You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@couchdb.apache.org by Paul Davis <pa...@gmail.com> on 2010/11/25 19:43:42 UTC
[VOTE] Apache CouchDB 1.0.2 release, Round 1
Hello,
I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
Changes since 1.0.1:
Futon:
* Make test suite work with Safari and Chrome.
Storage System:
* Fix leaking file handles after compacting databases and views.
* Fix databases forgetting their validation function after compaction.
* Fix occasional timeout errors after successfully compacting large databases.
* Fix ocassional error when writing to a database that has just been compacted.
* Fix occasional timeout errors on systems with slow or heavily loaded IO.
Log System:
* Reduce lengthy stack traces.
* Allow logging of native <xml> types.
HTTP Interface:
* Allow reduce=false parameter in map-only views.
* Fix parsing of Accept headers.
Replicator:
* Updated ibrowse library to 2.1.0 fixing numerous replication issues.
* Fix authenticated replication (with HTTP basic auth) of design documents
with attachments.
* Various fixes to make replication more resilient for edge-cases.
View Server:
* Don't trigger view updates when requesting `_design/doc/_info`.
* Fix for circular references in CommonJS requires.
* Made isArray() function available to functions executed in the query server.
* Documents are now sealed before being passed to map functions.
We encourage the whole community to download and test these release artifacts so
that any critical issues can be resolved before the release is made. Everyone is
free to vote on this release, so get stuck in!
We are voting on the following release artifacts:
http://people.apache.org/~davisp/dist/1.0.2/
These artifacts have been built from the 1.0.2 tag in Subversion:
http://svn.apache.org/repos/asf/couchdb/tags/1.0.2/
Happy voting,
Paul Davis
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Noah Slater <ns...@apache.org>.
On 28 Nov 2010, at 19:53, Paul Davis wrote:
> Is there a special incantation for aborting a vote? Something like:
I usually say something like:
> I'm aborting this vote for [some obvious reason.]
>
> I will prepare a new release artefact and call round two of the vote shortly.
This is canonicalised here:
http://wiki.apache.org/couchdb/Release_procedure
Where it says:
> The Release Manager has the power to abort a vote at any point and for any reason.
Notice how purposefully vague I left it?
THATS BECAUSE I HAVE TEH AWSUM POWARS.
I guess you could say:
> I'm feeling a little capricious today, so the vote is aborted.
>
> I'll prepare a new release and call round two of the vote whenever I damn well feel like it.
I would advise against that though.
Lest Rusty comes after you like a rabid hound, baying for your blood.
> It's not pinin'! It's passed on! This vote is no more! It has ceased
> to be! It's expired and gone to meet 'is maker! It's a stiff! Bereft
> of life, it rests in peace! If you hadn't nailed it to the perch it'd
> be pushing up the daisies! Its metabolic processes are now 'istory!
> It's off the twig! It's kicked the bucket, it's shuffled off 'is
> mortal coil, run down the curtain and joined the bleedin' choir
> invisibile!! THIS IS AN EX-VOTE!!
Or that.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Robert Newson <ro...@gmail.com>.
Yup, that's the usual incantation.
On Sun, Nov 28, 2010 at 7:53 PM, Paul Davis <pa...@gmail.com> wrote:
> On Sun, Nov 28, 2010 at 1:45 PM, Noah Slater <ns...@apache.org> wrote:
>> I'd like to note that Jan discovered an issue with the release. I was waiting for Paul to abort the vote, but I guess he's been busy. It's a busy season! Anyway, the NEWS and CHANGES files need to be updated to indicate the current release and to remove "This version has not been released yet." Because of this, I'm going to give the release my -1 for now.
>>
>
> Yeah, between family and food I haven't had a chance to respond. The
> NEWS and CHANGES and the recent bug report are enough to get me to -1
> the release on my own.
>
> Is there a special incantation for aborting a vote? Something like:
>
> It's not pinin'! It's passed on! This vote is no more! It has ceased
> to be! It's expired and gone to meet 'is maker! It's a stiff! Bereft
> of life, it rests in peace! If you hadn't nailed it to the perch it'd
> be pushing up the daisies! Its metabolic processes are now 'istory!
> It's off the twig! It's kicked the bucket, it's shuffled off 'is
> mortal coil, run down the curtain and joined the bleedin' choir
> invisibile!! THIS IS AN EX-VOTE!!
>
> Paul
>
>> On 28 Nov 2010, at 10:04, Dirkjan Ochtman wrote:
>>
>>> On Fri, Nov 26, 2010 at 11:34, Dirkjan Ochtman <di...@ochtman.nl> wrote:
>>>> Browser test suite fails in a few places, but this might be because
>>>> I'm running it through an SSH tunnel. Report at
>>>> http://couchdb.couchdb.org/_utils/document.html?test_suite_reports/ecc3f61712723f766c818ca17774f382.
>>>
>>> Without tunnel, it looks better, but reader_acl and replication still
>>> fail (with Firefox 4b7).
>>>
>>> I tried to run the 1.0.1 test suite too, same box, same browser, and
>>> it also fails reader_acl, but not replication.
>>>
>>> Cheers,
>>>
>>> Dirkjan
>>
>>
>
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Paul Davis <pa...@gmail.com>.
On Sun, Nov 28, 2010 at 1:45 PM, Noah Slater <ns...@apache.org> wrote:
> I'd like to note that Jan discovered an issue with the release. I was waiting for Paul to abort the vote, but I guess he's been busy. It's a busy season! Anyway, the NEWS and CHANGES files need to be updated to indicate the current release and to remove "This version has not been released yet." Because of this, I'm going to give the release my -1 for now.
>
Yeah, between family and food I haven't had a chance to respond. The
NEWS and CHANGES and the recent bug report are enough to get me to -1
the release on my own.
Is there a special incantation for aborting a vote? Something like:
It's not pinin'! It's passed on! This vote is no more! It has ceased
to be! It's expired and gone to meet 'is maker! It's a stiff! Bereft
of life, it rests in peace! If you hadn't nailed it to the perch it'd
be pushing up the daisies! Its metabolic processes are now 'istory!
It's off the twig! It's kicked the bucket, it's shuffled off 'is
mortal coil, run down the curtain and joined the bleedin' choir
invisibile!! THIS IS AN EX-VOTE!!
Paul
> On 28 Nov 2010, at 10:04, Dirkjan Ochtman wrote:
>
>> On Fri, Nov 26, 2010 at 11:34, Dirkjan Ochtman <di...@ochtman.nl> wrote:
>>> Browser test suite fails in a few places, but this might be because
>>> I'm running it through an SSH tunnel. Report at
>>> http://couchdb.couchdb.org/_utils/document.html?test_suite_reports/ecc3f61712723f766c818ca17774f382.
>>
>> Without tunnel, it looks better, but reader_acl and replication still
>> fail (with Firefox 4b7).
>>
>> I tried to run the 1.0.1 test suite too, same box, same browser, and
>> it also fails reader_acl, but not replication.
