#66 Improve performance of backtrace filtering - mocha

Type	To find
responsible:me	tickets assigned to you
tagged:"@high"	tickets tagged @high
milestone:next	tickets in the upcoming milestone
state:invalid	tickets with the state invalid
created:"last week"	tickets created last week
sort:number, importance, updated	tickets sorted by #, importance or updated
Combine keywords for powerful searching.
Use advanced searching »

#66 ✓resolved

Improve performance of backtrace filtering

Reported by James Mead | 2010-03-17 19:11:23 UTC

Nikolai Weibull posted the following to the Mocha mailing list :-

Mocha::BacktraceFilter creates a new Regexp object for every line of every backtrace it filters. It also calls File.expand_path for each such line.

The Regexp object creation can easily be removed by creating it once in #initialize. File.expand_path might be more difficult to get rid of, but File.expand_path is slow, at least on Windows. Getting rid of all those calls to it would thus save a lot of time. Rewriting BacktraceFilter to that below makes running my 108 tests with Expectations run in less than a third of the time:

  def initialize(lib_directory = LIB_DIRECTORY)
    @lib_directory = Regexp.new('^' + Regexp.escape(lib_directory))
  end

  def filtered(backtrace)
    backtrace.reject { |location| @lib_directory.match(location) }
  end

I realize that this doesn’t filter backtraces correctly if Mocha’s paths are relative. I was, however, hoping that someone with more knowledge in this area would be able to resolve this. Perhaps setting LIB_DIRECTORY = File.dirname(File.dirname(__FILE__)) + File::SEPARATOR is a good way of doing it? Then both @lib_directory and location will both be relative or absolute, depending on how Mocha was required.

Comments and changes to this ticket

You flagged this item as spam.
Nikolai Weibull 2010-03-22 11:03:39 UTC
Here’s a patch that should work.
- backtrace.patch 798 Bytes
James Mead 2010-11-26 11:41:31 UTC
- State changed from “new” to “open”
@nikolai: Apologies for the (very) slow response. I've just been looking at this and thinking about it a bit. I've made the change so that the Regexp instance is now only created once in the BacktraceFilter#initialize method.

However, I've also realized that (a) this probably won't improve performance because a new BacktraceFilter instance is created for every expectation error; and (b) in any case, any performance improvement will only be apparent for people who have a lot of test failing with expectation errors. This latter case seems like a bit of an edge case, so I'm not going to make the other changes.

I'm closing this ticket. Thanks, James.
James Mead 2010-11-26 11:41:40 UTC
- State changed from “open” to “resolved”

Please Sign in or create a free account to add a new ticket.

With your very own profile, you can contribute to projects, track your activity, watch tickets, receive and update tickets through your email and much more.