Release 0.9 bug candidates

Eli Stevens (Gmail)

unread,

Sep 20, 2012, 12:50:58 PM9/20/12

to couchdb-python

Alexander Shorin

unread,

Sep 20, 2012, 1:51:56 PM9/20/12

to couchdb...@googlegroups.com

> http://code.google.com/p/couchdb-python/issues/detail?id=194

Hm, I forgot to post a patch with tests. Got it.

> http://code.google.com/p/couchdb-python/issues/detail?id=179

Also added to todo list.

> http://code.google.com/p/couchdb-python/issues/detail?id=204

self.cache.pop(url, None) would be much more Pythonic way to do this.

> http://code.google.com/p/couchdb-python/issues/detail?id=213

Actually, there is nothing to do if those error is the only one since
it doesn't relevant to couchdb-python.

> http://code.google.com/p/couchdb-python/issues/detail?id=210
> http://code.google.com/p/couchdb-python/issues/detail?id=175

Exception masking fixed in 2908fd72e112
http://code.google.com/p/couchdb-python/source/detail?r=2908fd72e112b2ed9669733938ead1932a1eeefe
Anything else if JSON library implementation details. For example,
simplejson / stdlib doesn't support datetime encoding out of box while
ujson has, but your expectation upon result may fails (result would be
unix timestamp, surprise!).

About any others I have no comments, so silently agreed with you.

Also, there is missed one awesome thing, that is must have for normal
work with real CouchDB databases. It was mentioned in mail list a long
time ago and relevant to next issue:
http://code.google.com/p/couchdb-python/issues/detail?id=162
I'd like to highlight Matt's Database.iterview() method since it's
really useful for large views and works stable. I'm sure everyone had
implemented his own view paginator to reduce memory usage and speedup
result processing, so why not to have such common things out of box?
That's the only one new feature that I'd like to add to your's, Eli,
list, but it really worth it.

--
,,,^..^,,,

On Thu, Sep 20, 2012 at 8:50 PM, Eli Stevens (Gmail)
<wicke...@gmail.com> wrote:
> Either looks fixed already or has a patch available, should be applied
> and closed (or a note added explaining why what's provided isn't
> sufficient):
>
> http://code.google.com/p/couchdb-python/issues/detail?id=121
> http://code.google.com/p/couchdb-python/issues/detail?id=126 (waiting
> on an SVG?)
> http://code.google.com/p/couchdb-python/issues/detail?id=188

>

> --
> You received this message because you are subscribed to the Google Groups "CouchDB-Python" group.
> To post to this group, send email to couchdb...@googlegroups.com.
> To unsubscribe from this group, send email to couchdb-pytho...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/couchdb-python?hl=en.
>

Dirkjan Ochtman

unread,

Sep 21, 2012, 4:43:14 AM9/21/12

to couchdb...@googlegroups.com, Matt Goodall

Thanks for doing this!

On Thu, Sep 20, 2012 at 6:50 PM, Eli Stevens (Gmail)
<wicke...@gmail.com> wrote:
> Either looks fixed already or has a patch available, should be applied
> and closed (or a note added explaining why what's provided isn't
> sufficient):
>
> http://code.google.com/p/couchdb-python/issues/detail?id=121
> http://code.google.com/p/couchdb-python/issues/detail?id=126 (waiting
> on an SVG?)
> http://code.google.com/p/couchdb-python/issues/detail?id=188
> http://code.google.com/p/couchdb-python/issues/detail?id=194

Marked 121 as works for me, 188 as fixed. 126 is still waiting for
SVG, 194 should get some attention from Alexander.

> Needs attention:
>
> http://code.google.com/p/couchdb-python/issues/detail?id=144 (now sure
> why this is low; implies unittest breakage on 2.6+)

We need to come up with a solution to the issue that works with both
stdlib json and simplejson. I marked it as low because it's an issue
with the test suite itself, not with the underlying functionality.

> http://code.google.com/p/couchdb-python/issues/detail?id=178 (might be
> the same as a fixed bug)

Right, marked as a duplicate.

> http://code.google.com/p/couchdb-python/issues/detail?id=179

The latest patch from the comments seems to have a number of
unresolved issues (Alexander has indicated he will take a look).

> http://code.google.com/p/couchdb-python/issues/detail?id=190

I've ripped out validate_dbname(), so this is fixed, IMO.

> http://code.google.com/p/couchdb-python/issues/detail?id=196

I've brought the docs in line with other methods, being less specific
helps here.

> http://code.google.com/p/couchdb-python/issues/detail?id=204

Alexander, can you outline your ideas on the issue itself?

