 2013-12-12 22:51:34]

https://codereview.chromium.org/111953007/diff/20001/.gitignore
File .gitignore (right):

https://codereview.chromium.org/111953007/diff/20001/.gitignore#newcode209
.gitignore:209: !/third_party/clang_format/bin/*/*.sha1
Is this needed? gn doesn't have this

[ncarter (slow), 2013-12-12 23:06:45]

Just responding, no fixes just yet -- the Windows binary (build with VS2013
using the MinSizeRel config) is 1.5MB. The linux and mac binaries are something
like 11 or 13 megs, which makes me unhappy. I have a hunch that if we do a CMake
build there will be a knob-equivalent to the windows MinSizeRel, that'll
hopefully drop the sizes on mac/linux.

I'll do another cycle and generate fresh binaries, play around with the above,
and document the procedure. Albert actually generated the Linux binaries, and
the Win binaries are at an earlier revision than the others.

https://codereview.chromium.org/111953007/diff/20001/.gitignore
File .gitignore (right):

https://codereview.chromium.org/111953007/diff/20001/.gitignore#newcode209
.gitignore:209: !/third_party/clang_format/bin/*/*.sha1
On 2013/12/12 22:51:35, Nico wrote:
> Is this needed? gn doesn't have this

gn excludes everything, meaning there's a gitignore for the tracked .sha1 files
under gn. 'git add' gives warnings if you're editing ignored files, even if
they're already tracked:

"""
  D:\src\crgit2\src\tools\gn\bin\linux>git add gn.sha1
  The following paths are ignored by one of your .gitignore files:
tools/gn/bin/linux
Use -f if you really want to add them.
fatal: no files added
"""

The pattern I used avoids this, provided the wildcard */*/ doesn't somehow slow
down common git commands (I don't have evidence either way). If you concur I'd
be happy to update the gn rule.

[ncarter (slow), 2014-01-10 01:18:41]

PTAL

[Nico, 2014-01-10 01:27:41]

lgtm

Please add to the CL description that we want to pull the scripts from deps
eventually once a mirror exists (with a link to that bug)

Thanks! Woot!

[awong, 2014-01-10 01:28:00]

LGTM

Comments in clang-format-diff.py apparently should go to upstream. I'm writing
them here, but feel to ignore for this review.

https://codereview.chromium.org/111953007/diff/100001/third_party/clang_forma...
File third_party/clang_format/scripts/clang-format-diff.py (right):

https://codereview.chromium.org/111953007/diff/100001/third_party/clang_forma...
third_party/clang_format/scripts/clang-format-diff.py:29: import StringIO
sort order? If you're doing case-insenstive, StringIO should sitll be above
subprocess.

https://codereview.chromium.org/111953007/diff/100001/third_party/clang_forma...
third_party/clang_format/scripts/clang-format-diff.py:63: match =
re.search('^\+\+\+\ (.*?/){%s}(\S*)' % args.p, line)
Can you add a comment with an example of what the regexp is trying to match?
Same comment for other regexps.

https://codereview.chromium.org/111953007/diff/100001/third_party/clang_forma...
third_party/clang_format/scripts/clang-format-diff.py:66: if filename == None:
filename is None?

http://google-styleguide.googlecode.com/svn/trunk/pyguide.html?showone=True/F...

[ncarter (slow), 2014-01-10 07:01:11]

I tested the current version out on all three platforms, so I think this is good
to go.

+brettw for third_party OWNERS

[Nico, 2014-01-10 20:34:52]

On 2014/01/10 07:01:11, ncarter wrote:
> I tested the current version out on all three platforms, so I think this is
good
> to go.
> 
> +brettw for third_party OWNERS

brettw tells me it's ok to TBR him.

[commit-bot: I haz the power, 2014-01-10 20:53:34]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/nick@chromium.org/111953007/100001

[commit-bot: I haz the power, 2014-01-10 23:02:22]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/nick@chromium.org/111953007/100001

[Nico, 2014-01-10 23:48:23]

:-(

https://codereview.chromium.org/111953007/diff/100001/third_party/clang_forma...
File third_party/clang_format/README.chromium (right):

https://codereview.chromium.org/111953007/diff/100001/third_party/clang_forma...
third_party/clang_format/README.chromium:7: License: NCSA
I'm afraid you also need a
  License File: NOT_SHIPPED
line, else the check_license step is going to complain :-(

Also, checklicenses.py contains "University of Illinois/NCSA Open Source License
(BSD like)", so that should probably be what's on the License: line (but that
doesn't seem to matter, according to `tools/licenses.py scan`)

[ncarter (slow), 2014-01-11 01:06:24]

Such boilerplate, very enforcing, woww.

I'll have a new patch up shortly.


On Fri, Jan 10, 2014 at 3:48 PM, <thakis@chromium.org> wrote:

> :-(
>
>
> https://codereview.chromium.org/111953007/diff/100001/
> third_party/clang_format/README.chromium
> File third_party/clang_format/README.chromium (right):
>
> https://codereview.chromium.org/111953007/diff/100001/
> third_party/clang_format/README.chromium#newcode7
> third_party/clang_format/README.chromium:7: License: NCSA
> I'm afraid you also need a
>   License File: NOT_SHIPPED
> line, else the check_license step is going to complain :-(
>
> Also, checklicenses.py contains "University of Illinois/NCSA Open Source
> License (BSD like)", so that should probably be what's on the License:
> line (but that doesn't seem to matter, according to `tools/licenses.py
> scan`)
>
> https://codereview.chromium.org/111953007/
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[commit-bot: I haz the power, 2014-01-11 02:06:54]

Retried try job too often on android_dbg for step(s) slave_steps
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=android_db...

