couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF IRC Services" <asf...@wilderness.apache.org>
Subject Summary of IRC meeting in #couchdb-meeting, Wed Mar 27 20:09:05 2013
Date Wed, 27 Mar 2013 20:54:44 GMT
Members present: jan____, benoitc, ryan_ramage, wendall911, chewbranca, tilgovi, nslater_,
Wohali

----------------
Meeting summary:
----------------

1. Preface

2. 1.3.0 release
  a. wohali to create a ./configure check for >=R16B (jan____, 2)
  b. benoitc and wendall911 concerned about test suite stability. (jan____, 2)
  c. Noah to abourt 1.3.0.rc.2 and wait for fixes to go int 1.3.x (jan____, 2)
  d. EVERYBODY TEST RC.3 when it comes out (jan____, 2)
  e. or else... (nslater_, 2)

3. 1.2.2
  a. someone explain xylophone (Wohali, 3)
  b. ship 1.2.2 (jan____, 3)

4. spidermonkey 17
  a. mozilla started the release of https://developer.mozilla.org/en-US/docs/SpiderMonkey/17
(jan____, 4)
  b. benoitc & tilgovi to look into SM 17 (seventeen) compat and deprecation of support
for SM 1.7 (that’s one dot seven) (jan____, 4)
  c. deprecate SM 1.7 via ./configure (jan____, 4)
  d. Jan opens a beer (jan____, 4)

5. Google Summer of Code (GSoC)
  a. Yay GSoC. filed a bunch of issues, waiting for process now. dch in charge. (jan____,
5)
  b. some gsoc tagged issues need to be deduplicated against ongoing efforts (rcouch merge
etc) (jan____, 5)

6. Any Other Business
  a. wendall911 tor each out to dev@ for help with test suite issues (jan____, 6)


--------
Actions:
--------
- wohali to create a ./configure check for >=R16B (jan____, 20:18:50)
- Noah to abourt 1.3.0.rc.2 and wait for fixes to go int 1.3.x (jan____, 20:26:11)
- EVERYBODY TEST RC.3 when it comes out (jan____, 20:26:18)
- ship 1.2.2 (jan____, 20:32:56)

IRC log follows:


# 1. Preface #
20:09:06 [tilgovi]: oh, hey
20:09:27 [jan____]: let’s roll :)
20:09:27 [chewbranca]: jan____: not that I know of, been caught up on other things since last
week
20:09:41 [jan____]: chewbranca: aye