> http://code.google.com/p/couchdb-python/issues/detail?id=205

This definitely needs some attention, but is not really within my expertise.

> http://code.google.com/p/couchdb-python/issues/detail?id=208

Alexander seems to have ideas about this stuff.

> http://code.google.com/p/couchdb-python/issues/detail?id=210

Looks like it should be fixed.

> http://code.google.com/p/couchdb-python/issues/detail?id=213

Closed as works for me.

> Candidates for rejection:
>
> http://code.google.com/p/couchdb-python/issues/detail?id=173

Rejected.

> http://code.google.com/p/couchdb-python/issues/detail?id=175

Should be fixed.

> http://code.google.com/p/couchdb-python/issues/detail?id=187

Closed as invalid.

> http://code.google.com/p/couchdb-python/issues/detail?id=191

Closed as invalid.

> http://code.google.com/p/couchdb-python/issues/detail?id=192

Closed as invalid.

> http://code.google.com/p/couchdb-python/issues/detail?id=197

Closed as works for me.

> I purposely have not included any bugs for either the mapping portion
> of the code (I am not a customer, and so my opinion of what's
> important there is suspect), nor for the python view server (same, and
> others seem to have the situation well in hand). Would it be possible
> to get a classification of bugs by component or something? It would
> be useful to be able to filter the bug list by
> core/mapping/viewserver/tools.

I think we could just reflect this in the issue titles for now (e.g.
prefix title by "core: ", "mapping: ", "tools: ").

> It might also be nice to have some way to indicate "this bug
> represents a significant change, and needs a champion" in the same way
> that Alexander has stepped up to work on the view server. Perhaps a
> placeholder "Needs Champion" milestone?

I've added you and Alexander as Project Contributors and given you
EditIssue privileges, so you can now go through the issue tracker and
do your worst (I'd have done it yesterday, but it took some time to
find the right config panel -- GCode is a little convoluted
sometimes). Matt, I hope that's okay with you!

So, I think that leaves about 6 things in scope for 0.9:

http://code.google.com/p/couchdb-python/issues/detail?id=126
terrible logo, waiting for SVG version

http://code.google.com/p/couchdb-python/issues/detail?id=144
doctests fail on stdlib json vs simplejson

http://code.google.com/p/couchdb-python/issues/detail?id=179
couchdb-dump fails with unicode doc ids

http://code.google.com/p/couchdb-python/issues/detail?id=194
couchdb-dump doesn't propagate credentials correctly

http://code.google.com/p/couchdb-python/issues/detail?id=204
http.Session is not threadsafe or something

http://code.google.com/p/couchdb-python/issues/detail?id=205
don't fail with forked processes (resetting the connection pool)

Cheers,

Dirkjan

Eli Stevens (Gmail)

unread,

Sep 21, 2012, 1:50:26 PM9/21/12

to couchdb...@googlegroups.com

On Fri, Sep 21, 2012 at 1:43 AM, Dirkjan Ochtman <dir...@ochtman.nl> wrote:
> I've added you and Alexander as Project Contributors and given you
> EditIssue privileges, so you can now go through the issue tracker and
> do your worst

Awesome, thank you. :) I'll try and not abuse the privilege.
However, I'd like to state that my base nature is to be ruthlessly
pragmatic, and so (for example) my first inclination is to reject bugs
like the logo one, and let interested parties re-open it if / when a
.svg is available, etc. If I'm too hasty with something like that, I
100% will not be offended if a change gets reversed.

> (I'd have done it yesterday, but it took some time to
> find the right config panel -- GCode is a little convoluted
> sometimes).

Indeed. I can't seem to figure out how to change the milestone from
---- to Release-0.9. After poking around for a bit I discovered that
a bunch of input options appear when going down to add a comment, but
even the text-box-turns-into-a-select-box ... thingys ... that they
have don't seem to show release 0.9 as an option (just 1.0). Any
ideas?

> So, I think that leaves about 6 things in scope for 0.9:
>
> http://code.google.com/p/couchdb-python/issues/detail?id=126
> terrible logo, waiting for SVG version

How difficult is it to re-open an issue after it's been rejected?

> http://code.google.com/p/couchdb-python/issues/detail?id=144
> doctests fail on stdlib json vs simplejson

How important is supporting versions prior to 2.6?

