Adjusted to check language3[0] only and ignore the summary return.

Adjusted to check language3[0] only and ignore the summary return.
Currently the is_valid_language check is on the summary language.

I'm trying to fix what seems to be a bug introduced here:
https://codereview.chromium.org/1263613002

When that change switched from the summary language to language3[0] (The
description shows this was intentional, with "Edit CLD2 result to return top
language instead of summary language.")

When mcindy@ switched from summary language to top language, shouldn't they have
also used that language to validate unknownness? I don't understand why you
currently want to check the summary language when you're returning the
language3[0] as the result.

BUG=599940
Committed: https://crrev.com/80f25b767e1a4b060f085b82e32f016ea4db881e
Cr-Commit-Position: refs/heads/master@{#391255}

[rkaplow, 2016-04-01 18:55:16]

rkaplow@chromium.org changed reviewers:
+ andrewhayden@chromium.org

[Andrew Hayden (chromium.org), 2016-04-11 10:20:51]

I'd like to hear riesa@'s thoughts on this. I looked through the CLD2 code but
it isn't immediately obvious what the difference between the summary language
and languages[0] is, and there appears to be non-trivial calculation involved.

Why don't we want to use the summary language? Can you discuss more about why
this change is necessary?

[rkaplow, 2016-04-11 14:07:31]

On 2016/04/11 10:20:51, Andrew Hayden (chromium.org) wrote:
> I'd like to hear riesa@'s thoughts on this. I looked through the CLD2 code but
> it isn't immediately obvious what the difference between the summary language
> and languages[0] is, and there appears to be non-trivial calculation involved.
> 
> Why don't we want to use the summary language? Can you discuss more about why
> this change is necessary?

I'm trying to fix what seems to be a bug introduced here:
https://codereview.chromium.org/1263613002

When that change switched from the summary language to language3[0] (The
description shows this was intentional, with "Edit CLD2 result to return top
language instead of summary language.")

When mcindy@ switched from summary language to top language, shouldn't they have
also used that language to validate unknownness? I don't understand why you
currently want to check the summary language when you're returning the
language3[0] as the result.

[Andrew Hayden (chromium.org), 2016-04-11 14:12:46]

Thanks for the context! Can you put some of that into the related bug? This
makes more sense now.

I'm surprised there's no regression associated with this, but the nuance of
language detection makes it plausible. LGTM for finishing what micindy@'s
re-wiring.

[rkaplow, 2016-04-11 14:15:49]

On 2016/04/11 14:12:46, Andrew Hayden (chromium.org) wrote:
> Thanks for the context! Can you put some of that into the related bug? This
> makes more sense now.
> 
> I'm surprised there's no regression associated with this, but the nuance of
> language detection makes it plausible. LGTM for finishing what micindy@'s
> re-wiring.

Yes, sorry. I did have the context in the original bug's description, but forgot
to move that over.
Will add my comment to the description.

[rkaplow, 2016-04-11 14:16:05]

Description was changed from

==========
Adjusted to check language3[0] only and ignore the summary return.
Currently the is_valid_language check is on the summary language

BUG=599940
==========

to

==========
Adjusted to check language3[0] only and ignore the summary return.
Currently the is_valid_language check is on the summary language.

I'm trying to fix what seems to be a bug introduced here:
https://codereview.chromium.org/1263613002

When that change switched from the summary language to language3[0] (The
description shows this was intentional, with "Edit CLD2 result to return top
language instead of summary language.")

When mcindy@ switched from summary language to top language, shouldn't they have
also used that language to validate unknownness? I don't understand why you
currently want to check the summary language when you're returning the
language3[0] as the result.

BUG=599940
==========

[rkaplow, 2016-05-03 15:21:41]

The CQ bit was checked by rkaplow@chromium.org

[commit-bot: I haz the power, 2016-05-03 15:21:56]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1849323002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1849323002/1

[commit-bot: I haz the power, 2016-05-03 16:21:31]

Description was changed from

==========
Adjusted to check language3[0] only and ignore the summary return.
Currently the is_valid_language check is on the summary language.

I'm trying to fix what seems to be a bug introduced here:
https://codereview.chromium.org/1263613002

When that change switched from the summary language to language3[0] (The
description shows this was intentional, with "Edit CLD2 result to return top
language instead of summary language.")

When mcindy@ switched from summary language to top language, shouldn't they have
also used that language to validate unknownness? I don't understand why you
currently want to check the summary language when you're returning the
language3[0] as the result.

BUG=599940
==========

to

==========
Adjusted to check language3[0] only and ignore the summary return.
Currently the is_valid_language check is on the summary language.

I'm trying to fix what seems to be a bug introduced here:
https://codereview.chromium.org/1263613002

When that change switched from the summary language to language3[0] (The
description shows this was intentional, with "Edit CLD2 result to return top
language instead of summary language.")

When mcindy@ switched from summary language to top language, shouldn't they have
also used that language to validate unknownness? I don't understand why you
currently want to check the summary language when you're returning the
language3[0] as the result.

BUG=599940
==========

[commit-bot: I haz the power, 2016-05-03 16:21:32]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-05-03 16:23:30]

Description was changed from

==========
Adjusted to check language3[0] only and ignore the summary return.
Currently the is_valid_language check is on the summary language.

I'm trying to fix what seems to be a bug introduced here:
https://codereview.chromium.org/1263613002

When that change switched from the summary language to language3[0] (The
description shows this was intentional, with "Edit CLD2 result to return top
language instead of summary language.")

When mcindy@ switched from summary language to top language, shouldn't they have
also used that language to validate unknownness? I don't understand why you
currently want to check the summary language when you're returning the
language3[0] as the result.

BUG=599940
==========

to

==========
Adjusted to check language3[0] only and ignore the summary return.
Currently the is_valid_language check is on the summary language.

I'm trying to fix what seems to be a bug introduced here:
https://codereview.chromium.org/1263613002

When that change switched from the summary language to language3[0] (The
description shows this was intentional, with "Edit CLD2 result to return top
language instead of summary language.")

When mcindy@ switched from summary language to top language, shouldn't they have
also used that language to validate unknownness? I don't understand why you
currently want to check the summary language when you're returning the
language3[0] as the result.

BUG=599940

Committed: https://crrev.com/80f25b767e1a4b060f085b82e32f016ea4db881e
Cr-Commit-Position: refs/heads/master@{#391255}
==========

[commit-bot: I haz the power, 2016-05-03 16:23:31]

Patchset 1 (id:??) landed as
https://crrev.com/80f25b767e1a4b060f085b82e32f016ea4db881e
Cr-Commit-Position: refs/heads/master@{#391255}