incubator-callback-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Joe Bowser <bows...@gmail.com>
Subject Re: Chicken and the Egg: Proposed process for Corodova JS releases
Date Tue, 10 Apr 2012 19:30:29 GMT
I don't care about contacts right now, because those are at least failing
consistency across the four devices I'm testing.  What I do care about is
the fact that we're getting inconsistent tests across multiple Android
devices.

Samsung Galaxy S II (2.3.4): 9 Failures
Galaxy Nexus (4.0.2): 9 Failures
Motorola RAZR (2.3.5): 17 Failures
Samsung Nexus S (2.3.6): 23 Failures

All these devices were factory reset before we started testing them, and
DirectoryEntry and GeoLocation tests are timing out.  I'm OK with tagging
this, but this is something that needs to be looked into, and I'm wondering
if this is an issue with other platforms as well.

Also, I believe Simon mentioned that there was something he fixed in the JS
earlier.

On Tue, Apr 10, 2012 at 12:25 PM, Filip Maj <fil@adobe.com> wrote:

> It looks like the first contacts.save test fails because the contact
> returned in the save success callback is the wrong one.
>
> Looks like a native Android issue and not a JS issue.
>
> IMO JS and Docs can be tagged.
>
> On 4/10/12 11:55 AM, "Joe Bowser" <bowserj@gmail.com> wrote:
>
> >OK, I'm getting 9 failures on the Samsung Galaxy S II.  Testing appears to
> >be completely inconsistent.  I'm going to factory reset the Galaxy Nexus
> >and see if I get the same results.
> >
> >On Tue, Apr 10, 2012 at 11:40 AM, Joe Bowser <bowserj@gmail.com> wrote:
> >
> >> I'm getting the same thing on Gingerbread.  The thing is that on my
> >>Galaxy
> >> Nexus running 4.0.2, I'm only getting 13 failures.
> >>
> >>
> >> On Tue, Apr 10, 2012 at 11:40 AM, Filip Maj <fil@adobe.com> wrote:
> >>
> >>> Looks like ICS is having issues with saving a contact, but only a
> >>>couple
> >>> of the tests are failing in that. The round trip (heavy) test that
> >>>saves,
> >>> searches, removes, then searches again passes.. So.. Not sure.
> >>>
> >>> On 4/10/12 11:36 AM, "Joe Bowser" <bowserj@gmail.com> wrote:
> >>>
> >>> >I'm still testing Android 2.3.6, because that's what most people
> >>>have.  I
> >>> >do think that 21 tests is rather high for us to release, IMO.  Why
> >>>did it
> >>> >jump up like that?
> >>> >
> >>> >On Tue, Apr 10, 2012 at 11:35 AM, Filip Maj <fil@adobe.com> wrote:
> >>> >
> >>> >> All manual tests pass with latest js + framework commit on Android.
> >>>21
> >>> >> failed Qunit tests. 4.0.2 Galaxy Nexus.
> >>> >>
> >>> >> From what I can tell Shaz says iOS is good to go, Jesse says the
> >>>same
> >>> >>for
> >>> >> WP7. I know BB and Android are good.
> >>> >>
> >>> >> I'm going to tag the JS and update the docs with a 1.6.0 directory,
> >>> then
> >>> >> tag the docs.
> >>> >>
> >>> >> On 4/10/12 11:23 AM, "Filip Maj" <fil@adobe.com> wrote:
> >>> >>
> >>> >> >We have not committed anything new in cordova-js, we are just
> >>>picking
> >>> a
> >>> >> >new commit to tag to 1.6.0, so assuming all of us have been
working
> >>> >>with
> >>> >> >the cordova-js master in our platforms, we are not introducing
> >>> anything
> >>> >> >new.
> >>> >> >
> >>> >> >Every time any cordova developer touches the common code in
> >>>cordova-js
> >>> >> >that dev should be testing across all platforms. We have to
stop
> >>> >>working
> >>> >> >in our little native silos; that is not in the spirit of this
> >>>project.
> >>> >>We
> >>> >> >write a cross-platform tool, any of us need to be comfortable
> >>>testing
> >>> >>on
> >>> >> >all supported platforms.
> >>> >> >
> >>> >> >On 4/10/12 11:05 AM, "Joe Bowser" <bowserj@gmail.com>
wrote:
> >>> >> >
> >>> >> >>Sounds good. But if the changes somehow break Android,
what
> >>>happens?