# 2. 1.3.0 release #
20:09:56 [nslater_]: i am still poorly. please consider me available for answering direct
questions only
20:10:26 [jan____]: nslater_: ack
20:10:42 [Wohali]: woo spidermonkey 17
20:11:11 [tilgovi]: "JavaScript 17 includes significant updates to language features, yo."
20:11:20 [tilgovi]: Thanks, Mozilla :-P. (Source: https://developer.mozilla.org/en-US/docs/SpiderMonkey/17)
20:11:27 [jan____]: 1.3.0: we are currently in rc2 and consider pulling it due to a test that
fails for many testers, but doesnt constitute a bug in CouchDB itself. This one could be annoying
for downstream packagers and from source/make check users though
20:11:27 [benoitc]: mmm
20:11:35 [jan____]: the fix is increasing a timeout, that is all.
20:11:42 [jan____]: the question is whether to cancel rc.2 and rol rc.3 with the fix
20:11:50 [jan____]: tilgovi: on topic pls
20:12:00 [benoitc]: so for some reason wouldn't write to this chan
20:12:15 [benoitc]: or read. what are the topics?
20:12:37 [tilgovi]: 1.3.0, 1.2.2, release testing, GSoC, spidermonkey 17....
20:12:42 [jan____]: I would like to hear from Wohali / rnewson / wendall911 (he already posted
on dev@) about this
20:12:57 [ryan_ramage]: jan____: well I think to send a good build message to the community
a rc3 would be good (and easier now I hope)
20:12:58 [jan____]: benoitc: [21:07:26] <jan____>	 1.3.0 release, 1.2.2 release, release
testing in general, sm17, gsoc
20:13:12 [tilgovi]: I think what Benoit said about noting the erlang incompatibility is a
good idea, too
20:13:13 [benoitc]: ok
20:13:35 [Wohali]: cancel and do rc3, and note we require erlang >= r14b01 < r16
20:13:35 [benoitc]: there is a patch oin master
20:13:50 [wendall911]: I'm here now. I am for aborting if this is something that packagers
are running into. Personally, this is not an issue with my package building. I'm having serious
issues with the javascript test runner.
20:13:51 [tilgovi]: but, if I can get a clarification on release procedure, does the first
RC to pass a vote *become* the final release, or is there another vote for 1.3.0 after the
release candidates?
20:13:59 [jan____]: tilgovi: we werent gonna not note it, just not in NEWS/CHANGES, but it
doesn’t hurt fixing that, when we roll anew
20:14:04 [benoitc]: for some reason i coulcn't apply it easily this tmorning on 1.3.x and
didn't have time time to investigate
20:14:06 [Wohali]: recently went through chef hell on something similar, they ended up specifying
a MAXIMUM version for a dependency, seems fair to me.
20:14:18 [jan____]: wendall911: notethat your errors are unrelated to this particular issue
20:14:20 [wendall911]: tilgovi: the rc becomes the release if +1
20:14:26 [jan____]: tilgovi: the rc *becomes* the final release
20:14:33 [Wohali]: in semver terms, a major release increase implies a possible abi breakage,
which effectively we're at right now.
20:14:35 [wendall911]: jan____: noted
20:14:36 [Wohali]: and yes bump to fix the timeout
20:14:48 [Wohali]: because CI matters.
20:15:13 [jan____]: Wohali: that means we want a ./configure level check for >=R16B as
well
20:15:26 [Wohali]: imo yes
20:15:27 [benoitc]: i'm confused about the thing that people can vote +1 or a release while
using patched tests on their own
20:15:35 [jan____]: Wohali: can you provide one?
20:15:36 [benoitc]: imo we should clarify the release procedure as well
20:15:48 [jan____]: benoitc: you can vote whatever on whatever
20:15:56 [jan____]: it is pretty clear
20:15:57 [benoitc]: for me make check should always work
20:16:03 [jan____]: you can vote even if tests fail
20:16:14 [jan____]: or iv you have to jump though a hoop to make it pass for your configuration
20:16:18 [benoitc]: yeah and i'm confused how you can vote  +1 on a broken release
20:16:20 [jan____]: I think that is snesible
20:16:34 [benoitc]: when it's supposed to be testing sign and make check
20:16:41 [jan____]: benoitc: in an idael world yes, but are you gonna sit down the three weeks
straigt to make the test suite work perfect everywhere?
20:16:49 [jan____]: a failing test does not constitute a failing release
20:16:56 [jan____]: *broken
20:16:56 [benoitc]: jan____: i make sure make check run
20:17:03 [Wohali]: jan____: Sure, I can do that.
20:17:03 [benoitc]: even if it's about disabling tests
20:17:11 [wendall911]: benoitc: I do the same, and that's been my issue also.