>>
>> Cheers,
>>
>> Dirkjan
>
>
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Noah Slater <ns...@apache.org>.
I'd like to note that Jan discovered an issue with the release. I was waiting for Paul to abort the vote, but I guess he's been busy. It's a busy season! Anyway, the NEWS and CHANGES files need to be updated to indicate the current release and to remove "This version has not been released yet." Because of this, I'm going to give the release my -1 for now.
On 28 Nov 2010, at 10:04, Dirkjan Ochtman wrote:
> On Fri, Nov 26, 2010 at 11:34, Dirkjan Ochtman <di...@ochtman.nl> wrote:
>> Browser test suite fails in a few places, but this might be because
>> I'm running it through an SSH tunnel. Report at
>> http://couchdb.couchdb.org/_utils/document.html?test_suite_reports/ecc3f61712723f766c818ca17774f382.
>
> Without tunnel, it looks better, but reader_acl and replication still
> fail (with Firefox 4b7).
>
> I tried to run the 1.0.1 test suite too, same box, same browser, and
> it also fails reader_acl, but not replication.
>
> Cheers,
>
> Dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Dirkjan Ochtman <di...@ochtman.nl>.
On Fri, Nov 26, 2010 at 11:34, Dirkjan Ochtman <di...@ochtman.nl> wrote:
> Browser test suite fails in a few places, but this might be because
> I'm running it through an SSH tunnel. Report at
> http://couchdb.couchdb.org/_utils/document.html?test_suite_reports/ecc3f61712723f766c818ca17774f382.
Without tunnel, it looks better, but reader_acl and replication still
fail (with Firefox 4b7).
I tried to run the 1.0.1 test suite too, same box, same browser, and
it also fails reader_acl, but not replication.
Cheers,
Dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Dirkjan Ochtman <di...@ochtman.nl>.
On Thu, Nov 25, 2010 at 19:43, Paul Davis <pa...@gmail.com> wrote:
> We encourage the whole community to download and test these release artifacts so
> that any critical issues can be resolved before the release is made. Everyone is
> free to vote on this release, so get stuck in!
>
> We are voting on the following release artifacts:
>
> http://people.apache.org/~davisp/dist/1.0.2/
Looking fairly good on Gentoo. etap tests all pass.
Browser test suite fails in a few places, but this might be because
I'm running it through an SSH tunnel. Report at
http://couchdb.couchdb.org/_utils/document.html?test_suite_reports/ecc3f61712723f766c818ca17774f382.
Cheers,
Dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Noah Slater <ns...@apache.org>.
Give me your home address and I'll draw you one and post it.
On 25 Nov 2010, at 23:54, Paul Davis wrote:
> On Thu, Nov 25, 2010 at 5:40 PM, Noah Slater <ns...@apache.org> wrote:
>>
>> On 25 Nov 2010, at 18:43, Paul Davis wrote:
>>
>>> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>>
>> Good job, Paul.
>>
>> This release gets the official Noah Slater Seal of Approval.
>>
>> +1
>>
>>
>
> When can I expect my official copy in the mail?
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Paul Davis <pa...@gmail.com>.
On Thu, Nov 25, 2010 at 5:40 PM, Noah Slater <ns...@apache.org> wrote:
>
> On 25 Nov 2010, at 18:43, Paul Davis wrote:
>
>> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>
> Good job, Paul.
>
> This release gets the official Noah Slater Seal of Approval.
>
> +1
>
>
When can I expect my official copy in the mail?
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Noah Slater <ns...@apache.org>.
On 25 Nov 2010, at 18:43, Paul Davis wrote:
> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
Good job, Paul.
This release gets the official Noah Slater Seal of Approval.
+1
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Benoit Chesneau <bc...@gmail.com>.
On Thu, Nov 25, 2010 at 7:43 PM, Paul Davis <pa...@gmail.com> wrote:
> Hello,
>
> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>
+1
tested on ubuntu 10.10 amd64 & macosx 10.6.5 with erlang r13b03,
r13b04 & r14b .
make check ok
javascript tests ok
sh1 & md5 signatures ok
- benoit
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Robert Newson <ro...@gmail.com>.
+1
* checksums and sigs match.
* make distcheck passes.
* Futon test suite passes in Firefox, Chrome and Safari.
OS X (10.6.5), erlang 14B, spidermonkey 1.8.5, ICU 4.4.1.
B.
On Thu, Nov 25, 2010 at 6:43 PM, Paul Davis <pa...@gmail.com> wrote:
> Hello,
>
> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>
> Changes since 1.0.1:
>
> Futon:
>
> * Make test suite work with Safari and Chrome.
>
> Storage System:
>
> * Fix leaking file handles after compacting databases and views.
> * Fix databases forgetting their validation function after compaction.
> * Fix occasional timeout errors after successfully compacting large databases.
> * Fix ocassional error when writing to a database that has just been compacted.
> * Fix occasional timeout errors on systems with slow or heavily loaded IO.
>
> Log System:
>
> * Reduce lengthy stack traces.
> * Allow logging of native <xml> types.
>
> HTTP Interface:
>
> * Allow reduce=false parameter in map-only views.
> * Fix parsing of Accept headers.
>
> Replicator:
>
> * Updated ibrowse library to 2.1.0 fixing numerous replication issues.
> * Fix authenticated replication (with HTTP basic auth) of design documents
> with attachments.
> * Various fixes to make replication more resilient for edge-cases.
>
> View Server:
>
> * Don't trigger view updates when requesting `_design/doc/_info`.
> * Fix for circular references in CommonJS requires.
> * Made isArray() function available to functions executed in the query server.
> * Documents are now sealed before being passed to map functions.
>
> We encourage the whole community to download and test these release artifacts so
> that any critical issues can be resolved before the release is made. Everyone is
> free to vote on this release, so get stuck in!
>
> We are voting on the following release artifacts:
>
> http://people.apache.org/~davisp/dist/1.0.2/
>
> These artifacts have been built from the 1.0.2 tag in Subversion:
>
> http://svn.apache.org/repos/asf/couchdb/tags/1.0.2/
>
> Happy voting,
> Paul Davis
>
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Robert Dionne <di...@dionne-associates.com>.
+1
OS X 10.6.5 Erlang R14B
everything works except the futon test auth_cache in Chrome
On Nov 25, 2010, at 1:43 PM, Paul Davis wrote:
> Hello,
>
> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>
> Changes since 1.0.1:
>
> Futon:
>
> * Make test suite work with Safari and Chrome.
>
> Storage System:
>
> * Fix leaking file handles after compacting databases and views.
> * Fix databases forgetting their validation function after compaction.
> * Fix occasional timeout errors after successfully compacting large databases.
> * Fix ocassional error when writing to a database that has just been compacted.
> * Fix occasional timeout errors on systems with slow or heavily loaded IO.
>
> Log System:
>
> * Reduce lengthy stack traces.
> * Allow logging of native <xml> types.