> >>> >> >>
> >>> >> >>On Tue, Apr 10, 2012 at 11:04 AM, Jesse MacFadyen
> >>> >> >><purplecabbage@gmail.com>wrote:
> >>> >> >>
> >>> >> >>> None.
> >>> >> >>> Is none the new +1?
> >>> >> >>>
> >>> >> >>> Cheers,
> >>> >> >>>  Jesse
> >>> >> >>>
> >>> >> >>> Sent from my iPhone5
> >>> >> >>>
> >>> >> >>> On 2012-04-10, at 11:00 AM, Shazron <shazron@gmail.com>
wrote:
> >>> >> >>>
> >>> >> >>> > None
> >>> >> >>> >
> >>> >> >>> > 2012/4/10 Filip Maj <fil@adobe.com>:
> >>> >> >>> >> I was gonna tag it 1.6.0.. Objections?
> >>> >> >>> >>
> >>> >> >>> >> On 4/10/12 10:50 AM, "Joe Bowser" <bowserj@gmail.com>
wrote:
> >>> >> >>> >>
> >>> >> >>> >>> Are you going to tag it 1.6.0? or 1.6.0rc3?
> >>> >> >>> >>>
> >>> >> >>> >>> On Tue, Apr 10, 2012 at 10:50 AM, Filip
Maj <fil@adobe.com>
> >>> >>wrote:
> >>> >> >>> >>>
> >>> >> >>> >>>> .... Already notes in docs. Durrr.
> >>> >> >>> >>>>
> >>> >> >>> >>>> Tag?
> >>> >> >>> >>>>
> >>> >> >>> >>>> On 4/10/12 10:48 AM, "Filip Maj"
<fil@adobe.com> wrote:
> >>> >> >>> >>>>
> >>> >> >>> >>>>> AhhhŠ actually Compass is not
available in BlackBerry
> >>>before
> >>> >> >>>7.0.. So
> >>> >> >>> >>>> that
> >>> >> >>> >>>>> would explain why it's not working
on 6.0 :)
> >>> >> >>> >>>>>
> >>> >> >>> >>>>> I'm going to file an issue for
that in JIRA.
> >>> >> >>> >>>>>
> >>> >> >>> >>>>> I'm going to update the docs
to note this, and then, we
> >>> >>should be
> >>> >> >>> good
> >>> >> >>> >>>> to
> >>> >> >>> >>>>> tag, ya?
> >>> >> >>> >>>>>
> >>> >> >>> >>>>> On 4/10/12 10:44 AM, "Filip Maj"
<fil@adobe.com> wrote:
> >>> >> >>> >>>>>
> >>> >> >>> >>>>>> 37 failing tests on a Torch
running 6.0. The accel
> >>>callback
> >>> >>test
> >>> >> >>> >>>> failed
> >>> >> >>> >>>>>> but when I run the manual
tests for accel they all check
> >>> >>out, so
> >>> >> >>>the
> >>> >> >>> >>>> 37
> >>> >> >>> >>>>>> failing tests might be a
little blown up.
> >>> >> >>> >>>>>>
> >>> >> >>> >>>>>> The file API looks fine,
Drew.
> >>> >> >>> >>>>>>
> >>> >> >>> >>>>>> Looks to me like Compass
may be a little f'ed. The manual
> >>> >>tests
> >>> >> >>>for
> >>> >> >>> >>>>>> Compass keep returning "[object
object]" so there seems
> >>>to
> >>> >>be a
> >>> >> >>> little
> >>> >> >>> >>>>>> mistake in there somewhere.
Gord and I are looking into
> >>> that.
> >>> >> >>> >>>>>>
> >>> >> >>> >>>>>> If we resolve the compass
issue IMO we're good to tag. We
> >>> >>pass
> >>> >> >>>on
> >>> >> >>> >>>> both a
> >>> >> >>> >>>>>> 9900 (runs 7.0) and a Torch
(runs 6.0).
> >>> >> >>> >>>>>>
> >>> >> >>> >>>>>>
> >>> >> >>> >>>>>> On 4/10/12 10:31 AM, "Filip
Maj" <fil@adobe.com> wrote:
> >>> >> >>> >>>>>>
> >>> >> >>> >>>>>>> No worries Jesse.
> >>> >> >>> >>>>>>>
> >>> >> >>> >>>>>>> I got my hands on an
OS6 device so I'll try to
> >>>reproduce +
> >>> >>fix
> >>> >> >>>what
> >>> >> >>> >>>>>>> you're
> >>> >> >>> >>>>>>> seeing, Drew.
