823900 - Rename "Android" builds to "Android 2.2"

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Ed Morley [:emorley]]

In bug 800047, the format for our Panda Android 4.0 builds was decided as (surprisingly) "Android 4.0".

It was also proposed that at a later point the current "Android {'','no-ionmonkey',armv6'}" be renamed to "Android 2.2 {...}".

TBPL's regexes are ready for this change:
{
        /android(?: 2\.2)? no-ionmonkey/i.test(name) ? "android-2-2-noion" :
        /android(?: 2\.2)? armv6/i.test(name) ? "android-2-2-armv6" :
        /android(?: 2\.2)? x86/i.test(name) ? "android-2-2-x86" :
        /android(?: 2\.2)?/i.test(name) ? "android-2-2" :
}

We just need to change this in buildbot-configs/custom.

Comment 1

[Chris AtLee [:catlee]]

This is pretty easy - just change the base_name in the build platforms (e.g. http://hg.mozilla.org/build/buildbot-configs/file/default/mozilla/config.py#l905) and for tests (http://hg.mozilla.org/build/buildbot-configs/file/default/mozilla-tests/mobile_config.py#l72)

Comment 2

[Ed Morley [:emorley]]

Ryan, up for taking a look at this? Hopefully is just a case of changing the two strings listed in comment 0 :-)

Comment 3

[Ryan VanderMeulen [:RyanVM]]

Sure

Comment 4

[Ryan VanderMeulen [:RyanVM]]

Attachment: add Android version number to all Android builds and tests

This patch:
1.) Adds "2.2" to all version names and drops the "250" part off the Tegras.
2.) Adds "4.2" to the Android x86 job for consistency's sake. Geoff, the feedback? request is just to confirm that 4.2 is correct and you're OK with this change.

Can't say I'm a huge fan of how no-ion builds present themselves, but given that their time is coming to an end, I can't say I care that much either.

Comment 5

[Ryan VanderMeulen [:RyanVM]]

Ed, assuming you don't have an issue with these changes, I will file a TBPL bug for tweaking the x86 regex a bit to support both the "with" and "without" 4.2 versions until we only need to support the "with" case.

Comment 6

[Geoff Brown [:gbrown]]

Comment on attachment 801977 [details] [diff] [review]
add Android version number to all Android builds and tests

Review of attachment 801977 [details] [diff] [review]:
-----------------------------------------------------------------

Yes, the x86 emulators are running Android 4.2. (I hope to upgrade them to 4.3, but I doubt that will happen this year.)

I like all these names.

Comment 7

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/build/buildbot-configs/rev/261eacaccb8d

Comment 8

[Rail Aliiev [:rail]]

in production

Comment 9

[Justin Wood (:Callek)]

Attachment: [braindump] build faster report

Coop, is there anything more than "checkin" I need to do to deploy this?

Comment 10

[Ryan VanderMeulen [:RyanVM]]

Comment on attachment 802531 [details] [diff] [review]
[braindump] build faster report

Android 4.2 x86 is a lower-case X86, in case that matters.

Comment 11

[Justin Wood (:Callek)]

Attachment: [buildapi] patch for buildapi (used in reports)

Comment 12

[Justin Wood (:Callek)]

Attachment: [slave_health] add x86 while here

I peeked into slave health for if I needed to change stuff, and saw missing x86 jobs, so added them, didn't need to tweak panda or tegra ones

Comment 13

[Chris Cooper [:coop] (he/him)]

Comment on attachment 802626 [details] [diff] [review]
[slave_health] add x86 while here

https://hg.mozilla.org/users/coop_mozilla.com/slave_health/rev/a63b48bfbea8

Comment 14

[Chris Cooper [:coop] (he/him)]

Comment on attachment 802531 [details] [diff] [review]
[braindump] build faster report

Review of attachment 802531 [details] [diff] [review]:
-----------------------------------------------------------------