>
> HTTP Interface:
>
> * Allow reduce=false parameter in map-only views.
> * Fix parsing of Accept headers.
>
> Replicator:
>
> * Updated ibrowse library to 2.1.0 fixing numerous replication issues.
> * Fix authenticated replication (with HTTP basic auth) of design documents
> with attachments.
> * Various fixes to make replication more resilient for edge-cases.
>
> View Server:
>
> * Don't trigger view updates when requesting `_design/doc/_info`.
> * Fix for circular references in CommonJS requires.
> * Made isArray() function available to functions executed in the query server.
> * Documents are now sealed before being passed to map functions.
>
> We encourage the whole community to download and test these release artifacts so
> that any critical issues can be resolved before the release is made. Everyone is
> free to vote on this release, so get stuck in!
>
> We are voting on the following release artifacts:
>
> http://people.apache.org/~davisp/dist/1.0.2/
>
> These artifacts have been built from the 1.0.2 tag in Subversion:
>
> http://svn.apache.org/repos/asf/couchdb/tags/1.0.2/
>
> Happy voting,
> Paul Davis
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Sebastian Cohnen <se...@googlemail.com>.
-1 b/c of COUCHDB-968
On 25.11.2010, at 19:43, Paul Davis wrote:
> Hello,
>
> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>
> Changes since 1.0.1:
>
> Futon:
>
> * Make test suite work with Safari and Chrome.
>
> Storage System:
>
> * Fix leaking file handles after compacting databases and views.
> * Fix databases forgetting their validation function after compaction.
> * Fix occasional timeout errors after successfully compacting large databases.
> * Fix ocassional error when writing to a database that has just been compacted.
> * Fix occasional timeout errors on systems with slow or heavily loaded IO.
>
> Log System:
>
> * Reduce lengthy stack traces.
> * Allow logging of native <xml> types.
>
> HTTP Interface:
>
> * Allow reduce=false parameter in map-only views.
> * Fix parsing of Accept headers.
>
> Replicator:
>
> * Updated ibrowse library to 2.1.0 fixing numerous replication issues.
> * Fix authenticated replication (with HTTP basic auth) of design documents
> with attachments.
> * Various fixes to make replication more resilient for edge-cases.
>
> View Server:
>
> * Don't trigger view updates when requesting `_design/doc/_info`.
> * Fix for circular references in CommonJS requires.
> * Made isArray() function available to functions executed in the query server.
> * Documents are now sealed before being passed to map functions.
>
> We encourage the whole community to download and test these release artifacts so
> that any critical issues can be resolved before the release is made. Everyone is
> free to vote on this release, so get stuck in!
>
> We are voting on the following release artifacts:
>
> http://people.apache.org/~davisp/dist/1.0.2/
>
> These artifacts have been built from the 1.0.2 tag in Subversion:
>
> http://svn.apache.org/repos/asf/couchdb/tags/1.0.2/
>
> Happy voting,
> Paul Davis
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Volker Mische <vo...@gmail.com>.
+1
* Md5 sum matches
* make dist works
* Browser tests tested on FF and Chromium
* my system:
* Debian
* AMD64
* Erlang R13B04
* SpiderMonkey 1.9.1.9
* FF (IceWeasel) 3.6.4
* Chromimum 6.0.420.0 (48550)
On 11/25/2010 07:43 PM, Paul Davis wrote:
> Hello,
>
> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>
> Changes since 1.0.1:
>
> Futon:
>
> * Make test suite work with Safari and Chrome.
>
> Storage System:
>
> * Fix leaking file handles after compacting databases and views.
> * Fix databases forgetting their validation function after compaction.
> * Fix occasional timeout errors after successfully compacting large databases.
> * Fix ocassional error when writing to a database that has just been compacted.
> * Fix occasional timeout errors on systems with slow or heavily loaded IO.
>
> Log System:
>
> * Reduce lengthy stack traces.
> * Allow logging of native <xml> types.
>
> HTTP Interface:
>
> * Allow reduce=false parameter in map-only views.
> * Fix parsing of Accept headers.
>
> Replicator:
>
> * Updated ibrowse library to 2.1.0 fixing numerous replication issues.
> * Fix authenticated replication (with HTTP basic auth) of design documents
> with attachments.
> * Various fixes to make replication more resilient for edge-cases.
>
> View Server:
>
> * Don't trigger view updates when requesting `_design/doc/_info`.
> * Fix for circular references in CommonJS requires.
> * Made isArray() function available to functions executed in the query server.
> * Documents are now sealed before being passed to map functions.
>
> We encourage the whole community to download and test these release artifacts so
> that any critical issues can be resolved before the release is made. Everyone is
> free to vote on this release, so get stuck in!
>
> We are voting on the following release artifacts:
>
> http://people.apache.org/~davisp/dist/1.0.2/
>
> These artifacts have been built from the 1.0.2 tag in Subversion:
>
> http://svn.apache.org/repos/asf/couchdb/tags/1.0.2/
>
> Happy voting,
> Paul Davis
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Noah Slater <ns...@apache.org>.
On 25 Nov 2010, at 23:27, Robert Newson wrote:
> pretty sure ibrowse and mochiweb have important tweaks, using versions
> without them is broken.
>
> Not sure I understand the release process if it allows voting on
> things other than the release artifact.
Peter was voting on the release artefact. It just so happens that he's also modified it to suit his purposes. Say I downloaded the release artefact, and installed it into a custom set-up as well. If it pleased me, I'd vote a +1 because it works how I want it to. Seems fine to me.
If someone took the release artefact, modified it, and uploaded it somewhere else, and said "I've voted +1 on the release when it contains my changes." That would be a problem. But voting on the release artefact for its suitability for your own purposes is totally fine.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by till <ti...@php.net>.
On Fri, Nov 26, 2010 at 7:14 AM, Peter Lemenkov <le...@gmail.com> wrote:
> 2010/11/26 Robert Newson <ro...@gmail.com>:
>> pretty sure ibrowse and mochiweb have important tweaks, using versions
>> without them is broken.
>
> I'm not sure about mochiweb (since couchdb contains very old svn
> snapshot, and it's hard to check for the differences) but regarding
> ibrowse - couchdb is only 1 patch behind the 2.1.0. I'll plan to
> backport in into 2.1.0 shipped in Fedora.
>
> So
>
> a) These libraries are not so different from upstream ones (at least
> true for ibrowse)
> b) CouchDB in Fedora isn't so deadly broken as someone of you may think
>
> In fact I'm watching very closely the development of both ibrowse and
> mochiweb so I'm aware of the changes made or just proposed by CouchDB
> folks.
>
Please correct me if am I wrong (actually... :D) but what Peter is
doing is pretty common for packaging.
And in the end, there is a test suite to make sure everything works.
;-) If it doesn't catch an issue, it might need an improvement, or
two.
Till
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Peter Lemenkov <le...@gmail.com>.
