<?xml version="1.0" encoding="UTF-8"?><!-- generator="wordpress/2.0.7" -->
<rss version="2.0" 
	xmlns:content="http://purl.org/rss/1.0/modules/content/">
<channel>
	<title>Comments on: Subversion commit access</title>
	<link>http://jc.ngo.org.uk/blog/2007/01/18/subversion-commit-access/</link>
	<description>It'll all be over in 60 days</description>
	<pubDate>Mon, 08 Sep 2008 01:48:24 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.0.7</generator>

	<item>
		<title>by: nik</title>
		<link>http://jc.ngo.org.uk/blog/2007/01/18/subversion-commit-access/#comment-3311</link>
		<pubDate>Thu, 18 Jan 2007 23:42:19 +0000</pubDate>
		<guid>http://jc.ngo.org.uk/blog/2007/01/18/subversion-commit-access/#comment-3311</guid>
					<description>Oh, one more thing -- I'm interested in apps, libraries, utilities and the like that people are creating using the Perl bindings.  Can you tell me a bit about what you're using them for?</description>
		<content:encoded><![CDATA[<p>Oh, one more thing &#8212; I&#8217;m interested in apps, libraries, utilities and the like that people are creating using the Perl bindings.  Can you tell me a bit about what you&#8217;re using them for?
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: nik</title>
		<link>http://jc.ngo.org.uk/blog/2007/01/18/subversion-commit-access/#comment-3309</link>
		<pubDate>Thu, 18 Jan 2007 23:38:28 +0000</pubDate>
		<guid>http://jc.ngo.org.uk/blog/2007/01/18/subversion-commit-access/#comment-3309</guid>
					<description>You don't need to explicitly use pools at the moment.  The pool param to the calls is (AFAIK) always optional.

You &lt;b&gt;can&lt;/b&gt; use pools if you want to.  For example, if you're iterating over a lot of objects the "standard" approach is to allocate a subpool before you start the iteration, and clear it at the start of each iteration, so that memory usage doesn't grow, and operations tend to take O(1) memory rather than O(n).

You don't need to do that now if you don't want to though.

I agree that having to worry about this at all isn't terribly Perlish (although, I guess, not much more so than having to worry about references, cyclic references, weakening them, and so on).

The sort of stuff I'm immediately looking at is getting the API to support named parameters for calls as well as the positional parameters, and, where possible, to supply default values for those parameters so that your Perl code can be simpler and more readable.  I'm working on a patch (which will be committed) for that at the moment.</description>
		<content:encoded><![CDATA[<p>You don&#8217;t need to explicitly use pools at the moment.  The pool param to the calls is (AFAIK) always optional.</p>
<p>You <b>can</b> use pools if you want to.  For example, if you&#8217;re iterating over a lot of objects the &#8220;standard&#8221; approach is to allocate a subpool before you start the iteration, and clear it at the start of each iteration, so that memory usage doesn&#8217;t grow, and operations tend to take O(1) memory rather than O(n).</p>
<p>You don&#8217;t need to do that now if you don&#8217;t want to though.</p>
<p>I agree that having to worry about this at all isn&#8217;t terribly Perlish (although, I guess, not much more so than having to worry about references, cyclic references, weakening them, and so on).</p>
<p>The sort of stuff I&#8217;m immediately looking at is getting the API to support named parameters for calls as well as the positional parameters, and, where possible, to supply default values for those parameters so that your Perl code can be simpler and more readable.  I&#8217;m working on a patch (which will be committed) for that at the moment.
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: eric h</title>
		<link>http://jc.ngo.org.uk/blog/2007/01/18/subversion-commit-access/#comment-3306</link>
		<pubDate>Thu, 18 Jan 2007 21:40:14 +0000</pubDate>
		<guid>http://jc.ngo.org.uk/blog/2007/01/18/subversion-commit-access/#comment-3306</guid>
					<description>Does this mean that users of the perl client libraries won't need to explicitly allocate and de-allocate apr pools?  That'd be nice and perlish.</description>
		<content:encoded><![CDATA[<p>Does this mean that users of the perl client libraries won&#8217;t need to explicitly allocate and de-allocate apr pools?  That&#8217;d be nice and perlish.
</p>
]]></content:encoded>
				</item>
</channel>
</rss>