> >>> >> >>> >>>>>>>
> >>> >> >>> >>>>>>> On 4/10/12 10:04 AM,
"Jesse MacFadyen"
> >>> >> >>><purplecabbage@gmail.com>
> >>> >> >>> >>>> wrote:
> >>> >> >>> >>>>>>>
> >>> >> >>> >>>>>>>> I was the over anxious
js 1.6 tagger, in my rush to
> >>>have a
> >>> >> >>>long
> >>> >> >>> >>>>>>>> weekend.
> >>> >> >>> >>>>>>>> Sorry all.
> >>> >> >>> >>>>>>>>
> >>> >> >>> >>>>>>>> Cheers,
> >>> >> >>> >>>>>>>>  Jesse
> >>> >> >>> >>>>>>>>
> >>> >> >>> >>>>>>>> Sent from my iPhone5
> >>> >> >>> >>>>>>>>
> >>> >> >>> >>>>>>>> On 2012-04-10, at
9:50 AM, Joe Bowser
> >>><bowserj@gmail.com>
> >>> >> >>>wrote:
> >>> >> >>> >>>>>>>>
> >>> >> >>> >>>>>>>>> I deleted the
1.6.0 tag from Android.  I'll put it
> >>>back
> >>> >>when
> >>> >> >>>we
> >>> >> >>> >>>> get
> >>> >> >>> >>>>>>>>> this
> >>> >> >>> >>>>>>>>> sorted out!
> >>> >> >>> >>>>>>>>>
> >>> >> >>> >>>>>>>>> On Tue, Apr 10,
2012 at 9:38 AM, Drew Walters <
> >>> >> >>> deedubbu@gmail.com>
> >>> >> >>> >>>>>>>>> wrote:
> >>> >> >>> >>>>>>>>>
> >>> >> >>> >>>>>>>>>> I'm still
testing on other versions of BB.  Seeing
> >>>some
> >>> >>odd
> >>> >> >>> >>>> behavior
> >>> >> >>> >>>>>>>>>> all of a
sudden in File API on OS 6.  Not sure if it
> >>>is
> >>> >>my
> >>> >> >>>test
> >>> >> >>> >>>> app
> >>> >> >>> >>>>>>>>>> or
> >>> >> >>> >>>>>>>>>> real bug.
> >>> >> >>> >>>>>>>>>>
> >>> >> >>> >>>>>>>>>> On Tue, Apr
10, 2012 at 11:36 AM, Brian LeRoux
> >>> >><b@brian.io>
> >>> >> >>> >>>> wrote:
> >>> >> >>> >>>>>>>>>>> +1
> >>> >> >>> >>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>> On Tuesday,
April 10, 2012, Filip Maj wrote:
> >>> >> >>> >>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>> OK
so I pulled the latest master from cordova-js
> >>>and
> >>> >> >>> integrated
> >>> >> >>> >>>>>>>>>>>> with
> >>> >> >>> >>>>>>>>>>>> latest
master for blackberry-webworks.
> >>> >> >>> >>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>> Tested
on the 9900, looks good. 18 tests failing.
> >>> >> >>> >>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>> Tag
it - ship it. Let's iron out the rest in 1.7.
> >>> >> >>> >>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>> On
4/10/12 8:59 AM, "Filip Maj" <fil@adobe.com>
> >>> wrote:
> >>> >> >>> >>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>
Agree with leaving the RC tags alone. Just have to
> >>> >> >>> >>>> remove/retag
> >>> >> >>> >>>>>>>>>>>>>
1.6.0
> >>> >> >>> >>>>>>>>>> IMO
> >>> >> >>> >>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>
On 4/10/12 8:54 AM, "Shazron" <shazron@gmail.com>
> >>> >>wrote:
> >>> >> >>> >>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>
Let's wait until BB is done and do a tag reset
> >>> >> >>>discussion?
> >>> >> >>> >>>> with
> >>> >> >>> >>>>>>>>>>>>>>
steps
> >>> >> >>> >>>>>>>>>> to
> >>> >> >>> >>>>>>>>>>>>>>
take
> >>> >> >>> >>>>>>>>>>>>>>
1.6.0rc1 should still be there though I think
> >>> >> >>> >>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>
On Tue, Apr 10, 2012 at 8:34 AM, Filip Maj
> >>> >> >>><fil@adobe.com>
> >>> >> >>> >>>>>>>>>>>>>>
wrote:
> >>> >> >>> >>>>>>>>>>>>>>>
I'm in the process of testing the latest BB
> >>>code so
> >>> >> >>>I'll
> >>> >> >>> let
> >>> >> >>> >>>>>>>>>>>>>>>
you
> >>> >> >>> >>>>>>>>>> guys
> >>> >> >>> >>>>>>>>>>>>>>>
know
> >>> >> >>> >>>>>>>>>>>>>>>
soon how we're looking there.
