| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
Comments:
<0> hey nola
<1> merlyn: hey! I might have a perl job.. maybe ...
<0> cool
<2> argh! competition!
<0> I spent yesterday afternoon refactoring a webserver that used POE::Component::Server::HTTP to use POE::Component::Server::SimpleHTTP
<1> Botje: you looking? what area? ... I'm in the beautiful city of chicago
<0> I then used my unit tests, and uncovered a bug in another part of my code. :)
<3> merlyn: what would you say is the best unit testing for a n00b to start out testing with?
<2> Nola: i'm in the beautiful (well, it's raining now) country of belgium :P
<1> merlyn: horray! ... I told this new gig about Test :: More and TAP ... they don't test now
<2> f00li5h: Test::Simple
<3> Botje: hrm, so that one should be the best one to play with,
<2> f00li5h: but read Test::Tutorial.
<3> Botje: did, it didn't make much sence to me.
<3> it may be the paridigm shift though
<0> I refactored so I could use SSL (which SimpleHTTP supports), and also so I could subcl***, because SimpleHTTP uses real states, not odd callbacks, to handle the URLs
<1> f00li5h: did you follow along and make some examples as you went?
<0> which means I can declare a derived cl*** from the object I just created, and get all the URLs handled *and* more
<3> Nola: well, a little, but i don't know where to put the tests, you see?
<4> Hi merlyn
<0> f00li5h - they generally go in t/*.t
<4> Ho f00li5h
<3> merlyn: do go on
<0> so you have bin/* your scripts, and lib/* your modules
<1> f00li5h: you just pu them in their own file .. where merlyn says :)
<3> rindolf: ho there!
<0> and t/*.t tests the stuff in lib/*, and to a lesser extent, the stuff in bin/*
<3> Nola: so grouped by?
<1> look at an existing module
<0> the trick is making yoru lib/* so that it *can* be tested.
<1> look in the t directory
<3> Nola: i don't have one <_<
<3> merlyn: yes!
<0> but that also means you're thinking about how isolated a "unit" can be, which is actually a *good* thing
<3> tell me more of this `can be tested'
<4> f00li5h, did you start with module-starter or whatever?
<3> rindolf: no, i started with mkdir
<4> f00li5h, heh.
<0> for example, I have a cl*** that used another cl***. to test the first cl***, I really didn't want it using the second cl***, so I added a hook so that I could "p*** in" the "second cl***" to use.
<4> f00li5h, are you trying to write a CPAN module?
<0> now in testing, I p*** in a dummy second cl***.
<4> f00li5h, or just a regular Perl module distribution?
<3> rindolf: heck no. nobody wants this crap
<0> which is really a little test dummy that reports what methods the first cl*** is calling.
<3> i just want to be able to look at green lights and go `^_^ its working'
<0> so that's an example of "designing to be tested"
<4> Nola, are you new here?
<0> the advantage is that my first cl*** is now reusable in a different system that might need a different "other cl***" to be useful.
<0> hardwired "other" cl***es are generally trouble.
<3> merlyn: like reading a config file, so i know what to expect a configurable cl*** will do?
<0> generally, it's in the parameters you p*** to ->new
<0> I have a "manager" cl*** that creates "worker" object
<0> so I p*** in my TestWorker cl*** name instead
<3> merlyn: and then sits there taking the credit?
<0> but by default, it would use the standard Worker cl***
<5> good morning.
<4> Hi Teratogen
<3> good day Teratogen
<0> I define TestWorker directly in the .t file, and it shares closure variables with things my test code can see
<0> so my .t code calls Manager which creates TestWorker which triggers vars which the .t code can monitor
<0> then I can make sure that Manager would do the right things to Worker
<3> merlyn: that's pretty ***y
<0> It's actually pretty standard... I can't take credit for it
<0> in fact, there's Test::MockObject to do this generically
<3> merlyn: well, i guess i'll just say that you're a lot better at this than i am ...
<0> Well - this is why $client went from simply bringing me in for a few weeks, to extending the contract to five months by reporting to upper management that I was "critical to the completion of the project"
<3> merlyn: ie you were the only one there that actually knew perl... the others were just mashing the keys, and faking it?
<0> No - everyone knows Perl more or less.
<0> but nobody knows it like me.
<6> Not even Larry!