2010/11/26 Peter Lemenkov <le...@gmail.com>:
> 2010/11/26 Peter Lemenkov <le...@gmail.com>:
>> just play for the stable API/ABI interfaces. Shipping internally a
>
> s,play,play,
s,play,pRay,
:)
--
With best regards, Peter Lemenkov.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Peter Lemenkov <le...@gmail.com>.
2010/11/26 Peter Lemenkov <le...@gmail.com>:
> just play for the stable API/ABI interfaces. Shipping internally a
s,play,play,
--
With best regards, Peter Lemenkov.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Robert Newson <ro...@gmail.com>.
I'm working on a branch to use mochiweb 1.4.1 in trunk for a the 1.2 release.
b.
On Fri, Nov 26, 2010 at 3:06 PM, Dirkjan Ochtman <di...@ochtman.nl> wrote:
> On Fri, Nov 26, 2010 at 15:57, Jan Lehnardt <ja...@apache.org> wrote:
>> I wonder what the best solution here is. The problem with updating
>> to the latest Mochiweb right before a release is that subtle issues
>> don't have time surfacing in trunk. To ensure stable releases we
>> should be conservative with updates (modulo critical bug fixes of
>> course) or upstream libraries. How can we best ensure to keep them
>> up to date enough while not compromising a release's stability?
>
> The best solution IMO is if you can rely on upstream releases but
> specify which versions have been tested with. I.e. I'm happy to keep
> an older version of mochiweb around if you're not going to want to
> upgrade to the shiny new version right before a CouchDB release.
>
> Cheers,
>
> Dirkjan
>
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Dirkjan Ochtman <di...@ochtman.nl>.
On Fri, Nov 26, 2010 at 15:57, Jan Lehnardt <ja...@apache.org> wrote:
> I wonder what the best solution here is. The problem with updating
> to the latest Mochiweb right before a release is that subtle issues
> don't have time surfacing in trunk. To ensure stable releases we
> should be conservative with updates (modulo critical bug fixes of
> course) or upstream libraries. How can we best ensure to keep them
> up to date enough while not compromising a release's stability?
The best solution IMO is if you can rely on upstream releases but
specify which versions have been tested with. I.e. I'm happy to keep
an older version of mochiweb around if you're not going to want to
upgrade to the shiny new version right before a CouchDB release.
Cheers,
Dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Jan Lehnardt <ja...@apache.org>.
On 26 Nov 2010, at 15:23, Peter Lemenkov wrote:
> Hello
>
> 2010/11/26 Jan Lehnardt <ja...@apache.org>:
>
>> We always patch Mochiweb to fit our needs that are divergent from the
>> upstream project. It isn't safe to assume an independent cut of Mochiweb
>> will just work.
>
> Well, it isn't even safe to assume that it will work on an arbitrary
> system with different set of underlying libraries and tools. So we
> just play for the stable API/ABI interfaces. Shipping internally a
> small set of libraries isn't enough - you need to ship js, libicu,
> erlang and so on.
Of course, the point I was making is that we do break* said interface
deliberately.
>> e.g. we're changing the default handling of encoding integers in mochijson2
>> among other things.
>
> Fortunately it's not very diverged from upstream - it just too old. I
> re-checked (as I did every couchdb release since 0.10.2) what was
> added in CouchDB internal copy - I don't see what changes from CouchDB
> copy are still missing in upstream repository (perhaps someone from
> Mochi is monitoring your work and quickly applies changes if any).
We usually propose patches upstream where they get reviewed. Historically
we have a pretty good acceptance rate.
* I did check back with the integer issue that I though didn't make it but
it seems latest mochijson2.erl has that patch, so sorry for the noise.
Robert is currently checking for other differences.
>> I'd strongly suggest to not build a package that is seemingly like our
>> release but has subtle differences or bugs.
>
> It hasn't any known bugs related to my changes (i mean I wasn't aware
> of them - not that they doesn't exist). So that's a pure speculation.
> On the other hand your copy of mochiweb does contains known defects
> already fixed in upstream. Not sure whether they affect CouchDB
> operation though.
There's plenty of defects we know of that ship with stable releases.
The issue that worries us here is that you may introduce subtle bugs
due to your packaging policy that ends up being an unhappy situation
for CouchDB users.
This is not a passing the blame situation, I understand that the
priorities of a package manager are quite different from a project.
I just want to make sure you understand my position. Besides, you're
quite the opposite than most distributions which ship out of date
dependencies that we can't rely on, so the forward/parity is rather
unique (at least for me :).
That said, your work is highly appreciated and we hope we can make
it easier for you down the road.
I wonder what the best solution here is. The problem with updating
to the latest Mochiweb right before a release is that subtle issues
don't have time surfacing in trunk. To ensure stable releases we
should be conservative with updates (modulo critical bug fixes of
course) or upstream libraries. How can we best ensure to keep them
up to date enough while not compromising a release's stability?
Cheers
Jan
--
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Peter Lemenkov <le...@gmail.com>.
Hello
2010/11/26 Jan Lehnardt <ja...@apache.org>:
> We always patch Mochiweb to fit our needs that are divergent from the
> upstream project. It isn't safe to assume an independent cut of Mochiweb
> will just work.
Well, it isn't even safe to assume that it will work on an arbitrary
system with different set of underlying libraries and tools. So we
just play for the stable API/ABI interfaces. Shipping internally a
small set of libraries isn't enough - you need to ship js, libicu,
erlang and so on.
> e.g. we're changing the default handling of encoding integers in mochijson2
> among other things.
Fortunately it's not very diverged from upstream - it just too old. I
re-checked (as I did every couchdb release since 0.10.2) what was
added in CouchDB internal copy - I don't see what changes from CouchDB
copy are still missing in upstream repository (perhaps someone from
Mochi is monitoring your work and quickly applies changes if any).
> I'd strongly suggest to not build a package that is seemingly like our
> release but has subtle differences or bugs.
It hasn't any known bugs related to my changes (i mean I wasn't aware
of them - not that they doesn't exist). So that's a pure speculation.
On the other hand your copy of mochiweb does contains known defects
already fixed in upstream. Not sure whether they affect CouchDB
operation though.
--
With best regards, Peter Lemenkov.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Jan Lehnardt <ja...@apache.org>.
On 26 Nov 2010, at 07:14, Peter Lemenkov wrote:
> 2010/11/26 Robert Newson <ro...@gmail.com>:
>> pretty sure ibrowse and mochiweb have important tweaks, using versions
>> without them is broken.
>
> I'm not sure about mochiweb (since couchdb contains very old svn
> snapshot, and it's hard to check for the differences) but regarding
> ibrowse - couchdb is only 1 patch behind the 2.1.0. I'll plan to
> backport in into 2.1.0 shipped in Fedora.
We always patch Mochiweb to fit our needs that are divergent from the
upstream project. It isn't safe to assume an independent cut of Mochiweb
will just work.
e.g. we're changing the default handling of encoding integers in mochijson2
among other things.
I'd strongly suggest to not build a package that is seemingly like our
release but has subtle differences or bugs.
