#pyramid

yah-veh-skript

[pyramid_tm] mmerickel pushed 1 new commit to master: http://git.io/vBRqM

pyramid_tm/master f188c5b Michael Merickel: run docs only against python3

[pyramid] mmerickel pushed 3 new commits to master: http://git.io/vBROX

pyramid/master a708d35 Michael Merickel: remove py2-docs from tox.ini and reorder so coverage is last

pyramid/master 0030fba Michael Merickel: add docs to travis builds

pyramid/master 2d9ed3f Michael Merickel: Merge pull request #2152 from mmerickel/fix/drop-py2-docs...

[pyramid] mmerickel pushed 2 new commits to 1.6-branch: http://git.io/vBR36

pyramid/1.6-branch 04d2ae7 Michael Merickel: remove py2-docs from tox.ini and reorder so coverage is last

pyramid/1.6-branch 7bff487 Michael Merickel: add docs to travis builds

stevepiercy_away: couldn't help but notice that HACKING.txt is slightly different on 1.6-branch and master

dunno if it's something backports missed or whatever else, but just pointing it out fwiw

[pyramid] mmerickel pushed 1 new commit to 1.6-branch: http://git.io/vBRsj

pyramid/1.6-branch 739788a Michael Merickel: oops, drop pep8 from 1.6-branch

[pyramid_jinja2] mmerickel pushed 1 new commit to master: http://git.io/vBRci

pyramid_jinja2/master f978ec6 Michael Merickel: use tox that tests py2+py3 coverage

[pyramid_jinja2] mmerickel pushed 1 new commit to master: http://git.io/vBRcS

pyramid_jinja2/master e55445c Michael Merickel: add full coverage and docs to travis

hi, anyone using statsd + pyramid?

ricoalpha: substanced (a pyramid app server) uses statsd

so tentative plan for pyramid_tm is to refactor the tween into 2 tweens

interesting

one retryable tween under excview and a new tween over excview that handles abort/commit

this way you can still use the transactions inside excviews

this also gives certain exception views the ability to override the exception and mark the transaction to be committed, possibly

anyway... just saying it out loud before I start playing with it

oh cool

that makes sense

cguardia: what library did you use?

ricoalpha: statsd

and did you configure the settings through pyramid?

did you use this library too? https://pypi.python.org/pypi/pyramid_metrics/0.1.4

ricoalpha: no pyramid_metrics...you can see the integration code here: https://github.com/Pylons/substanced/blob/maste...

it never gets old http://www.commitstrip.com/wp-content/uploads/2...

cool

would be perfect to have it as a seperate interface to include in pyramid

[pyramid_tm] mmerickel pushed 1 new commit to master: http://git.io/vBRQH

pyramid_tm/master 8b0a1b5 Michael Merickel: remove py2.5 syntax

raydeo ok, i'll have a look

raydeo: lol =)

oddly I wrote (wht I can tell) is a very sane integration test for pyramid_tm and it fails on 1.2 and 1.3

i would say that the guy in that comic strip is me, but he's probably wearing pants

Is 1.6-branch not pep8 BTW?

I see what you did there

x58: I thought it was but apparently not so meh

raydeo: Time to move on from 1.3 :P

x58: well I'm fairly certain these tests shouldn't fail on 1.3

Speaking of pants, got pizza coming so I am going to find some.

raydeo yeah, looks like master was not backported to 1.6 for HACKING.txt, as we banished submodules

No more version branching until we are ready to release...

i'll fix

no request methods in 1.2 sucks

x58 raydeo this is a policy question for you: https://github.com/Pylons/pyramid/pull/2148

I'm not super excited about it

I'm a no on it.

considering adding it to that file usually means work for me... and I don't use dash

I liked Dash at one point, but then it became nasty shareware that didn't work right on the docsets I was already using.

RELEASING.txt is a one-man affiar generally

[pyramid] stevepiercy pushed 2 new commits to 1.6-branch: http://git.io/vBRNE

pyramid/1.6-branch 3bf245b Steve Piercy: backport master to 1.6-branch

pyramid/1.6-branch ab49af1 Steve Piercy: Merge pull request #2153 from stevepiercy/1.6-branch...

ok, i wasn't super excited about it myself, i'll close the PR

ah fun... damn I don't like a lot of our tests

there's such an anti-integration testing stigma in our official projects

I feel like any time I get into them and look at the tests there's so much mocked out that it's hard to tell what's going on

like earlier this week when I got into pyramid_nacl_session and the tests weren't even using nacl's SecretBox

lol

or now in pyramid_tm where I write literally one integration tests that tries the tm on a real request and it fails immediately on pyramid 1.2/1.3, despite lots of tests testing the feature that fails because just too much is mocked out

Remember that fix I was trying to make where I was complaining about the tests not actually testing a code path?

yeah

That's cause so much was mocked out that it was incidentally testing it.

But in a way that didn't properly test that the functionality was correct

it just happened to hit the code path and thus be covered.

yeah not shipping a single integration test is a big no no to me

I agree using unit tests for all the corner cases is great... but you've gotta make sure the damn thing actually works

So would pyramid_tm even work on 1.3?

not atm!

well

I'm surprised we haven't had any complaints about it.

it's supposed to work on 1.3, dunno what's going on

but it's definitely broken on 1.2

despite us having tox explicitly test against 1.2 ;-)

special.

but it also seems like the request property isn't working on 1.3 :(

honestly not sure why 1.3 is broken atm

set_request_property seems to not be working or something

How are your getting your request object

?

webtest

router

Gotcha.

test is literally just sending a request into a dummy app