Update freetype sources (2.6.2) in third party.

Update freetype sources (version 2.6.2) in third party.

Required by the Newlib PNaCl toolchain when building skia,
as the native system freetype cannot be used.

Everything in "third_party/freetype2/src" is unmodified
from the original tarball.

BUG=#431
R=viettrungluu@chromium.org

Committed: https://chromium.googlesource.com/external/mojo/+/e57e4a3a1de3b6c4f6fc0af18cf25179be4c404e

[Sean Klein, 2015-12-14 22:54:14]

Description was changed from

==========
Added freetype sources (2.6.2) to third party.

BUG=#594
==========

to

==========
Added freetype sources (version 2.6.2) to third party.

Required by the Newlib PNaCl toolchain when building skia,
as the native system freetype cannot be used.

BUG=#594
==========

[Sean Klein, 2015-12-14 22:55:04]

Description was changed from

==========
Added freetype sources (version 2.6.2) to third party.

Required by the Newlib PNaCl toolchain when building skia,
as the native system freetype cannot be used.

BUG=#594
==========

to

==========
Added freetype sources (version 2.6.2) to third party.

Required by the Newlib PNaCl toolchain when building skia,
as the native system freetype cannot be used.

Everything in "third_party/freetype2/src" is unmodified
from the original tarball.

BUG=#594
==========

[Sean Klein, 2015-12-15 00:06:33]

Description was changed from

==========
Added freetype sources (version 2.6.2) to third party.

Required by the Newlib PNaCl toolchain when building skia,
as the native system freetype cannot be used.

Everything in "third_party/freetype2/src" is unmodified
from the original tarball.

BUG=#594
==========

to

==========
Added freetype sources (version 2.6.2) to third party.

Required by the Newlib PNaCl toolchain when building skia,
as the native system freetype cannot be used.

Everything in "third_party/freetype2/src" is unmodified
from the original tarball.

BUG=#431
==========

[Sean Klein, 2015-12-15 00:10:52]

smklein@chromium.org changed reviewers:
+ phosek@chromium.org, viettrungluu@chromium.org

[viettrungluu, 2015-12-15 00:21:19]

viettrungluu@chromium.org changed reviewers:
+ jamesr@chromium.org, kulakowski@chromium.org

[viettrungluu, 2015-12-15 00:21:22]

Adding jamesr & kulakowski....

So, I was thinking about doing this today:
* Probably, on Linux, we should also our own freetype2, rather than the system
one.
* Possibly, on Android, we should use the same freetype2. (Advantage: sanity.
Disadvantage: rendering may not match the system.)
* Don't we already have (some older version of) freetype2 in third_party? (Now,
it happens to be unused, which is pretty crazy....)
* But we should probably find a way of sharing freetype2 with, e.g., flutter.
* So maybe freetype2 should live in its own repo and shared the jiri-way
(whatever that is), in future-world?

[Sean Klein, 2015-12-15 01:36:57]

On 2015/12/15 00:21:22, viettrungluu wrote:
> Adding jamesr & kulakowski....
> 
> So, I was thinking about doing this today:
> * Probably, on Linux, we should also our own freetype2, rather than the system
> one.
> * Possibly, on Android, we should use the same freetype2. (Advantage: sanity.
> Disadvantage: rendering may not match the system.)
> * Don't we already have (some older version of) freetype2 in third_party?
(Now,
> it happens to be unused, which is pretty crazy....)
> * But we should probably find a way of sharing freetype2 with, e.g., flutter.
> * So maybe freetype2 should live in its own repo and shared the jiri-way
> (whatever that is), in future-world?

The older version of freetype2 was left over from copying and pasting Chromium.
It hadn't been updated in over a year, and I can't find any targets that used
the old version.
(That's why I replaced the "third_party/freetype2" directory to be something
useful)

I don't have any strong feelings about where the freetype2 code lives, but I
need the
freetype2 BUILD.gn target to be visible as a static library to build Skia with
PNaCl.

[kulakowski1, 2015-12-15 17:42:50]

On 2015/12/15 00:21:22, viettrungluu wrote:
> Adding jamesr & kulakowski....
> 
> So, I was thinking about doing this today:
> * Probably, on Linux, we should also our own freetype2, rather than the system
> one.
> * Possibly, on Android, we should use the same freetype2. (Advantage: sanity.
> Disadvantage: rendering may not match the system.)
> * Don't we already have (some older version of) freetype2 in third_party?
(Now,
> it happens to be unused, which is pretty crazy....)
> * But we should probably find a way of sharing freetype2 with, e.g., flutter.
> * So maybe freetype2 should live in its own repo and shared the jiri-way
> (whatever that is), in future-world?