> >>> >> >>> >>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>
Is that the last thing need before we're all
> >>>good
> >>> to
> >>> >> >>>tag
> >>> >> >>> >>>> this
> >>> >> >>> >>>>>>>>>> release?
> >>> >> >>> >>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>
On 4/10/12 8:30 AM, "Simon MacDonald"
> >>> >> >>> >>>>>>>>>>>>>>>
<simon.macdonald@gmail.com>
> >>> >> >>> >>>>>>>>>>>>>>>
wrote:
> >>> >> >>> >>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>
I think we should delete all the 1.6.0 tags. We
> >>> >> >>>haven't
> >>> >> >>> >>>>>>>>>>>>>>>>
released
> >>> >> >>> >>>>>>>>>>>>>>>>
any
> >>> >> >>> >>>>>>>>>>>>>>>>
build artifacts from 1.6.0 so there shouldn't
> >>>be a
> >>> >> >>>problem
> >>> >> >>> >>>>>>>>>>>>>>>>
with
> >>> >> >>> >>>>>>>>>> that.
> >>> >> >>> >>>>>>>>>>>>>>>>
So I agree with Fil's steps.
> >>> >> >>> >>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>
Simon Mac Donald
> >>> >> >>> >>>>>>>>>>>>>>>>
http://hi.im/simonmacdonald
> >>> >> >>> >>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>
On Tue, Apr 10, 2012 at 11:20 AM, Filip Maj <
> >>> >> >>> fil@adobe.com>
> >>> >> >>> >>>>>>>>>>>>>>>>
wrote:
> >>> >> >>> >>>>>>>>>>>>>>>>>
I like the general process Joe lays out.
> >>> >> >>> >>>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>>
I'm not sure how vendoring-in a tagged
> >>>cordova.js
> >>> >> >>>file is
> >>> >> >>> >>>>>>>>>>>>>>>>>
error
> >>> >> >>> >>>>>>>>>> prone
> >>> >> >>> >>>>>>>>>>>>>>>>>
though, Bryce. Is it just the manual process
> >>>of
> >>> >> >>>checking
> >>> >> >>> >>>> out
> >>> >> >>> >>>>>>>>>>>>>>>>>
a
> >>> >> >>> >>>>>>>>>> tag in
> >>> >> >>> >>>>>>>>>>>>>>>>>
cordova-js, building, and copying the file
> >>>over
> >>> to
> >>> >> >>>the
> >>> >> >>> >>>>>>>>>>>>>>>>>
platform
> >>> >> >>> >>>>>>>>>>>>>>>>>
implementation? If this is the concern then
> >>> >> >>>certainly,
> >>> >> >>> the
> >>> >> >>> >>>>>>>>>>>>>>>>>
release
> >>> >> >>> >>>>>>>>>>>>>>>>>
tool
> >>> >> >>> >>>>>>>>>>>>>>>>>
should be set up to do that automatically.
> >>> >> >>> >>>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>>
For some reason 1.6.0 tag in cordova-js was
> >>>added
> >>> >>4
> >>> >> >>>days
> >>> >> >>> >>>> ago,
> >>> >> >>> >>>>>>>>>>>>>>>>>
but
> >>> >> >>> >>>>>>>>>>>>>>>>>
1.6.0rc2
> >>> >> >>> >>>>>>>>>>>>>>>>>
was added ~ 1 day ago. Not sure what happened
> >>> >>there.
> >>> >> >>> >>>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>>
In light of the tags not being ordered
> >>>properly
> >>> >>and
> >>> >> >>>the
> >>> >> >>> >>>> file
> >>> >> >>> >>>>>>>>>>>>>>>>>
seek
> >>> >> >>> >>>>>>>>>> bug
> >>> >> >>> >>>>>>>>>>>>>>>>>
creeping in, I propose, just for the 1.6.0
> >>> >>release,
> >>> >> >>>that
> >>> >> >>> >>>> we:
> >>> >> >>> >>>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>>
1) Delete the old 1.6.0 tag in cordova-js.
