<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0"><channel><title>Disqus - Latest Comments for dlsspy</title><link xmlns="http://www.w3.org/2005/Atom" rel="http://api.friendfeed.com/2008/03#sup" href="http://disqus.com/sup/all.sup#usercomments-043e29c9" type="application/json"/><link>http://disqus.com/people/dlsspy/</link><description></description><language>en</language><lastBuildDate>Fri, 23 Oct 2009 17:34:50 -0000</lastBuildDate><item><title>Re: ZFS for MacOS X</title><link>http://dustin.github.com/2009/10/23/mac-zfs.html#comment-20885848</link><description>This is based on the old zfs-119 builds from Apple.  It's zpool version 8.  zfs version 2.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Fri, 23 Oct 2009 17:34:50 -0000</pubDate></item><item><title>Re: Memcached Report Card</title><link>http://dustin.github.com/2009/10/22/memcached-reportcard.html#comment-20799704</link><description>Thanks.  :)  I used &lt;a href="http://www.omnigroup.com/applications/omnigraphsketcher/" rel="nofollow"&gt;http://www.omnigroup.com/applications/omnigraph...&lt;/a&gt;  Makes it pretty easy to polish up the data I was bringing together in spreadsheets and on paper and such.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Thu, 22 Oct 2009 15:57:49 -0000</pubDate></item><item><title>Re: EMemcached</title><link>http://dustin.github.com/2009/10/11/ememcached.html#comment-20021656</link><description>That's not so much what it does.  The primary point is to provide a bridge to other services.  It's the storage services that have to deal with  memory fragmentation.&lt;br&gt;&lt;br&gt;If you care about such things, use the other server I work on:  &lt;a href="http://code.google.com/p/memcached/" rel="nofollow"&gt;http://code.google.com/p/memcached/&lt;/a&gt;  :)</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Wed, 14 Oct 2009 02:23:01 -0000</pubDate></item><item><title>Re: EMemcached</title><link>http://dustin.github.com/2009/10/11/ememcached.html#comment-19940653</link><description>Yes.  My code may be valuable to that project, but the goals are different.&lt;br&gt;&lt;br&gt;My goal was not to produce a memcached server, but a memcached protocol parser/dispatcher.&lt;br&gt;&lt;br&gt;The cacherl guys are welcome to add my code to their project for binary protocol support.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Mon, 12 Oct 2009 22:02:20 -0000</pubDate></item><item><title>Re: EMemcached</title><link>http://dustin.github.com/2009/10/11/ememcached.html#comment-19926558</link><description>I know all about it.  :)&lt;br&gt;&lt;br&gt;Usually my process of innovation involves thinking of something really awesome, doing a google search and spending the rest of my day evaluating implementations of things from people who think them out far better than I do.&lt;br&gt;&lt;br&gt;It's a gift and a curse.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Mon, 12 Oct 2009 17:14:50 -0000</pubDate></item><item><title>Re: Tornado on Twisted</title><link>http://dustin.github.com/2009/09/12/tornado.html#comment-17385616</link><description>That's true, it's not needed.  I still have stuff to remove.&lt;br&gt;&lt;br&gt;Someone else has picked up some work on this and may help me move it further forward.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Fri, 25 Sep 2009 13:04:33 -0000</pubDate></item><item><title>Re: Tornado on Twisted</title><link>http://dustin.github.com/2009/09/12/tornado.html#comment-16544014</link><description>I don't have a benchmark.  I expect it to be ballpark, but not as good as it could be.  I introduced some overhead to maintain API compatibility -- that is, if it had been built on twisted, a few things would have been a bit more simple and just kind of run.&lt;br&gt;&lt;br&gt;I expect it to perform better on systems that aren't Linux, but between Tornado's optimizations for Linux and the aforementioned API compatibility wrappers, it's going to be slightly slower.&lt;br&gt;&lt;br&gt;That said, when you start needing to access a ton of external resources (xmpp, http as a client, databases, memcached, etc...) and you get to use all of twisted's async interfaces to these things, it should maintain consistent performance much better.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Sun, 13 Sep 2009 15:26:17 -0000</pubDate></item><item><title>Re: Tornado on Twisted</title><link>http://dustin.github.com/2009/09/12/tornado.html#comment-16522284</link><description>It's a 10pt monaco.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Sat, 12 Sep 2009 21:14:42 -0000</pubDate></item><item><title>Re: The technology behind Tornado, FriendFeed's web server - Bret Taylor's blog</title><link>http://bret.appspot.com/entry/tornado-web-server#comment-16506643</link><description>I've used the twisted enterprise stuff in some apps for non-blocking RDBMS action.  twitterspy supports both the twisted async interface to sqlite and a couchdb interface (I use the latter).  