<6> or Damian!
<0> <melbrooks>It's Good To Be The King</melbrooks>
<0> well, Larry and Damian know it "differently" :)
<6> heh
<0> Larry and I have similar programming histories, and I've *used* Perl actually a lot more than he has. :)
<0> I haven't been burdened by constantly looking at the insides. :)
<0> "are you gonna drive, or just spend all your time under the hood?"
<3> aah, i see
<7> I wish the code controlling the replacement part of s/// wasn't so confusing to me.
<7> I'm slipping in my C-fu.
<0> Ilya-code. :)
<0> obfu-russian
<7> merlyn: all I wanted to do was patch s/// to accept /r as a modifier, whereby the modified string is returned, and the original variable remains untouched
<0> Eeek.
<7> but that is proving way harder than I thought
<8> oh, I tried that once
<0> sounds great, but I bet it's a nightmare
<8> it seemed to build
<7> mauke: you retained your sanity!
<0> you can't lose what you don't have in the first place.
<0> the fact that he *tried* already demonstrates that.
<8> no brain, no pain
<3> merlyn: credit card.
<7> actually, I've been looking at it the wrong way.
<7> I should copy the scalar first, replace TARG with that copy, and then return TARG.
<7> I'll try it again this afternoon...
<0> oddly, I thought /r would be ***eger. :)
<7> merlyn: no, that's a whole nother can of worms.
<7> I've since changed my mind: you want a reversed regex, you pay me ;)
<7> or DIY
<8> add it to tr/// too, while you're at it
<0> and chomp()!
<8> chomp/$foo/r
<0> people often do $a = chomp($a)
<0> oops. :)
<8> $foo =~ chomp//r
<6> I wish chomp worked like that :|
<3> merlyn: i find that i don't do that, out of habbit. php's chop doesn't work in place. and it takes ages to hunt down
<0> just avoid php :)
<0> Pretty Home Page. :)
<6> s/\'s chop//; s/ in place//;
<3> merlyn: they've put my job up for `internal advertising, followed by public [me] advertising' and perhaps i'll find a perl job before i re-apply.
<9> so they are trying to get rid of you?
<3> they say no, but they're replacing contractors with permi's
<3> so more or less... i looks that wya
<4> merlyn: I'm writing PHP now.
<3> rindolf: i feel for you.
<7> you see, that's why I like about ruby...
<7> y = chomp(x);
<7> chomp!(x);
<3> japhy: that it's not php?
<0> You're probably not writing PHP like Rasmus writes PHP. :)
<4> f00li5h: no, that you can add ! to do stuff in place.
<7> err, methods, though
<7> y = x.chomp;
<7> x.chomp!;
<3> rindolf: how does that work?
<4> f00li5h: in Ruby?
<3> rindolf: f(x,y):z
<7> f00li5h: you have to IMPLEMENT the in-place thing.
<3> japhy: and say which arg it inplaces?
<7> 'chomp' and 'chomp!' are two different methods, I would expect. chomp! merely says x = x.chomp
<7> but I could be totally wrong, too
<0> we could *all* be totally wrong!
<8> use subs 'chomp'; sub chomp { unless (defined wantarray) { CORE::chomp(@_); return } CORE::chomp(my @x = @_); wantarray ? @_ : $_[0] }
<5> *chomp*
<7> it might be that THING.xyz! defaults to being: THING = THING.xyz
<10> you guys have the knowledge
<7> unless .xyz! is defined otherwise
<0> maybe *python* actually makes sense, if you've used it long enough
<4> /~ If you've got the knowledge honey, I've got the ties... ~/
<10> i hope someday i can progress to have the knowledge also
<7> nah, I don't think python makes enough sense.
<7> python can't make sense without importing it first ;)
<0> or TCL, which I call, Bracket Hell
<0> all the syntax of Lisp with very little of the power.
<7> someone can feel free to quote me on that, by the way.
<0> apparently - people quote me whether I give permission or not. :)
<0> damn pedestals. :)
Return to
#perl
or
Go to some related
logs:
reinit raid mdadm
#python
xgl conkey
Eulex log
what do numbers mean linux syslog
trouble wlanassistant
fglrx dapper 9200 printfps
ubuntu how does gdm authenticate?
ethereal sudoer
riotgeek
|
|