Yes, I think in future-world there should be one place for freetype, with
perhaps a different revision in an Android manifest if that becomes an important
concern.

This seems as good a way as any to me to pick the freetype we all use in the
future. *shrug*

[viettrungluu, 2015-12-15 21:15:36]

On 2015/12/15 17:42:50, kulakowski1 wrote:
> On 2015/12/15 00:21:22, viettrungluu wrote:
> > Adding jamesr & kulakowski....
> > 
> > So, I was thinking about doing this today:
> > * Probably, on Linux, we should also our own freetype2, rather than the
system
> > one.
> > * Possibly, on Android, we should use the same freetype2. (Advantage:
sanity.
> > Disadvantage: rendering may not match the system.)
> > * Don't we already have (some older version of) freetype2 in third_party?
> (Now,
> > it happens to be unused, which is pretty crazy....)
> > * But we should probably find a way of sharing freetype2 with, e.g.,
flutter.
> > * So maybe freetype2 should live in its own repo and shared the jiri-way
> > (whatever that is), in future-world?
> 
> Yes, I think in future-world there should be one place for freetype, with
> perhaps a different revision in an Android manifest if that becomes an
important
> concern.
> 
> This seems as good a way as any to me to pick the freetype we all use in the
> future. *shrug*

OK, let's do this. It should probably be reorganized later (into its own repo,
or whatever), but for now I guess it's OK.

(I was confused by the fact that the files are nearly all "adds" with almost no
"modifies".)

Sean, could you:
* Update the CL description to say "update" or "replace" instead of "add".
* Update the README.chromium file.
* Update or remove the freetype2.patch file. (Is this exactly upstream freetype2
2.6.2, or did you make changes?)

To do later:
* Switch to use this freetype2 on Linux and FNL, instead of the "system's".
* Ditto for Android (after investigating the effects of this).

[Sean Klein, 2015-12-15 21:22:43]

Description was changed from

==========
Added freetype sources (version 2.6.2) to third party.

Required by the Newlib PNaCl toolchain when building skia,
as the native system freetype cannot be used.

Everything in "third_party/freetype2/src" is unmodified
from the original tarball.

BUG=#431
==========

to

==========
Update freetype sources (version 2.6.2) in third party.

Required by the Newlib PNaCl toolchain when building skia,
as the native system freetype cannot be used.

Everything in "third_party/freetype2/src" is unmodified
from the original tarball.

BUG=#431
==========

[Sean Klein, 2015-12-15 21:38:31]

On 2015/12/15 21:15:36, viettrungluu wrote:
> On 2015/12/15 17:42:50, kulakowski1 wrote:
> > On 2015/12/15 00:21:22, viettrungluu wrote:
> > > Adding jamesr & kulakowski....
> > > 
> > > So, I was thinking about doing this today:
> > > * Probably, on Linux, we should also our own freetype2, rather than the
> system
> > > one.
> > > * Possibly, on Android, we should use the same freetype2. (Advantage:
> sanity.
> > > Disadvantage: rendering may not match the system.)
> > > * Don't we already have (some older version of) freetype2 in third_party?
> > (Now,
> > > it happens to be unused, which is pretty crazy....)
> > > * But we should probably find a way of sharing freetype2 with, e.g.,
> flutter.
> > > * So maybe freetype2 should live in its own repo and shared the jiri-way
> > > (whatever that is), in future-world?
> > 
> > Yes, I think in future-world there should be one place for freetype, with
> > perhaps a different revision in an Android manifest if that becomes an
> important
> > concern.
> > 
> > This seems as good a way as any to me to pick the freetype we all use in the
> > future. *shrug*
> 
> OK, let's do this. It should probably be reorganized later (into its own repo,
> or whatever), but for now I guess it's OK.
> 
> (I was confused by the fact that the files are nearly all "adds" with almost
no
> "modifies".)
> 
> Sean, could you:
> * Update the CL description to say "update" or "replace" instead of "add".
> * Update the README.chromium file.
> * Update or remove the freetype2.patch file. (Is this exactly upstream
freetype2
> 2.6.2, or did you make changes?)
> 
> To do later:
> * Switch to use this freetype2 on Linux and FNL, instead of the "system's".
> * Ditto for Android (after investigating the effects of this).