20:17:11 [jan____]: also, this is a derail at this point, can we focus on what we want to
do  for rc.3?
20:17:41 [jan____]: Wohali: any time soon, too? no rush, just want to get this over with quick
20:17:43 [Wohali]: can do it in the next 24h
20:17:48 [Wohali]: will take offline a concern i have with you
20:17:50 [Wohali]: not for this group
20:18:11 [nslater_]: for clarification: you can vote +1 based on any criteria you choose.
shipping with broken tests or even broken software is fine from a policy stand point. the
only thing you can't do is ship anything that is not compatible with the apache license
20:18:18 [jan____]: we might want to tie in the general lack of tests/votes in the Release
process. what can we do to improve that, so we don’t need an rc.4?
20:18:26 [jan____]: Wohali: ack
20:18:33 [jan____]: #actin wohali to create a ./configure check for >=R16B
20:18:50 [jan____]: #action wohali to create a ./configure check for >=R16B
20:19:11 [jan____]: #info benoitc and wendall911 concerned about test suite stability.
20:19:20 [nslater_]: i am happy to do rc.3 if people agree we need to abort
20:19:33 [nslater_]: i am just waiting for that message. 2 x -1 votes are pretty loud. a third
would convince me
20:19:48 [jan____]: nslater_: that seems to be the consenus for now
20:20:03 [nslater_]: okay then i have a request:
20:20:11 [nslater_]: please can we test the release thoroughly *now* so we dont end up doing
an rc.4
20:20:28 [nslater_]: anyone who's in this channel who hasn't tested/voted - please do so!
20:20:41 [jan____]: remsg: we might want to tie in the general lack of tests/votes in the
Release process. what can we do to improve that, so we don’t need an rc.4?
20:20:48 [nslater_]: aborting constantly because the testing is so fragmented wastes everybody's
time
20:20:48 [Wohali]: nslater_: consider this my official -1, i can email to dev@ if you need
20:21:12 [nslater_]: jan____: saw that - i dont think this is a process thing. no way you
can solve engagement issues with process
20:21:33 [nslater_]: jan____: we have what 14 committers at this stage? like 20% of them voted
20:21:48 [jan____]: nslater_: I am not asking for process
20:21:57 [nslater_]: okay
20:22:03 [chewbranca]: I went and looked at rc.2 thread last night, and given the number of
-1s, I assumed it would be aborted
20:22:11 [nslater_]: bare minimum activity from a pmc member should be testing releases imo
20:22:26 [chewbranca]: I might have time to test tomorrow night, at the very least I'll have
more time friday night and over the weekend
20:22:33 [nslater_]: if they can't do that, i dont see what business they have being on the
pmc ;)
20:22:56 [nslater_]: (modulo time/family/personal constraints. obviously. i am being hyperbolic.
but you get my point)
20:22:58 [jan____]: nslater_: pmc issues are out of scope for this meeting
20:23:03 [Wohali]: i will remind folks it's easter weekend
20:23:11 [nslater_]: Wohali: it wasn't 1 week ago ;)
20:23:13 [Wohali]: public holidays for a lot of people (in commonwealth nations)
20:23:18 [Wohali]: fair.
20:23:18 [jan____]: nslater_: (also see pm)
20:23:33 [nslater_]: not sure why pmc activities are out of scope
20:23:41 [tilgovi]: I'll run it, but probably -0 since I'd prefer we raise that timeout and
add a configure check for R16
20:23:48 [tilgovi]: I did run the last rc pretty thoroughly
20:23:56 [jan____]: nslater_: meta, xylophone whatever, lets move on.
20:24:03 [nslater_]: what is this xylophone reference? second time i see it now
20:24:11 [benoitc]: tilgovi: you also have fraction in votes 
20:24:14 [nslater_]: sure, move on. i am just highlighting an engagement issue we have
20:24:19 [nslater_]: not sure how to solve it without cajoling people
20:24:41 [jan____]: tilgovi: did  you tell dev@ (currently not seeing you there, might be
me)
20:24:48 [wendall911]: On a positive note, looks like 1.2.2 is gtg :)
20:24:48 [tilgovi]: nope, I will do so after I do the testing
20:24:56 [tilgovi]: sorry for confusion
20:25:13 [nslater_]: wendall911: yes i am good to call that as soon as i dont feel super-poorly
20:25:26 [nslater_]: thanks everyone for pitching in on that one!
20:25:26 [jan____]: tilgovi: thanks
20:26:11 [jan____]: #action Noah to abourt 1.3.0.rc.2 and wait for fixes to go int 1.3.x
20:26:18 [jan____]: #action EVERYBODY TEST RC.3 when it comes out
20:26:48 [nslater_]: #info or else...
20:27:12 [jan____]: ok