Cheers
Jan
--
>
> So
>
> a) These libraries are not so different from upstream ones (at least
> true for ibrowse)
> b) CouchDB in Fedora isn't so deadly broken as someone of you may think
>
> In fact I'm watching very closely the development of both ibrowse and
> mochiweb so I'm aware of the changes made or just proposed by CouchDB
> folks.
>
> --
> With best regards, Peter Lemenkov.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Filipe David Manana <fd...@apache.org>.
On Fri, Nov 26, 2010 at 6:14 AM, Peter Lemenkov <le...@gmail.com> wrote:
> 2010/11/26 Robert Newson <ro...@gmail.com>:
>> pretty sure ibrowse and mochiweb have important tweaks, using versions
>> without them is broken.
>
> I'm not sure about mochiweb (since couchdb contains very old svn
> snapshot, and it's hard to check for the differences) but regarding
> ibrowse - couchdb is only 1 patch behind the 2.1.0. I'll plan to
> backport in into 2.1.0 shipped in Fedora.
CouchDB's ibrowse is version 2.1.0 with one extra patch (already
submitted upstream long time ago):
http://svn.apache.org/viewvc?view=revision&revision=1034405
>
> So
>
> a) These libraries are not so different from upstream ones (at least
> true for ibrowse)
> b) CouchDB in Fedora isn't so deadly broken as someone of you may think
>
> In fact I'm watching very closely the development of both ibrowse and
> mochiweb so I'm aware of the changes made or just proposed by CouchDB
> folks.
>
> --
> With best regards, Peter Lemenkov.
>
--
Filipe David Manana,
fdmanana@gmail.com, fdmanana@apache.org
"Reasonable men adapt themselves to the world.
Unreasonable men adapt the world to themselves.
That's why all progress depends on unreasonable men."
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Peter Lemenkov <le...@gmail.com>.
2010/11/26 Robert Newson <ro...@gmail.com>:
> pretty sure ibrowse and mochiweb have important tweaks, using versions
> without them is broken.
I'm not sure about mochiweb (since couchdb contains very old svn
snapshot, and it's hard to check for the differences) but regarding
ibrowse - couchdb is only 1 patch behind the 2.1.0. I'll plan to
backport in into 2.1.0 shipped in Fedora.
So
a) These libraries are not so different from upstream ones (at least
true for ibrowse)
b) CouchDB in Fedora isn't so deadly broken as someone of you may think
In fact I'm watching very closely the development of both ibrowse and
mochiweb so I'm aware of the changes made or just proposed by CouchDB
folks.
--
With best regards, Peter Lemenkov.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Benoit Chesneau <bc...@gmail.com>.
On Fri, Nov 26, 2010 at 11:59 AM, Peter Lemenkov <le...@gmail.com> wrote:
> 2010/11/26 Benoit Chesneau <bc...@gmail.com>:
>> On Fri, Nov 26, 2010 at 11:30 AM, Peter Lemenkov <le...@gmail.com> wrote:
>>
>>> Unfortunately, right now CouchDB's own copy of mochiweb misses many
>>> bugfixes already available in upstream. So it's worth to rebase
>>> couchdb on top of recently released 1.4.1 rather than trying to find
>>> something valuable in src/mochiweb.
>
>> Can you open a ticket about that an deventually list "important" bits
>> we missed ? That could help :)
>
> I don't see any point in duplicating upstream bug-tracker, so you may
> take a look right there:
>
> https://github.com/mochi/mochiweb/issues/closed
>
Sorry but you made statements we should rebase against another version
because of change in upstream mohiweb. So if you consider there is an
issue in couchdb,that need to be fixed let us know. We already
imported some changes in our code base:
https://github.com/apache/couchdb/tree/trunk/src/mochiweb/
Some other changes in latest mochiweb aren't only fixes. 1.0.2 is a
release for fixes as far as I can tell. 1.1 is another deal.
- benoit
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Peter Lemenkov <le...@gmail.com>.
2010/11/26 Benoit Chesneau <bc...@gmail.com>:
> On Fri, Nov 26, 2010 at 11:30 AM, Peter Lemenkov <le...@gmail.com> wrote:
>
>> Unfortunately, right now CouchDB's own copy of mochiweb misses many
>> bugfixes already available in upstream. So it's worth to rebase
>> couchdb on top of recently released 1.4.1 rather than trying to find
>> something valuable in src/mochiweb.
> Can you open a ticket about that an deventually list "important" bits
> we missed ? That could help :)
I don't see any point in duplicating upstream bug-tracker, so you may
take a look right there:
https://github.com/mochi/mochiweb/issues/closed
--
With best regards, Peter Lemenkov.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Benoit Chesneau <bc...@gmail.com>.
On Fri, Nov 26, 2010 at 11:30 AM, Peter Lemenkov <le...@gmail.com> wrote:
> Unfortunately, right now CouchDB's own copy of mochiweb misses many
> bugfixes already available in upstream. So it's worth to rebase
> couchdb on top of recently released 1.4.1 rather than trying to find
> something valuable in src/mochiweb.
>
Can you open a ticket about that an deventually list "important" bits
we missed ? That could help :)
- benoît
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Jan Lehnardt <ja...@apache.org>.
On 26 Nov 2010, at 11:30, Peter Lemenkov wrote:
> 2010/11/26 Dirkjan Ochtman <di...@ochtman.nl>:
>> On Fri, Nov 26, 2010 at 00:27, Robert Newson <ro...@gmail.com> wrote:
>>> pretty sure ibrowse and mochiweb have important tweaks, using versions
>>> without them is broken.
>>
>> Are these being pushed upstream?
>
> Unfortunately, right now CouchDB's own copy of mochiweb misses many
> bugfixes already available in upstream. So it's worth to rebase
> couchdb on top of recently released 1.4.1 rather than trying to find
> something valuable in src/mochiweb.
The flip side of that is that you can't know if newer than imported
versions introduce any subtle issues with CouchDB. In an ideal world
our test suite would catch these, but if we were to live in an ideal
world, we wouldn't have dependency hell like this. Besides, tests can
only verify the existence of bugs, not their absence (Dijkstra IIRC).
Cheers
Jan
--
>
> Much better scenario would be to disallow generally bad practice of
> bundling sources from different projects, but this would be painful
> because even many popular Linux distributions (not to mention of Mac
> OS X and Windoze) lacks of good Erlang support right of the box.
>
> --
> With best regards, Peter Lemenkov.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Jan Lehnardt <ja...@apache.org>.
On 26 Nov 2010, at 13:22, Dirkjan Ochtman wrote:
> On Fri, Nov 26, 2010 at 12:55, Jan Lehnardt <ja...@apache.org> wrote:
>> This breaks the basic requirement of not requiring networks access for
>> an installation which I remember Noah is advocating.
>
> Well, with doing it in the build system I really meant something that
> pulls them in for the tarball (or a tarball -- have one with, and one
> without the deps).