> http://code.google.com/p/couchdb-python/issues/detail?id=179
> couchdb-dump fails with unicode doc ids
>
> http://code.google.com/p/couchdb-python/issues/detail?id=194
> couchdb-dump doesn't propagate credentials correctly
>
> http://code.google.com/p/couchdb-python/issues/detail?id=204
> http.Session is not threadsafe or something
>
> http://code.google.com/p/couchdb-python/issues/detail?id=205
> don't fail with forked processes (resetting the connection pool)

I can try and see about working on this one. It might take me a day
or so to get a test env set up.

Eli

Dirkjan Ochtman

unread,

Sep 21, 2012, 4:04:42 PM9/21/12

to couchdb...@googlegroups.com

On Fri, Sep 21, 2012 at 7:50 PM, Eli Stevens (Gmail)
<wicke...@gmail.com> wrote:
> On Fri, Sep 21, 2012 at 1:43 AM, Dirkjan Ochtman <dir...@ochtman.nl> wrote:
>> I've added you and Alexander as Project Contributors and given you
>> EditIssue privileges, so you can now go through the issue tracker and
>> do your worst
>
> Awesome, thank you. :) I'll try and not abuse the privilege.
> However, I'd like to state that my base nature is to be ruthlessly
> pragmatic, and so (for example) my first inclination is to reject bugs
> like the logo one, and let interested parties re-open it if / when a
> .svg is available, etc. If I'm too hasty with something like that, I
> 100% will not be offended if a change gets reversed.

I tend to also err on the side of cleaning things up, though perhaps
not as much as you do (see also below).

> Indeed. I can't seem to figure out how to change the milestone from
> ---- to Release-0.9. After poking around for a bit I discovered that
> a bunch of input options appear when going down to add a comment, but
> even the text-box-turns-into-a-select-box ... thingys ... that they
> have don't seem to show release 0.9 as an option (just 1.0). Any
> ideas?

I think you can just type it as a label, and it should start with
"Milestone-". Again, GCode can be a little convoluted. ;)

>> So, I think that leaves about 6 things in scope for 0.9:
>>
>> http://code.google.com/p/couchdb-python/issues/detail?id=126
>> terrible logo, waiting for SVG version
>
> How difficult is it to re-open an issue after it's been rejected?

Did you actually look at it today? It's getting some movement. I agree
that it's not that important, but it's also not *that* important that
we keep the bug list that short, IMO. Legitimate bugs are legitimate,
no need to close them if the only problem is that no one's willing to
do the work right now. If what's bugging you is that there's no nice
list to drive down before we get to a release, go ahead with the
tagging and just only look at the milestone bug list. :)

>> http://code.google.com/p/couchdb-python/issues/detail?id=144
>> doctests fail on stdlib json vs simplejson
>
> How important is supporting versions prior to 2.6?

I really don't care about anything pre-2.7 myself, not sure about
others. Perhaps support 2.5 for 0.9 and drop it after that?

> I can try and see about working on this one. It might take me a day
> or so to get a test env set up.

Cool, thanks?

Cheers,

Dirkjan

Eli Stevens (Gmail)

unread,

Sep 21, 2012, 4:16:35 PM9/21/12

to couchdb...@googlegroups.com

On Fri, Sep 21, 2012 at 1:04 PM, Dirkjan Ochtman <dir...@ochtman.nl> wrote:
> I think you can just type it as a label, and it should start with
> "Milestone-". Again, GCode can be a little convoluted. ;)

I was only able to get "Milestone-Release1.0" to show up when I did
that. :/ Is there some admin difference between the two? I haven't
seen where those get created (I suspect I don't have privs); maybe
there's a check^Wtextbox that needs a magic value?

> Did you actually look at it today?

I hadn't seen the movement when I wrote that; now that I have, I would
not use it as an example the way I did. Like I said, if you ever
disagree with something I've changed, I won't be put off by a revert.
:)

> I really don't care about anything pre-2.7 myself, not sure about
> others. Perhaps support 2.5 for 0.9 and drop it after that?

I also am only interested in 2.7+, but I will try not to be too noisy about it.

Cheers,
Eli

Dirkjan Ochtman

unread,

Sep 21, 2012, 4:22:43 PM9/21/12

to couchdb...@googlegroups.com

On Fri, Sep 21, 2012 at 10:16 PM, Eli Stevens (Gmail)
<wicke...@gmail.com> wrote:
> I was only able to get "Milestone-Release1.0" to show up when I did
> that. :/ Is there some admin difference between the two? I haven't
> seen where those get created (I suspect I don't have privs); maybe
> there's a check^Wtextbox that needs a magic value?

Ah yes, I found a widget to predefine a new label. Milestone-0.9
should be in there now.

Cheers,

Dirkjan

Reply all

Reply to author

Forward