# 3. 1.2.2 #
20:27:26 [jan____]: SHIP IT
20:27:41 [nslater_]: yep
20:27:56 [Wohali]: #info someone explain xylophone
20:28:20 [nslater_]: apparently it means "time is up. let's move on."
20:29:41 [Wohali]: ship it
20:32:48 [tilgovi]: did I get disconnected or did this just die somehow?
20:32:56 [tilgovi]: what's next?
20:32:56 [jan____]: #action ship 1.2.2


# 4. spidermonkey 17 #
20:33:03 [tilgovi]: good
20:33:11 [tilgovi]: so, I just glanced at the SM 17 docs
20:33:18 [jan____]: #info mozilla started the release of https://developer.mozilla.org/en-US/docs/SpiderMonkey/17
20:33:26 [jan____]: we need to asses what that means for us
20:33:26 [jan____]: s
20:33:43 [tilgovi]: I will say for sure that if I do any work to make it happen I am cutting
SM 1.7 support
20:33:56 [tilgovi]: that's 1 DOT 7 not 17
20:34:11 [benoitc]: tilgovi: apaprently there are 1/2 lines to change in our code
20:34:14 [jan____]: tilgovi: just out of principle?
20:34:18 [tilgovi]: yeah
20:34:26 [benoitc]: or at least in the only sm185 version of rcouch
20:34:33 [tilgovi]: okay
20:34:33 [jan____]: tilgovi: I thoguht the sm/abstraction was fairly self contained?
20:34:36 [tilgovi]: if it's only 1/2 lines changed from 1.8.5
20:34:41 [benoitc]: +1 for removing 1.7
20:34:48 [jan____]: cool
20:34:48 [tilgovi]: then we can probably support it all still
20:34:56 [tilgovi]: but it does feel a bit ridiculous
20:34:56 [benoitc]: i will make a full test tonight
20:35:03 [jan____]: we can deprecate it at least
20:35:11 [tilgovi]: ahh! that's the way to do it!
20:35:11 [benoitc]: i've prepared a build but was qide tracked by a customer
20:35:18 [wendall911]: tilgovi: we'll need to have an "official" 1.8.5 package for RHEL, etc.
and some older Debian versions, as 1.7 is default there.
20:35:18 [tilgovi]: deprecate with a configure warning, remove in the next version
20:35:43 [jan____]: #info benoitc & tilgovi to look into SM 17 (seventeen) compat and
deprecation of support for SM 1.7 (that’s one dot seven)
20:36:03 [wendall911]: tilgovi: Wohali has one she supports for cloudant, and I have a working
spec on my github for RHEL
20:36:11 [jan____]: #info deprecate SM 1.7 via ./configure
20:36:18 [tilgovi]: I had a working spec for 1.8.5 somewhere once, too
20:36:26 [jan____]: #info Jan opens a beer
20:36:26 [tilgovi]: we'll talk it on the list
20:36:33 [tilgovi]: or wherever
20:36:33 [tilgovi]: out of meeting
20:36:41 [jan____]: aye
20:36:48 [jan____]: eot?
20:36:50 [tilgovi]: i think so