> >>> >> >>> >>>>>>>>>>>>>>>>>
2) Retag cordova-js 1.6.0 to the latest commit
> >>> >>(that
> >>> >> >>> >>>> includes
> >>> >> >>> >>>>>>>>>>>>>>>>>
the
> >>> >> >>> >>>>>>>>>>>>>>>>>
file
> >>> >> >>> >>>>>>>>>>>>>>>>>
seek bug fix) - now our tags are at least in
> >>>the
> >>> >> >>>right
> >>> >> >>> >>>> order
> >>> >> >>> >>>>>>>>>>>>>>>>>
3) rebuild, reintegrate into platforms
> >>> >> >>> >>>>>>>>>>>>>>>>>
4) unfortunately, retag the platform
> >>> >>implementations
> >>> >> >>> 1.6.0
> >>> >> >>> >>>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>>
If retagging is too unholy then f it, I say we
> >>> tag
> >>> >> >>> >>>> everything
> >>> >> >>> >>>>>>>>>> 1.6.1.
> >>> >> >>> >>>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>>
On 4/10/12 7:19 AM, "Bryce Curtis"
> >>> >> >>> >>>> <curtis.bryce@gmail.com>
> >>> >> >>> >>>>>>>>>> wrote:
> >>> >> >>> >>>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>>>
As Joe eluded to, checking cordova-js into
> >>>the
> >>> >> >>>various
> >>> >> >>> >>>>>>>>>>>>>>>>>>
platform
> >>> >> >>> >>>>>>>>>>>>>>>>>>
repositories holds up the release.  It is
> >>>also
> >>> >>error
> >>> >> >>> >>>> prone -
> >>> >> >>> >>>>>>>>>>>>>>>>>>
not
> >>> >> >>> >>>>>>>>>> to
> >>> >> >>> >>>>>>>>>>>>>>>>>>
mention pushing to each repository every time
> >>> >>there
> >>> >> >>>is a
> >>> >> >>> >>>>>>>>>>>>>>>>>>
change
> >>> >> >>> >>>>>>>>>> takes
> >>> >> >>> >>>>>>>>>>>>>>>>>>
a lot of time & can get out of of sync.
> >>> >> >>> >>>>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>>>
Any thoughts on having the release build
> >>>script
> >>> >> >>>handle
> >>> >> >>> >>>> this?
> >>> >> >>> >>>>>>>>>>>>>>>>>>
As
> >>> >> >>> >>>>>>>>>> far
> >>> >> >>> >>>>>>>>>>>>>>>>>>
as during normal development and testing, we
> >>>are
> >>> >>all
> >>> >> >>> >>>>>>>>>>>>>>>>>>
building
> >>> >> >>> >>>>>>>>>>>>>>>>>>
cordova.js anyway, and keep current in our
> >>>own
> >>> >>ways.
> >>> >> >>> >>>>>>>>>>>>>>>>>>
> >>> >> >>> >>>>>>>>>>>>>>>>>>
On Tue, Apr 10, 2012 at 9:04 AM, Simon
> >>>MacDonald
> >>> >> >>> >>>>>>>>>>>>>>>>>>
<simon.macdonald@gmail.com> wrote:
> >>> >> >>> >>>>>>>>>>>>>>>>>>>
I just fixed what seems to be a zero day
> >>>bug in
> >>> >>our
> >>> >> >>> >>>>>>>>>> implementation
> >>> >> >>> >>>>>>>>>>>>>>>>>>>
of
> >>> >> >>> >>>>>>>>>>>>>>>>>>>
the
> >>> >> >>> >>>>>>>>>>>>>>>>>>>
FileWriter. If possible it would be good to
> >>>get
> >>> >> >>>this
> >>> >> >>> bug
> >>> >> >>> >>>>>>>>>>>>>>>>>>>
fix
> >>> >> >>> >>>>>>>>>> into
> >>> >> >>> >>>>>>>>>>>>>>>>>>>
all
> >>> >> >>> >>>>>>>>>>>>>>>>>>>
the
> >>> >> >>> >>>>>>>>>>>>>>>>>>>
platform
> >>> >> >>> >>>>>>>>>>
> >>> >> >>> >>>>>>>
> >>> >> >>> >>>>>>
> >>> >> >>> >>>>>
> >>> >> >>> >>>>
> >>> >> >>> >>>>
> >>> >> >>> >>
> >>> >> >>>
> >>> >> >
> >>> >>
> >>> >>
> >>>
> >>>
> >>
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message