>
>> I think we can agree on that packaging and distribution is a tricky
>> business and that there are not general right or wrong ways of doing
>> it, just different priorities for certain trade-offs.
>
> Certainly agreed, though I think that, if you're putting deps in your
> version control tree, a certain effort is required to:
>
> (a) track how our version and upstream's compare, e.g. at the least
> make clear which version or revision our version is based on (oauth
> has none of this, apparently), and
>
> (b) spend quite some time making sure patches we need (or some
> variant of them) go upstream (for example, I'm not seeing any upstream
> issue about getting SSL into mochiweb -- maybe there has been other
> contact).
I think handle both well (erlang_oauth has been neglected, but I'm working
on updating it as we speak, it didn't progress much) and if you follow
dev@ and commits@ you can clearly see that.
As for SSL an Mochiweb, it got SSL first. We (Robert) just updated the
bundled Mochiweb version and started using the new SSL APIs. It's not that
we added SSL to Mochiweb and then didn't push it back (modulo what Robert
just wrote about that patch).
Cheers
Jan
--
>
> Cheers,
>
> dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Dirkjan Ochtman <di...@ochtman.nl>.
On Fri, Nov 26, 2010 at 14:32, Jan Lehnardt <ja...@apache.org> wrote:
> Understandable. We could make things clearer in CHANGES.
That would be much appreciated.
Cheers,
Dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Jan Lehnardt <ja...@apache.org>.
On 26 Nov 2010, at 14:23, Dirkjan Ochtman wrote:
> On Fri, Nov 26, 2010 at 13:53, Robert Newson <ro...@gmail.com> wrote:
>> Obivously I agree we should be using official releases of mochiweb and
>> ibrowse but that's not the case today, discarding the patches can
>> potentially break CouchDB. I don't like it, but the fact that part of
>> our source tree is imported from somewhere else is the current state
>> of play. Happily, I think the delta is getting smaller all the time as
>> the fixes and enhancements are applied upstream.
>
> That's good to hear.
>
> On Fri, Nov 26, 2010 at 14:13, Jan Lehnardt <ja...@apache.org> wrote:
>> I think handle both well (erlang_oauth has been neglected, but I'm working
>> on updating it as we speak, it didn't progress much) and if you follow
>> dev@ and commits@ you can clearly see that.
>
> Okay, sorry if I overreacted a bit, good to hear you're cleaning up
> oauth (and sorry, I don't have the bandwidth to follow commits@).
Understandable. We could make things clearer in CHANGES.
FWIW:
http://svn.apache.org/viewvc?rev=1039345&view=rev
https://github.com/tim/erlang-oauth/pull/2
Cheers
Jan
--
>
>> As for SSL an Mochiweb, it got SSL first. We (Robert) just updated the
>> bundled Mochiweb version and started using the new SSL APIs. It's not that
>> we added SSL to Mochiweb and then didn't push it back (modulo what Robert
>> just wrote about that patch).
>
> Okay, sounds great, it wasn't clear from the commit message that
> updated mochiweb.
>
> Cheers,
>
> Dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Dirkjan Ochtman <di...@ochtman.nl>.
On Fri, Nov 26, 2010 at 13:53, Robert Newson <ro...@gmail.com> wrote:
> Obivously I agree we should be using official releases of mochiweb and
> ibrowse but that's not the case today, discarding the patches can
> potentially break CouchDB. I don't like it, but the fact that part of
> our source tree is imported from somewhere else is the current state
> of play. Happily, I think the delta is getting smaller all the time as
> the fixes and enhancements are applied upstream.
That's good to hear.
On Fri, Nov 26, 2010 at 14:13, Jan Lehnardt <ja...@apache.org> wrote:
> I think handle both well (erlang_oauth has been neglected, but I'm working
> on updating it as we speak, it didn't progress much) and if you follow
> dev@ and commits@ you can clearly see that.
Okay, sorry if I overreacted a bit, good to hear you're cleaning up
oauth (and sorry, I don't have the bandwidth to follow commits@).
> As for SSL an Mochiweb, it got SSL first. We (Robert) just updated the
> bundled Mochiweb version and started using the new SSL APIs. It's not that
> we added SSL to Mochiweb and then didn't push it back (modulo what Robert
> just wrote about that patch).
Okay, sounds great, it wasn't clear from the commit message that
updated mochiweb.
Cheers,
Dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Robert Newson <ro...@gmail.com>.
for mochiweb at least one important tweak was in mochijson2.erl
(COUCHDB-796). Using a mochiweb without this fix undoes a resolved
ticket. Additionally I had to do some extra work to get the HTTPS
support working (though it's more than possible that this was also
fixed upstream since I upgraded; I haven't checked).
Obivously I agree we should be using official releases of mochiweb and
ibrowse but that's not the case today, discarding the patches can
potentially break CouchDB. I don't like it, but the fact that part of
our source tree is imported from somewhere else is the current state
of play. Happily, I think the delta is getting smaller all the time as
the fixes and enhancements are applied upstream.
I suggest this discussion should move out of this thread which should
focus on voting on the release artifacts that Paul kindly assembled.
B.
On Fri, Nov 26, 2010 at 12:22 PM, Dirkjan Ochtman <di...@ochtman.nl> wrote:
> On Fri, Nov 26, 2010 at 12:55, Jan Lehnardt <ja...@apache.org> wrote:
>> This breaks the basic requirement of not requiring networks access for
>> an installation which I remember Noah is advocating.
>
> Well, with doing it in the build system I really meant something that
> pulls them in for the tarball (or a tarball -- have one with, and one
> without the deps).
>
>> I think we can agree on that packaging and distribution is a tricky
>> business and that there are not general right or wrong ways of doing
>> it, just different priorities for certain trade-offs.
>
> Certainly agreed, though I think that, if you're putting deps in your
> version control tree, a certain effort is required to:
>
> (a) track how our version and upstream's compare, e.g. at the least
> make clear which version or revision our version is based on (oauth
> has none of this, apparently), and
>
> (b) spend quite some time making sure patches we need (or some
> variant of them) go upstream (for example, I'm not seeing any upstream
> issue about getting SSL into mochiweb -- maybe there has been other
> contact).
>
> Cheers,
>
> dirkjan
>
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Dirkjan Ochtman <di...@ochtman.nl>.
On Fri, Nov 26, 2010 at 12:55, Jan Lehnardt <ja...@apache.org> wrote:
> This breaks the basic requirement of not requiring networks access for
> an installation which I remember Noah is advocating.
Well, with doing it in the build system I really meant something that
pulls them in for the tarball (or a tarball -- have one with, and one
without the deps).
> I think we can agree on that packaging and distribution is a tricky
> business and that there are not general right or wrong ways of doing
> it, just different priorities for certain trade-offs.
Certainly agreed, though I think that, if you're putting deps in your
version control tree, a certain effort is required to:
(a) track how our version and upstream's compare, e.g. at the least
make clear which version or revision our version is based on (oauth
has none of this, apparently), and
(b) spend quite some time making sure patches we need (or some
variant of them) go upstream (for example, I'm not seeing any upstream
issue about getting SSL into mochiweb -- maybe there has been other
contact).