* CL description changed
* README.chromium file updated (should I identify freetype2 as security
critical? This doesn't seem exclusive to tests)
* The "freetype2.patch" file is needed for the instructions in the
README.chromium file, and will be used when updating to a new freetype version.
By keeping this patch around, everything in "third_party/freetype2/src" is
unmodified.

[viettrungluu, 2015-12-15 21:57:46]

On 2015/12/15 21:38:31, smklein1 wrote:
> On 2015/12/15 21:15:36, viettrungluu wrote:
> > On 2015/12/15 17:42:50, kulakowski1 wrote:
> > > On 2015/12/15 00:21:22, viettrungluu wrote:
> > > > Adding jamesr & kulakowski....
> > > > 
> > > > So, I was thinking about doing this today:
> > > > * Probably, on Linux, we should also our own freetype2, rather than the
> > system
> > > > one.
> > > > * Possibly, on Android, we should use the same freetype2. (Advantage:
> > sanity.
> > > > Disadvantage: rendering may not match the system.)
> > > > * Don't we already have (some older version of) freetype2 in
third_party?
> > > (Now,
> > > > it happens to be unused, which is pretty crazy....)
> > > > * But we should probably find a way of sharing freetype2 with, e.g.,
> > flutter.
> > > > * So maybe freetype2 should live in its own repo and shared the jiri-way
> > > > (whatever that is), in future-world?
> > > 
> > > Yes, I think in future-world there should be one place for freetype, with
> > > perhaps a different revision in an Android manifest if that becomes an
> > important
> > > concern.
> > > 
> > > This seems as good a way as any to me to pick the freetype we all use in
the
> > > future. *shrug*
> > 
> > OK, let's do this. It should probably be reorganized later (into its own
repo,
> > or whatever), but for now I guess it's OK.
> > 
> > (I was confused by the fact that the files are nearly all "adds" with almost
> no
> > "modifies".)
> > 
> > Sean, could you:
> > * Update the CL description to say "update" or "replace" instead of "add".
> > * Update the README.chromium file.
> > * Update or remove the freetype2.patch file. (Is this exactly upstream
> freetype2
> > 2.6.2, or did you make changes?)
> > 
> > To do later:
> > * Switch to use this freetype2 on Linux and FNL, instead of the "system's".
> > * Ditto for Android (after investigating the effects of this).
> 
> * CL description changed
> * README.chromium file updated (should I identify freetype2 as security
> critical? This doesn't seem exclusive to tests)

Yes, I think so.

> * The "freetype2.patch" file is needed for the instructions in the
> README.chromium file, and will be used when updating to a new freetype
version.

I don't think we need these changes if we're not building as a shared library.

> By keeping this patch around, everything in "third_party/freetype2/src" is
> unmodified.

[viettrungluu, 2015-12-15 21:58:10]

https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/B...
File third_party/freetype2/BUILD.gn (right):

https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/B...
third_party/freetype2/BUILD.gn:12: static_library("freetype2") {
Maybe just "source_set"?

https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/R...
File third_party/freetype2/README.chromium (right):

https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/R...
third_party/freetype2/README.chromium:2: URL:
http://download.savannah.gnu.org/releases/freetype/
I'd say "https", but their https is busted. D'oh. :-(

https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/R...
third_party/freetype2/README.chromium:6: License File: NOT_SHIPPED
Probably you should update this (to point to the LICENSE.TXT file).

[Sean Klein, 2015-12-15 22:32:25]

- Updated README to identify freetype2 as security critical
- Deleted freetype2.patch, undid impact of patch from include file, updated
README steps for updating to new freetype version.

Also, as a warning, I'm getting the following presubmit error:

"
Found ifdef check on always-defined macro[s]. Please fix your code
or check the list of ALWAYS_DEFINED_MACROS in src/PRESUBMIT.py.
      third_party/freetype2/src/src/gzip/zconf.h:218  TARGET_OS_MAC is always
defined. 
	Did you mean '#if TARGET_OS_MAC'? \
      third_party/freetype2/src/src/gzip/zutil.h:108  TARGET_OS_MAC is always
defined. 
	Did you mean '#if TARGET_OS_MAC'?
"


Should I modify these files to fix this issue? Or leave it the same as upstream
freetype?

https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/B...
File third_party/freetype2/BUILD.gn (right):

https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/B...
third_party/freetype2/BUILD.gn:12: static_library("freetype2") {
On 2015/12/15 21:58:10, viettrungluu wrote:
> Maybe just "source_set"?

