Beefy Boxes and Bandwidth Generously Provided by pair Networks
There's more than one way to do things

Re^2: Unit testing OS rich code

by Voronich (Hermit)
on Oct 12, 2011 at 15:54 UTC ( #931046=note: print w/replies, xml ) Need Help??

in reply to Re: Unit testing OS rich code
in thread Unit testing OS rich code

Eh. I really don't like having filesystem state for tests. I think I'm just avoiding the inevitable here. I've got the same issues coming fast on the horizon for database interactions as well.

Replies are listed 'Best First'.
Re^3: Unit testing OS rich code
by chromatic (Archbishop) on Oct 12, 2011 at 20:20 UTC
    I really don't like having filesystem state for tests.

    If your code manipulates the filesystem, how will you know if it works unless you test that it manipulates the filesystem?

    Improve your skills with Modern Perl: the free book.

      At some point you have to trust that the primitives are doing their job. Otherwise "full code coverage" would always include unit tests that cover perl itself. So keeping an onion-skin layer on top allows me to preserve testability.

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://931046]
[ambrus]: Corion: I think I parsed a HTTP header from a string with LWP once. You can definitely use that to create a HTTP message too. The problme is
[ambrus]: that if you do that, you'd have to find where each HTTP response ends, which is nontrivial if you want persistent connections (essential for performance if you have small requests).
[Corion]: ambrus: Yes, ideally an API that I feed the incoming data piece by piece and that I can ask "is that response done" and "what should I do next" and "please construct the appropriate redirect for me"
[Corion]: ambrus: Yes, ideally the module would do all that nasty stuff for me and give me a way to ask it what the current situation is
[ambrus]: Corion: you could also consider using some wrapper over the multi interface of curl, I think Net::Curl might be a good one, since implementing enough of what it expects from the event loop might be easier than a full AnyEvent interface.
[ambrus]: Corion: you could also consider using IO::Async and its POE driver and some HTTP module for it, but I don't know if that would solve your problems with AnyEvent+POE
[ambrus]: Corion: wait, you didn't say POE. You said Prima, let me look up what that is.
[ambrus]: Corion: have you considered just writing an AnyEvent integration for that thing?
[ambrus]: Or perhaps pushing schmorp to write one?
[ambrus]: Also, searching for an existing one on CPAN obviously

How do I use this? | Other CB clients
Other Users?
Others avoiding work at the Monastery: (11)
As of 2016-12-07 16:07 GMT
Find Nodes?
    Voting Booth?
    On a regular basis, I'm most likely to spy upon:

    Results (130 votes). Check out past polls.