Cheers,
dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Jan Lehnardt <ja...@apache.org>.
On 26 Nov 2010, at 11:34, Dirkjan Ochtman wrote:
> On Fri, Nov 26, 2010 at 11:30, Peter Lemenkov <le...@gmail.com> wrote:
>> Much better scenario would be to disallow generally bad practice of
>> bundling sources from different projects, but this would be painful
>> because even many popular Linux distributions (not to mention of Mac
>> OS X and Windoze) lacks of good Erlang support right of the box.
>
> The solution for that is a build system script to pull tarballs of the
> web, not keeping full copies (in some cases not associated with any
> upstream URL or revision -- for oauth) in the VCS, IMO.
This breaks the basic requirement of not requiring networks access for
an installation which I remember Noah is advocating.
I think we can agree on that packaging and distribution is a tricky
business and that there are not general right or wrong ways of doing
it, just different priorities for certain trade-offs.
Cheers
Jan
--
> I can
> understand that it made sense to bundle things while Couch was still
> getting started, but it should be moving to using released upstream
> versions. On Gentoo (where I maintain the ebuild) we definitely also
> frown on bundled packages.
>
> Cheers,
>
> Dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Dirkjan Ochtman <di...@ochtman.nl>.
On Fri, Nov 26, 2010 at 11:30, Peter Lemenkov <le...@gmail.com> wrote:
> Much better scenario would be to disallow generally bad practice of
> bundling sources from different projects, but this would be painful
> because even many popular Linux distributions (not to mention of Mac
> OS X and Windoze) lacks of good Erlang support right of the box.
The solution for that is a build system script to pull tarballs of the
web, not keeping full copies (in some cases not associated with any
upstream URL or revision -- for oauth) in the VCS, IMO. I can
understand that it made sense to bundle things while Couch was still
getting started, but it should be moving to using released upstream
versions. On Gentoo (where I maintain the ebuild) we definitely also
frown on bundled packages.
Cheers,
Dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Peter Lemenkov <le...@gmail.com>.
2010/11/26 Dirkjan Ochtman <di...@ochtman.nl>:
> On Fri, Nov 26, 2010 at 00:27, Robert Newson <ro...@gmail.com> wrote:
>> pretty sure ibrowse and mochiweb have important tweaks, using versions
>> without them is broken.
>
> Are these being pushed upstream?
Unfortunately, right now CouchDB's own copy of mochiweb misses many
bugfixes already available in upstream. So it's worth to rebase
couchdb on top of recently released 1.4.1 rather than trying to find
something valuable in src/mochiweb.
Much better scenario would be to disallow generally bad practice of
bundling sources from different projects, but this would be painful
because even many popular Linux distributions (not to mention of Mac
OS X and Windoze) lacks of good Erlang support right of the box.
--
With best regards, Peter Lemenkov.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Dirkjan Ochtman <di...@ochtman.nl>.
On Fri, Nov 26, 2010 at 00:27, Robert Newson <ro...@gmail.com> wrote:
> pretty sure ibrowse and mochiweb have important tweaks, using versions
> without them is broken.
Are these being pushed upstream?
Cheers,
Dirkjan
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Robert Newson <ro...@gmail.com>.
pretty sure ibrowse and mochiweb have important tweaks, using versions
without them is broken.
Not sure I understand the release process if it allows voting on
things other than the release artifact.
B.
On Thu, Nov 25, 2010 at 11:08 PM, Noah Slater <ns...@apache.org> wrote:
> People are free to give a +1 for any reason. :)
>
> However, I would be concerned that Fedora is about to ship a broken Couch.
>
> On 25 Nov 2010, at 22:07, Robert Newson wrote:
>
>> I'm pretty sure the vote can only relate to the artifact itself
>> without modifications.
>>
>> B.
>>
>> On Thu, Nov 25, 2010 at 10:05 PM, Sebastian Cohnen
>> <se...@googlemail.com> wrote:
>>> not sure about the 1.0.2 release, but wasn't at least ibrowse patched for couchdb? building against already packaged once may break some stuff.
>>>
>>> On 25.11.2010, at 22:17, Peter Lemenkov wrote:
>>>
>>>> 2010/11/25 Paul Davis <pa...@gmail.com>:
>>>>> Hello,
>>>>>
>>>>> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>>>>
>>>> Builds fine and passes all etap tests (no futon tests were performed) on
>>>>
>>>> * Fedora 12 for i686, x86_64, ppc and ppc64 with Erlang R13B04,
>>>> ibrowse-2.1.0, mochiweb-1.4.1, js-1.7.0, curl-7.19 and libicu-4.2.1:
>>>>
>>>> http://koji.fedoraproject.org/koji/taskinfo?taskID=2626994
>>>>
>>>> * Fedora 14 for i686 and x86_64 with Erlang R14B, ibrowse-2.1.0,
>>>> mochiweb-1.4.1, js-1.7.0, curl-7.21 and libicu-4.4.1:
>>>>
>>>> http://koji.fedoraproject.org/koji/taskinfo?taskID=2627004
>>>>
>>>> So my +1 here. Please, note, that according to Fedora guidelines I
>>>> patched out all bundled libraries in couchdb. Namely - oauth, ibrowse,
>>>> mochiweb,etap libraries were removed from tarball, and couchdb was
>>>> built against already packaged ones.
>>>> --
>>>> With best regards, Peter Lemenkov.
>>>
>>>
>
>
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Noah Slater <ns...@apache.org>.
People are free to give a +1 for any reason. :)
However, I would be concerned that Fedora is about to ship a broken Couch.
On 25 Nov 2010, at 22:07, Robert Newson wrote:
> I'm pretty sure the vote can only relate to the artifact itself
> without modifications.
>
> B.
>
> On Thu, Nov 25, 2010 at 10:05 PM, Sebastian Cohnen
> <se...@googlemail.com> wrote:
>> not sure about the 1.0.2 release, but wasn't at least ibrowse patched for couchdb? building against already packaged once may break some stuff.
>>
>> On 25.11.2010, at 22:17, Peter Lemenkov wrote:
>>
>>> 2010/11/25 Paul Davis <pa...@gmail.com>:
>>>> Hello,
>>>>
>>>> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>>>
>>> Builds fine and passes all etap tests (no futon tests were performed) on
>>>
>>> * Fedora 12 for i686, x86_64, ppc and ppc64 with Erlang R13B04,
>>> ibrowse-2.1.0, mochiweb-1.4.1, js-1.7.0, curl-7.19 and libicu-4.2.1:
>>>
>>> http://koji.fedoraproject.org/koji/taskinfo?taskID=2626994
>>>
>>> * Fedora 14 for i686 and x86_64 with Erlang R14B, ibrowse-2.1.0,
>>> mochiweb-1.4.1, js-1.7.0, curl-7.21 and libicu-4.4.1:
>>>
>>> http://koji.fedoraproject.org/koji/taskinfo?taskID=2627004
>>>
>>> So my +1 here. Please, note, that according to Fedora guidelines I
>>> patched out all bundled libraries in couchdb. Namely - oauth, ibrowse,
>>> mochiweb,etap libraries were removed from tarball, and couchdb was
>>> built against already packaged ones.