twitterspy as I run it == async couchdb + async memcached + async http streaming + async xmpp interface (and a bit more).&lt;br&gt;&lt;br&gt;I also saw that someone filed an issue against tornado about running async system commands and displaying their output.  If you go about three minutes into this video: &lt;a href="http://www.vimeo.com/5998733" rel="nofollow"&gt;http://www.vimeo.com/5998733&lt;/a&gt; you'll see Matt Ingenthron demonstrating a twisted-based web server I wrote that runs system commands and streams replicates of the output (processed through a python filter) to whomever happens to be looking.&lt;br&gt;&lt;br&gt;This is why Tornado is both interesting and mildly frustrating.  I've got a handful of little frameworks like the above that are doing useful realtime web stuff that could all use an easy to use web framework that doesn't block.  Hopefully by porting tornado to twisted, I can get some more web attention directed at these projects.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Sat, 12 Sep 2009 14:37:22 -0000</pubDate></item><item><title>Re: Tornado on Twisted</title><link>http://dustin.github.com/2009/09/12/tornado.html#comment-16505211</link><description>I'm pretty sure it can't go back upstream.  I've removed lots of their core.  :)  Applications they've written that deeply use their framework probably won't work.&lt;br&gt;&lt;br&gt;Ideally, I'd just take the web framework part, write at least one test for it, and move it towards twisted.  At the very least, we can get some twisted apps using it.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Sat, 12 Sep 2009 14:06:25 -0000</pubDate></item><item><title>Re: The technology behind Tornado, FriendFeed's web server - Bret Taylor's blog</title><link>http://bret.appspot.com/entry/tornado-web-server#comment-16486127</link><description>I've done a first pass implementation of tornado's web framework on top of twisted:&lt;br&gt;&lt;br&gt;&lt;a href="http://github.com/dustin/tornado" rel="nofollow"&gt;http://github.com/dustin/tornado&lt;/a&gt;&lt;br&gt;&lt;br&gt; 10 files changed, 92 insertions(+), 1389 deletions(-)</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Sat, 12 Sep 2009 01:42:41 -0000</pubDate></item><item><title>Re: The technology behind Tornado, FriendFeed's web server - Bret Taylor's blog</title><link>http://bret.appspot.com/entry/tornado-web-server#comment-16318104</link><description>I suppose it's an application-specific thing.  inlineDeferreds is very easy, though.  I wish some of the twisted community were able to address your issues early on.&lt;br&gt;&lt;br&gt;The bigger issue is that there are places where tornado could be used today to do great good if it were built on twisted (things like buildbot really need a web framework).&lt;br&gt;&lt;br&gt;Of course, I don't mean to sound like I don't appreciate the work you guys did and gave away.  I was very excited to see it.  I just can't use it for anything right now.  :(</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Thu, 10 Sep 2009 14:55:25 -0000</pubDate></item><item><title>Re: RockStarProgrammer - Git Tag Does the Wrong Thing by Default</title><link>http://www.rockstarprogrammer.org/post/2008/oct/16/git-tag-does-wrong-thing-default/#comment-15749315</link><description>tag without the -a creates something closer to a branch.  It's a kind of bookmark you can use to name a particular revision somewhat informally.&lt;br&gt;&lt;br&gt;Personally, I don't have much of a use for it.  I use branches when I want branches, and tags when I want tags.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Wed, 02 Sep 2009 00:56:30 -0000</pubDate></item><item><title>Re: Memcached 1.4.0</title><link>http://dustin.github.com/2009/07/16/memcached-1.4.html#comment-12931631</link><description>Glad you like it.  Several of us worked on this and we're all relieved to finally get this behind us.  :)</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Sun, 19 Jul 2009 20:14:00 -0000</pubDate></item><item><title>Re: Memcached 1.4.0</title><link>http://dustin.github.com/2009/07/16/memcached-1.4.html#comment-12900284</link><description>See this thread:  &lt;a href="http://groups.google.com/group/memcached/browse_thread/thread/8c7eea241def70f9" rel="nofollow"&gt;http://groups.google.com/group/memcached/browse...&lt;/a&gt;&lt;br&gt;&lt;br&gt;We're actively trying to get people to help out with Windows support.  There some to be few Windows users out there who are willing to participate.  :/</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Sun, 19 Jul 2009 01:02:12 -0000</pubDate></item><item><title>Re: Making Use of Caps Lock</title><link>http://dustin.github.com/2009/02/09/caps-lock.html#comment-12849265</link><description>Interesting.  I used vi/vim for about ten years, but the key to the left of A has always either been control or in my way.  :)  It actually *was* control on many of the older keyboards I have around the house.