Done.

https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/R...
File third_party/freetype2/README.chromium (right):

https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/R...
third_party/freetype2/README.chromium:2: URL:
http://download.savannah.gnu.org/releases/freetype/
On 2015/12/15 21:58:10, viettrungluu wrote:
> I'd say "https", but their https is busted. D'oh. :-(

Acknowledged :(

https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/R...
third_party/freetype2/README.chromium:6: License File: NOT_SHIPPED
On 2015/12/15 21:58:10, viettrungluu wrote:
> Probably you should update this (to point to the LICENSE.TXT file).

Done.

[viettrungluu, 2015-12-15 22:57:25]

On 2015/12/15 22:32:25, smklein1 wrote:
> - Updated README to identify freetype2 as security critical
> - Deleted freetype2.patch, undid impact of patch from include file, updated
> README steps for updating to new freetype version.
> 
> Also, as a warning, I'm getting the following presubmit error:
> 
> "
> Found ifdef check on always-defined macro[s]. Please fix your code
> or check the list of ALWAYS_DEFINED_MACROS in src/PRESUBMIT.py.
>       third_party/freetype2/src/src/gzip/zconf.h:218  TARGET_OS_MAC is always
> defined. 
> 	Did you mean '#if TARGET_OS_MAC'? \
>       third_party/freetype2/src/src/gzip/zutil.h:108  TARGET_OS_MAC is always
> defined. 
> 	Did you mean '#if TARGET_OS_MAC'?
> "
> 
> 
> Should I modify these files to fix this issue? Or leave it the same as
upstream
> freetype?

I guess it's OK to leave it alone? (Maybe get upstream to fix their stuff? I'm
rather horrified to see the ancient version of stuff they have in that
directory.)

(It'd be easier to accept/deal with modifications if we had this stuff in its
own repo. If we want to modify it, we should probably check in the exact
upstream version, and then check in fixes separately. Then at least it'd be a
properly versioned fork.)

> 
>
https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/B...
> File third_party/freetype2/BUILD.gn (right):
> 
>
https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/B...
> third_party/freetype2/BUILD.gn:12: static_library("freetype2") {
> On 2015/12/15 21:58:10, viettrungluu wrote:
> > Maybe just "source_set"?
> 
> Done.
> 
>
https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/R...
> File third_party/freetype2/README.chromium (right):
> 
>
https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/R...
> third_party/freetype2/README.chromium:2: URL:
> http://download.savannah.gnu.org/releases/freetype/
> On 2015/12/15 21:58:10, viettrungluu wrote:
> > I'd say "https", but their https is busted. D'oh. :-(
> 
> Acknowledged :(
> 
>
https://codereview.chromium.org/1524973002/diff/20001/third_party/freetype2/R...
> third_party/freetype2/README.chromium:6: License File: NOT_SHIPPED
> On 2015/12/15 21:58:10, viettrungluu wrote:
> > Probably you should update this (to point to the LICENSE.TXT file).
> 
> Done.

[viettrungluu, 2015-12-15 22:59:22]

lgtm

[Sean Klein, 2015-12-15 23:06:48]

Description was changed from

==========
Update freetype sources (version 2.6.2) in third party.

Required by the Newlib PNaCl toolchain when building skia,
as the native system freetype cannot be used.

Everything in "third_party/freetype2/src" is unmodified
from the original tarball.

BUG=#431
==========

to

==========
Update freetype sources (version 2.6.2) in third party.

Required by the Newlib PNaCl toolchain when building skia,
as the native system freetype cannot be used.

Everything in "third_party/freetype2/src" is unmodified
from the original tarball.

BUG=#431
R=viettrungluu@chromium.org

Committed:
https://chromium.googlesource.com/external/mojo/+/e57e4a3a1de3b6c4f6fc0af18cf...
==========

[Sean Klein, 2015-12-15 23:06:53]

Committed patchset #4 (id:30001) manually as
e57e4a3a1de3b6c4f6fc0af18cf25179be4c404e.