>>> --
>>> With best regards, Peter Lemenkov.
>>
>>
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Robert Newson <ro...@gmail.com>.
I'm pretty sure the vote can only relate to the artifact itself
without modifications.
B.
On Thu, Nov 25, 2010 at 10:05 PM, Sebastian Cohnen
<se...@googlemail.com> wrote:
> not sure about the 1.0.2 release, but wasn't at least ibrowse patched for couchdb? building against already packaged once may break some stuff.
>
> On 25.11.2010, at 22:17, Peter Lemenkov wrote:
>
>> 2010/11/25 Paul Davis <pa...@gmail.com>:
>>> Hello,
>>>
>>> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>>
>> Builds fine and passes all etap tests (no futon tests were performed) on
>>
>> * Fedora 12 for i686, x86_64, ppc and ppc64 with Erlang R13B04,
>> ibrowse-2.1.0, mochiweb-1.4.1, js-1.7.0, curl-7.19 and libicu-4.2.1:
>>
>> http://koji.fedoraproject.org/koji/taskinfo?taskID=2626994
>>
>> * Fedora 14 for i686 and x86_64 with Erlang R14B, ibrowse-2.1.0,
>> mochiweb-1.4.1, js-1.7.0, curl-7.21 and libicu-4.4.1:
>>
>> http://koji.fedoraproject.org/koji/taskinfo?taskID=2627004
>>
>> So my +1 here. Please, note, that according to Fedora guidelines I
>> patched out all bundled libraries in couchdb. Namely - oauth, ibrowse,
>> mochiweb,etap libraries were removed from tarball, and couchdb was
>> built against already packaged ones.
>> --
>> With best regards, Peter Lemenkov.
>
>
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Sebastian Cohnen <se...@googlemail.com>.
not sure about the 1.0.2 release, but wasn't at least ibrowse patched for couchdb? building against already packaged once may break some stuff.
On 25.11.2010, at 22:17, Peter Lemenkov wrote:
> 2010/11/25 Paul Davis <pa...@gmail.com>:
>> Hello,
>>
>> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
>
> Builds fine and passes all etap tests (no futon tests were performed) on
>
> * Fedora 12 for i686, x86_64, ppc and ppc64 with Erlang R13B04,
> ibrowse-2.1.0, mochiweb-1.4.1, js-1.7.0, curl-7.19 and libicu-4.2.1:
>
> http://koji.fedoraproject.org/koji/taskinfo?taskID=2626994
>
> * Fedora 14 for i686 and x86_64 with Erlang R14B, ibrowse-2.1.0,
> mochiweb-1.4.1, js-1.7.0, curl-7.21 and libicu-4.4.1:
>
> http://koji.fedoraproject.org/koji/taskinfo?taskID=2627004
>
> So my +1 here. Please, note, that according to Fedora guidelines I
> patched out all bundled libraries in couchdb. Namely - oauth, ibrowse,
> mochiweb,etap libraries were removed from tarball, and couchdb was
> built against already packaged ones.
> --
> With best regards, Peter Lemenkov.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Peter Lemenkov <le...@gmail.com>.
2010/11/25 Paul Davis <pa...@gmail.com>:
> Hello,
>
> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
Builds fine and passes all etap tests (no futon tests were performed) on
* Fedora 12 for i686, x86_64, ppc and ppc64 with Erlang R13B04,
ibrowse-2.1.0, mochiweb-1.4.1, js-1.7.0, curl-7.19 and libicu-4.2.1:
http://koji.fedoraproject.org/koji/taskinfo?taskID=2626994
* Fedora 14 for i686 and x86_64 with Erlang R14B, ibrowse-2.1.0,
mochiweb-1.4.1, js-1.7.0, curl-7.21 and libicu-4.4.1:
http://koji.fedoraproject.org/koji/taskinfo?taskID=2627004
So my +1 here. Please, note, that according to Fedora guidelines I
patched out all bundled libraries in couchdb. Namely - oauth, ibrowse,
mochiweb,etap libraries were removed from tarball, and couchdb was
built against already packaged ones.
--
With best regards, Peter Lemenkov.
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Jan Lehnardt <ja...@apache.org>.
On 25 Nov 2010, at 19:43, Paul Davis wrote:
> Hello,
>
> I would like call a vote for the Apache CouchDB 1.0.2 release, round 1.
> [...]
> We are voting on the following release artifacts:
>
> http://people.apache.org/~davisp/dist/1.0.2/
Great first round Paul (and thanks to Noah for helping :)
Signature checks out, both hashes, too.
Mac OS X 10.6.5, Erlang R14B
make check, Browser tests in Firefox and Chrome all pass
Ubuntu (I think 10.04) R13B2
make check
+1
Cheers
Jan
--
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by Robert Newson <ro...@gmail.com>.
FYI: I've updated couchdb to mochiweb 1.4.1 on a branch called mochiweb-1.4.1.
The only tweaks are to replace Makefile with our Makefile.am and to
retain our mochiweb.app.in (with updated vsn).
I've verified that Futon tests work, that the integer fix is
incorporated and that https support still works. I'm going to let it
sit on a branch for a few days and would appreciate it if others would
try it out.
Some Futon tests fail when run through https, notably
attachment_ranges, but since the tests pass reliably via http it'll
probably be something silly.
B.
On Fri, Nov 26, 2010 at 3:41 PM, till <ti...@php.net> wrote:
> On Thu, Nov 25, 2010 at 7:43 PM, Paul Davis <pa...@gmail.com> wrote:
> [ ... ]
>> We are voting on the following release artifacts:
>>
>> http://people.apache.org/~davisp/dist/1.0.2/
>>
>
> [one of my older production servers]
>
> FreeBSD 6.x/i386
> Erlang R13B04
> Spidermonkey 1.7.x
> gmake check = great success
>
> [in a virtualbox]
>
> FreeBSD 8.1/i386
> Erlang R14B
> Spidermonkey 1.7.x
> gmake check = great success
>
> +1
>
Re: [VOTE] Apache CouchDB 1.0.2 release, Round 1
Posted by till <ti...@php.net>.
On Thu, Nov 25, 2010 at 7:43 PM, Paul Davis <pa...@gmail.com> wrote:
[ ... ]
> We are voting on the following release artifacts:
>
> http://people.apache.org/~davisp/dist/1.0.2/
>
[one of my older production servers]
FreeBSD 6.x/i386
Erlang R13B04
Spidermonkey 1.7.x
gmake check = great success
[in a virtualbox]
FreeBSD 8.1/i386
Erlang R14B
Spidermonkey 1.7.x
gmake check = great success
+1