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Fri, 17 Jul 2009 18:23:47 -0000</pubDate></item><item><title>Re: Big database problem :  Notifixious down for at least a few hours</title><link>http://blog.notifixio.us/post/82563545#comment-6748982</link><description>Have you considered simpledb?  I'm thinking about giving it a go to simplify deployment of a few things, handle more scale-out, and relieve me from having to deal with sysadmin work around a DB.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Sun, 01 Mar 2009 16:11:45 -0000</pubDate></item><item><title>Re: RockStarProgrammer - What Matters in an Asynchronous Job Queue</title><link>http://www.rockstarprogrammer.org/post/2008/oct/04/what-matters-asynchronous-job-queue/#comment-6449864</link><description>I've not actually used SQS, but from looking in the documentation, it seems that getting a job isn't blocking.  I use beanstalkd in places where I need low-latency, so I either have to poll really fast, or ignore my latency requirements.&lt;br&gt;&lt;br&gt;beanstalkd currently does everything I need and it's quite simple.  One file, only libevent as a dependency, and it's currently using less than half a meg of RAM on my box.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Sat, 21 Feb 2009 00:26:15 -0000</pubDate></item><item><title>Re: Using Git Alternates</title><link>http://dustin.github.com/2008/12/30/git-alternates.html#comment-5706600</link><description>Yes, unpack objects will do it, and yes it'll make a pretty big mess.  I've done that recently with a rails project.  Unpacking rails after setting up as an alternate ended up using something like 600MB of space.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Fri, 30 Jan 2009 14:51:09 -0000</pubDate></item><item><title>Re: Git Timecard</title><link>http://dustin.github.com/2009/01/11/timecard.html#comment-5499229</link><description>Yeah, the google chart API seems to be the hard part in a lot of these attempts to plot stuff.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Fri, 23 Jan 2009 14:49:09 -0000</pubDate></item><item><title>Re: git ready &amp;raquo; count your commits</title><link>http://gitready.com/intermediate/2009/01/22/count-your-commits.html#comment-5490773</link><description>I wrote a tool to display these graphically:&lt;br&gt;&lt;br&gt;&lt;a href="http://dustin.github.com/2009/01/16/visualizing-contributors.html" rel="nofollow"&gt;http://dustin.github.com/2009/01/16/visualizing...&lt;/a&gt;&lt;br&gt;&lt;br&gt;You can pass ranges, etc... to it. I wrote another to generate changelogs for me:&lt;br&gt;&lt;br&gt;&lt;a href="http://dustin.github.com/2009/01/17/changelog.html" rel="nofollow"&gt;http://dustin.github.com/2009/01/17/changelog.html&lt;/a&gt;&lt;br&gt;&lt;br&gt;yields&lt;br&gt;&lt;br&gt;&lt;a href="http://dustin.github.com/java-memcached-client/changelog.html" rel="nofollow"&gt;http://dustin.github.com/java-memcached-client/...&lt;/a&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Fri, 23 Jan 2009 03:43:31 -0000</pubDate></item><item><title>Re: Git Timecard</title><link>http://dustin.github.com/2009/01/11/timecard.html#comment-5134232</link><description>I can understand why this would be desirable, but I'd rather build something new than rewrite this thing.  :)&lt;br&gt;&lt;br&gt;There's  good recipe here for *how* it works, so feel free to port it.  :)</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Thu, 15 Jan 2009 00:09:32 -0000</pubDate></item><item><title>Re: git ready &amp;raquo; interactive adding</title><link>http://gitready.com/intermediate/2009/01/14/interactive-adding.html#comment-5133914</link><description>That's the hard way.&lt;br&gt;&lt;br&gt;git add -p&lt;br&gt;&lt;br&gt;I think that functionality is generally more straightforward when you want to do an interactive add (fairly similar to the style of darcs).</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Wed, 14 Jan 2009 23:45:59 -0000</pubDate></item><item><title>Re: Emacs and Jabber: Happy Together</title><link>http://metajack.im/2009/01/07/emacs-and-jabber-happy-together/#comment-5101710</link><description>Those ugly red bars are from show-trailing-whitespace -- I should figure out how to apply that more selectively.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Tue, 13 Jan 2009 18:47:16 -0000</pubDate></item><item><title>Re: git ready &amp;raquo; fixing broken commit messages</title><link>http://gitready.com/advanced/2009/01/12/fixing-broken-commit-messages.html#comment-5088109</link><description>It's a lot easier to play a patch series in the right order if you just give all the patches to am:&lt;br&gt;&lt;br&gt;git am 00*&lt;br&gt;&lt;br&gt;On big fixes, I do tend to use filter-branch, although that's tedious in its own way.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">dlsspy</dc:creator><pubDate>Tue, 13 Jan 2009 01:26:42 -0000</pubDate></item></channel></rss>