::: reports/buildfaster_report.py
@@ +80,5 @@
> +        ('android_armv6', ['Android Armv6',
> +                           'Android 2.2 Armv6']),
> +        ('android_armv7', ['Android Tegra ',
> +                           'Android 2.2 Tegra ']),
> +        ('android_x86', ['Android X86',

Small nit: can we lowercase the 'X' in x86, please? (On the following line too)

Comment 15

[Chris Cooper [:coop] (he/him)]

Comment on attachment 802551 [details] [diff] [review]
[buildapi] patch for buildapi (used in reports)

Review of attachment 802551 [details] [diff] [review]:
-----------------------------------------------------------------

Wrong patch? Looks identical to https://bugzilla.mozilla.org/attachment.cgi?id=802531

Comment 16

[Justin Wood (:Callek)]

Attachment: [buildapi] patch for buildapi (used in reports)

bah indeed, this is the right patch

Comment 17

[Justin Wood (:Callek)]

Comment on attachment 802782 [details] [diff] [review]
[buildapi] patch for buildapi (used in reports)

https://hg.mozilla.org/build/buildapi/rev/f490db921804

Comment 18

[Justin Wood (:Callek)]

Comment on attachment 802531 [details] [diff] [review]
[braindump] build faster report

https://hg.mozilla.org/build/braindump/rev/b4fb9d50a87b

And deployed this.

Comment 19

[Nick Thomas [:nthomas] (UTC+12)]

Attachment: Busted wait report

(In reply to Justin Wood (:Callek) from comment #17)
> Comment on attachment 802782 [details] [diff] [review]
> [buildapi] patch for buildapi (used in reports)
> 
> https://hg.mozilla.org/build/buildapi/rev/f490db921804

This appears to have broken the test wait time report
* no tegra platform in break down
* replaced (?) by other platform with 8700+ jobs
* very long list for |Other platforms listed under 'other' platforms|

Ask me on IRC if you'd like help testing these types of changes.

Comment 20

[Justin Wood (:Callek)]

(In reply to Nick Thomas [:nthomas] from comment #19)
> Created attachment 804189 [details]
> Busted wait report
> 
> This appears to have broken the test wait time report

So looking at the patches in this bug, the existing report and comparing I realize the following:

* I did not break this
* it is indeed broken
* My patch was just incomplete due to not making '250' at end of string optional

Furthermore we realize the following
* We should have escaped . in the regex
* Since newly run reports with this script will not get the old names, we don't need to maintain backwards compat anymore.
* there was a Mac regex weirdness as well that I am going to fix here too.

Comment 21

[Justin Wood (:Callek)]

Attachment: [buildapi] part 2 (do it right)

Comment 22

[Nick Thomas [:nthomas] (UTC+12)]

Comment on attachment 804207 [details] [diff] [review]
[buildapi] part 2 (do it right)

Review of attachment 804207 [details] [diff] [review]:
-----------------------------------------------------------------

::: buildapi/model/util.py
@@ +23,4 @@
>          re.compile('^b2g .+_armv7a.+',  re.IGNORECASE),
>          re.compile('^b2g linux32_gecko .+',  re.IGNORECASE),
>          re.compile('^b2g_((?!(test|talos)).)+$',  re.IGNORECASE),
> +        re.compile('^Android (?!(?:2\.2 Tegra|2\.2 Armv6 Tegra|2\.2 no-ionmonkey Tegra 250|4\.0 Panda|4\.2 x86 Emulator)).+'),

Looks like you should zap the ' 250' here too, r+ with that.

Comment 23

[Justin Wood (:Callek)]

(In reply to Nick Thomas [:nthomas] from comment #22)
> Comment on attachment 804207 [details] [diff] [review]
> [buildapi] part 2 (do it right)
> 
> Looks like you should zap the ' 250' here too, r+ with that.

Done on checkin: https://hg.mozilla.org/build/buildapi/rev/c7703fc7deb3

This bug also looks finished now