The group you are posting to is a Usenet group. Messages posted to this group will make your email address visible to anyone on the Internet.
Your reply message has not been sent.
Your post was successful
From: Andrew Stewart <b...@airbladesoftware.com>
Date: Sat, 16 Feb 2008 12:34:03 +0000
Local: Sat, Feb 16 2008 7:34 am
Subject: Re: God's memory leak - a scientific treatment
On 16 Feb 2008, at 00:53, Tom Werner wrote: > Further testing shows that it's the condition.base_name part of the Well done for isolating this; a triumph of empirical legwork. > line that is responsible. Here's the code: > def base_name > Amazingly, this seemingly benign method rings a very familiar bell. Odd indeed. This looks like a bug in Ruby. Would Ruby Talk be the > Another developer at Powerset had identified a piece of code in his > app that looked very much like this and resulted in a HUGE memory leak > as well. The odd but functional fix that we identified was to create a > local scope by declaring a local variable. Here's the diff that I will > try for this test run: > def base_name best place to raise this in the hope of solving the problem at source? Regards, ------- You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
| ||||||||||||||