# 5. Google Summer of Code (GSoC) #
20:38:11 [jan____]: I am behind the email thread, but I saw dch and others came with a great
list of things, that got added to JIRA
20:38:18 [jan____]: and I am told we met any ASF/GSoC orga daedlines
20:38:26 [jan____]: *deadlines
20:38:33 [jan____]: to which I say: Fuck Yeah Go Team! <3
20:39:04 [benoitc]: i would just remove the already wip from the jora tickets tagged gsoc
20:39:18 [jan____]: benoitc: do it :)
20:40:13 [benoitc]: ok
20:40:20 [jan____]: #info Yay GSoC. filed a bunch of issues, waiting for process now. dch
in charge.
20:40:20 [wendall911]: It makes sense to have tickets for everything, but not having stuff
already going in with other code (rcouch, bigcouch) doesn't make sense for gsoc
20:40:50 [jan____]: #info some gsoc tagged issues need to be deduplicated against ongoing
efforts (rcouch merge etc)
20:40:59 [wendall911]: There was a good discussion around partial updates. tilgovi suggested
I make a thread about OT. Any interest there for gsoc?
20:41:26 [tilgovi]: I support it!
20:41:27 [wendall911]: OT meaning operational tranformation as a potential for partial updates
20:41:33 [Wohali]: ah
20:41:41 [benoitc]: wendall911: doubt anything can be achieved on that parts until the  some
internals are rewritten though
20:41:41 [jan____]: wendall911: it is not always clear what comes and doesn’t with BC and
rcouch, especially for the folks who are not into these projects. We can always kill JIRAs,
so I am happy we got it all in for now
20:41:48 [Wohali]: ah, not off topic
20:41:48 [jan____]: "about OT"?
20:41:48 [Wohali]: was confused
20:41:50 [jan____]: ah
20:41:58 [wendall911]: benoitc: agreed, and we had this discussion some months back as well
20:42:03 [benoitc]: imo we should only target things that can be done and help the prokect
in short terms
20:42:11 [benoitc]: project
20:42:26 [jan____]: I have a branch in my head that implements JSONPointer & JSON Patch
as regular ops for CouchDB
20:42:26 [wendall911]: ok, that was my thought as well
20:42:34 [jan____]: happy to have anyone beating me to it, but that’s the sane way forward
20:43:19 [jan____]: ok
20:43:26 [jan____]: EOT?
20:43:41 [benoitc]: jan____: also https://github.com/lechat-im/jsonq
20:43:41 [Wohali]: jan____: you mean HTTP PATCH?
20:43:56 [jan____]: whatever mnot made
20:44:18 [wendall911]: So, gsoc stuff should be obtainable implementations, not end up with
a bunch of partial code?
20:45:03 [jan____]: yea
20:45:19 [wendall911]: +1
20:45:33 [jan____]: EOT?
20:46:19 [wendall911]: +1


# 6. Any Other Business #
20:46:35 [jan____]: *sits back*
20:47:05 [tilgovi]: none from me
20:47:12 [wendall911]: I need some help troubleshooting some issues with the test suite. I'm
digging in, but am not seeing *why* some things are happening. I think a couple may be erlang
issues.
20:47:26 [wendall911]: What I'd like to know is where best to ask. dev@
20:47:26 [benoitc]: """
20:47:27 [benoitc]: Keep in mind that your ideas list should be a starting point for student
proposals; we've heard from past mentoring organization participants that some of their best
student projects are those that greatly expanded on a proposed idea or were blue-sky proposals
not mentioned on the ideas list at all. A link to a bug tracker for your open source organization
is NOT an ideas list.
20:47:34 [benoitc]: """
20:47:42 [jan____]: wendall911: shoot any requests for help on dev@, I’m happy to pitch
in where I can
20:47:56 [wendall911]: jan____: ok, will do
20:48:11 [benoitc]: http://www.google-melange.com/gsoc/document/show/gsoc_program/google/gsoc2013/help_page#3._What_are_the_goals_of_this_program
(source)
20:48:13 [jan____]: benoitc: we are collecting a prelimenary list, that is *not* the one we
send to GSoC.
20:48:33 [benoitc]: jan____: was answerin to idea vs partial code
20:48:33 [jan____]: benoitc: we are still a step before that
20:48:42 [benoitc]: not speaking about the last part
20:48:48 [jan____]: cool, yeah, perfect
20:48:49 [wendall911]: benoitc: appreciated, answers it well
20:49:21 [jan____]: #info wendall911 tor each out to dev@ for help with test suite issues
20:49:41 [jan____]: cool, letting this all sit for three minutes, then closing the meeting.
20:50:50 [benoitc]: wendall911: about testing & co i think we need to open some discussions
on @dev as well
20:51:18 [wendall911]: benoitc: agreed
20:51:28 [benoitc]: we really lack of some simple logging and tools to troubleshoot couchdb
as well
20:51:41 [jan____]: yes
20:51:42 [jan____]: +100000


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