another code sample

[juan_gandhi]

  def parameters(href): Map[String, String] = href split("\\?",2) toList match {
    case url::params::Nil => params.split("&") map (_ split "=" toList) collect {
      case k::v::Nil => k->v
    } toMap

    case otherwise => Map.empty  
  }

Or?...

Comments

[jdevelop.livejournal.com]

wouldn't catch the case when param1=value1&param1=value2

also & will break everything

I'd suggest to use URL from Apache HttpClient.

[juan-gandhi.livejournal.com]

Why would not it catch the case?
And I despise the whole Apache HttpClient anyway.

[jdevelop.livejournal.com]

because you have multiple values for same parameter, which is legal - and you'r code will get only last one.

as for & - this is a valid code block for encoding the ampersand in the parameter name or value.

[juan-gandhi.livejournal.com]

Right. Well, have to think about it.

[http://users.livejournal.com/_xacid_/]

use List[(String, String)] as result

:)

[juan-gandhi.livejournal.com]

Well, this is more serious, of course. :)

[mikkim08.livejournal.com]

I'd suggest to use

Только без "to". Можно говорить "suggest (you) use" или "suggest using", а "suggest to use" -- нельзя.

Я это говорю только потому, что когда исправляют, лучше запоминается.

[juan-gandhi.livejournal.com]

http://www.englishteachermelanie.com/grammar-how-to-use-the-verb-suggest/

[yussouf.livejournal.com]

what does these toLists toMaps cost?

[juan-gandhi.livejournal.com]

I would not say I care. Converting max 1k of data from one form to another...

[yussouf.livejournal.com]

that's apparently a wicked thinking to me

which you functionalistz are (very) usually tend to fall into
everything is to be sacrificed to the that notorious "expressiveness" again

PS even if you're personally very aware of what exactly is going on in this particular case and are pretty 100% sure that the data structures won't exceed the reasonable size of that 1K (and the number of that stuff processed in parallel would also be reasonable) - it is STILL a wicked thinking

[juan-gandhi.livejournal.com]

Ok, but I don't care! I have other things on mind.

[badula.livejournal.com]

At some point in our talks I asked Schwarze­negger how much time he had spent, as governor, grappling with the on-the-ground local implications of the big state crisis. The question pretty clearly bored him. “I’m not into the local stuff,” he’d said. “I was born for the world.” (http://www.vanityfair.com/business/features/2011/11/michael-lewis-201111)

[http://users.livejournal.com/_xacid_/]

I know several imperativists who don't care even more

[yussouf.livejournal.com]

I know several dozens of 'em - so what the heck does it prove? :-)

[http://users.livejournal.com/_xacid_/]

nothing, of course

[badula.livejournal.com]

the costs are mainly infrastructural, because all data is immutable, the only added cost is the one of calling through the shitload of intermediate iterators
not sure how lazy the "split" thing is in scala. the sample goes through all iterators sequentially

[yussouf.livejournal.com]

the split produces just mere Java mutable Array of strings afair

so, that happened to be the full copy of data, no reuse

[badula.livejournal.com]

The complexity is still linear. That is in most cases what matters. In some other cases instead of 5000 servers one may have to have 10000 suckers. At which point the additional operational expenses should be but trivial.

[http://users.livejournal.com/_xacid_/]

why not parsers combinators? :)

[juan-gandhi.livejournal.com]

Sure it would work; but this is a regular language anyway (why not a regex?)

[http://users.livejournal.com/_xacid_/]

indeed :) moreover, they both work great together