#pyramid

how?

whats the complete price of a kit?

You can literally buy them off the shelf.

not 35$

I'm not going to argue about the unit price. See my post on The Register for exactly the same point.

now, you can slap linux on that

its better than RPI for your tox stuff

and you just buy it and plug it

2gb of ram

For real work you'd want an A57 based device and if you want a lot you wouldn't buy them from the electronics store with chassis

but you just suggested pi's are great for work

Yes, I know that I can get a dedicated device.

i saw how rpi's struggle with xbmc even

You are still missing the point. The π is driving the market for these devices. Without the π you wouldn't be seeing them,

they were on the market before pi actually ;-)

XBMC runs fine on mine. NFS support is crap

Charlie_X: with a nice skin?

FFS.

those things are basicly android tablets without the screen

Why the fuck would I care about the skin? I switch my screen off when I'm not watching a film.

Charlie_X: i like nice looking things and improved navigation, im just saying PI is underpowered for xbmc

Phones are driving ARM. π has grabbed headlines for cheap ARM hardware

i think it often had issues with 1080p?

I've no problems with 1080p. But 100 MBit might be a bottleneck

100mbit?

I think average blu-ray disc is ~20mbit stream

(I might be mistaken here)

That's the π's network. I've not measured it

anyhoo - right now I have an ouya for

xbmc

A π is just a cheap bit of kit to do whatever you want with. Nice project on using one for a robot arm in c't a couple of issues back.

yes, that is a great purpose for pi

The "chromecast" (ie. comparable harware) are continually updating the chips. The π has some silly economic stuff that takes it outside the usual market: making stuff in Wales as opposed to Guangdong.

dstufft: woo

inklesspen: it seems to work pretty ok so far, though some things (pyramid_mailer for example) still depends on the thread local mananger, but that jsut means more PRs!

yay

now just to get the pyramid_tm PR accepted and released so I can stop using a git checkout :D

re ?

bah - "git clean -xdf && python3 -m venv . && bin/pip install -e ." and suddenly entrypoints work again

Don't know if it's relevant but I have find . -name "*.pyc" -delete in my config

damn, pyramid's view config is so much more complex than zope's publisher

interesting, that pyramid bug is not related to ABCs. With normal base classes I get the exact same error

Charlie_X: PYTHONDONTWRITEBYTECODE=1 ftw

Had that been a Perl flag it would be $[=1 .

you made this up, right?

PYTHONDONTWRITEBYTECODE is real.

dstufft: I know that but that's just a loaded gun

how is it a loaded gun

postlogic: we learnt yesterday that Perl 6 will let you modify the grammar of the language at runtime.

NO. NO NO NO.

>:-|

do3cc: afraid not

Well technically you can do similar thigns with AST I guess, but it's just not right :-/

dstufft: because it almost always not what you want do to. You want a command to cleanup a virtualenv when switching branches.

Charlie_X: eh, it's almost always exactly what I want to do ;)

my virtual environments don't live in my repo root

Neither do mine (I think)

my commands shouldn't go modifying random things in my virtualenv to clean out pycs

The problem with all environment variables is that you have to know they're there.

it's always there for me, it's in my shell rc file :D

hi, I'm trying to hide links based on the permissions of the target view, how can I safely check that?

I'm very happy for you. Mine is in my .hgrc file

There are lots of times when I'm very glad that there are .pyc files around

But whatever works best for you.

Charlie_X: in what situation are you glad there are .pyc files around? I can't say I've ever been happy for them to be around so I'm curious

So the bytecode doesn't need recompiling all the time.

They are just implementation artefacts that I normally don't care about.

Ok, I guess I just don't value caching the bytecode on my development machine since there's basically no noticeable speed difference to me and it means I never have to get confused about why code isn't doing the right thing and then get all like "oh whoops need to clean out my .pycs"

hmm, problems with old .pycs happens once every 5 years I would guess

I'm not sure i've even beenw riting Python for 5 years and it's happened to me enough to get frustrated enough to find a solution

uhm, does no one know how to check permissions on routes/views? e.g. asking pyramid "can I see route 'foo'?"

dstufft: there are incidences where people have been very happy that the .pycs are around. The performance will be relative to what you're doing. You're obviously very happy with your solution but I think it has pitfalls for the unwary, which is why it's not the default. But enough bikeshedding

Justus: http://docs.pylonsproject.org/projects/pyramid/...

Wow, way wrong version

way to go postlogic

Stupid google

postlogic: that is explicitly not what is required. I do not want to check a specific permission against a known request and resource, I want to know if a route leads to a resource that requires a specific permission and if yes if I have it

Boom

1.4?

Charlie_X: IT IS MONDAY

Justus: http://docs.pylonsproject.org/projects/pyramid/...

I just wanna be done at work ;_;

postlogic: thank you :)

There's that funny word again…

That was 1.4 again

20 minutes and I can leaaaaave

Charlie_X: I don't think that API has changed, though

detention for postlogic

I doubt it has