[ncarter (slow), 2014-01-11 03:23:47]

I've updated README.chromium, and added a licensecheck.py suppression (as the
editor scripts don't have copyright headers)

Nico, PTAL and feel free to CQ if it LGTY.

[ncarter (slow), 2014-01-11 03:26:57]

On 2014/01/11 03:23:47, ncarter wrote:
> I've updated README.chromium, and added a licensecheck.py suppression (as the
> editor scripts don't have copyright headers)
> 
> Nico, PTAL and feel free to CQ if it LGTY.

Oh also, I got rid of LICENSE.txt because I figured it wouldn't be there when we
pull in via DEPS. I can add a copy back in. It's not clear which is the right
thing -- whether 'NOT_SHIPPED' is meant to mean "the project has no such file"
or "it's not shipped with the subset of the source we're concerned with here"

[Nico, 2014-01-11 03:28:46]

On 2014/01/11 03:23:47, ncarter wrote:
> I've updated README.chromium, and added a licensecheck.py suppression (as the
> editor scripts don't have copyright headers)

Is that necessary with the "License File: NOT_SHIPPED" bit? Oh well.

phajdan: I'm going to file an llvm bug for license headers in these scripts
right now and reply here with the link.


lgtm

[ncarter (slow), 2014-01-11 03:29:35]

On 2014/01/11 03:26:57, ncarter wrote:
> On 2014/01/11 03:23:47, ncarter wrote:
> > I've updated README.chromium, and added a licensecheck.py suppression (as
the
> > editor scripts don't have copyright headers)
> > 
> > Nico, PTAL and feel free to CQ if it LGTY.
> 
> Oh also, I got rid of LICENSE.txt because I figured it wouldn't be there when
we
> pull in via DEPS. I can add a copy back in. It's not clear which is the right
> thing -- whether 'NOT_SHIPPED' is meant to mean "the project has no such file"
> or "it's not shipped with the subset of the source we're concerned with here"

Ah got it: """License File: (OPTIONAL) File that contains a copy of the
package's license. Use the special value NOT_SHIPPED to indicate that the
package is not included in the shipped product, so its license does not need to
be included in about:credits and no license file is required."""

[Nico, 2014-01-11 03:30:32]

Filed http://llvm.org/bugs/show_bug.cgi?id=18442

[commit-bot: I haz the power, 2014-01-11 03:32:40]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/nick@chromium.org/111953007/610001

[commit-bot: I haz the power, 2014-01-11 22:40:11]

Change committed as 244372

[cmp, 2014-01-11 23:51:43]

A revert of this CL has been created in
https://codereview.chromium.org/135773002/ by cmp@chromium.org.

The reason for reverting is: Broke Linux build in
http://build.chromium.org/p/chromium/builders/Linux/builds/46507..

[cmp, 2014-01-11 23:56:29]

In failing build, check_perms failed with this output:


Found subversion checkout at /b/build/slave/Linux/build/src
Processed 53854 files, 5431 files where tested for shebang/ELF header

FAILED

/b/build/slave/Linux/build/src/third_party/clang_format/scripts/clang-format-diff.py:
Has shebang but not executable bit

[Nico, 2014-01-12 00:14:06]

Yeah, I had fixed this in
http://src.chromium.org/viewvc/chrome?revision=244392&view=revision .
Please check if stuff is already fixed on trunk before reverting.


On Sat, Jan 11, 2014 at 3:56 PM, <cmp@chromium.org> wrote:

> In failing build, check_perms failed with this output:
>
>
> Found subversion checkout at /b/build/slave/Linux/build/src
> Processed 53854 files, 5431 files where tested for shebang/ELF header
>
> FAILED
>
> /b/build/slave/Linux/build/src/third_party/clang_format/
> scripts/clang-format-diff.py:
> Has shebang but not executable bit
>
>
> https://codereview.chromium.org/111953007/
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[Nico, 2014-01-12 00:14:12]

(And please revert your revert.)


On Sat, Jan 11, 2014 at 4:14 PM, Nico Weber <thakis@chromium.org> wrote:

> Yeah, I had fixed this in
> http://src.chromium.org/viewvc/chrome?revision=244392&view=revision .
> Please check if stuff is already fixed on trunk before reverting.
>
>
> On Sat, Jan 11, 2014 at 3:56 PM, <cmp@chromium.org> wrote:
>
>> In failing build, check_perms failed with this output:
>>
>>
>> Found subversion checkout at /b/build/slave/Linux/build/src
>> Processed 53854 files, 5431 files where tested for shebang/ELF header
>>
>> FAILED
>>
>> /b/build/slave/Linux/build/src/third_party/clang_format/
>> scripts/clang-format-diff.py:
>> Has shebang but not executable bit
>>
>>
>> https://codereview.chromium.org/111953007/
>>
>
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.