<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	>
<channel>
	<title>Comments on: Upstart 0.5: Job Lifecycle</title>
	<atom:link href="http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/</link>
	<description>Just another WordPress weblog</description>
	<pubDate>Sat, 05 Jul 2008 08:49:08 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.5.1</generator>
		<item>
		<title>By: James Scott jr</title>
		<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/#comment-855</link>
		<dc:creator>James Scott jr</dc:creator>
		<pubDate>Thu, 12 Jun 2008 05:37:38 +0000</pubDate>
		<guid isPermaLink="false">http://www.netsplit.com/?p=144#comment-855</guid>
		<description>Question:
Where can I find documentation and examples for this new world?  Namely, I need and example or explaination of how to write a the ideal daemon for UpStart - should it fork, contact dbus or create a pid file?

Next how do I write a Upstart init.d file to replace a SYSV formatted file I used under Fedora, or a Debian based init.d for Ubuntu?  Again, an example would be nice. I can supply an autotools package with debian&#124;redhat.initd templates, and two versions of the same C template daemon; libc &#38; glib based with only --userid username --version --force (pid create) --help, if that would help. Its on my sourceforge page with gapcmon.  gdaemons-0.1.1.tar.bz2 https://sourceforge.net/project/showfiles.php?group_id=157888&#38;package_id=280067&#38;release_id=606274

I ask because I am interested in what your doing and there is a benefit for me; one init.d script for all linux platforms (WooHoo).  Thus, I want to understand how the two or three daemons I have released will be advantaged by upstart.  I have no problem adding dbus, although I already create a pid file.  Also I always fork once, but that two can be changed.  I'm looking for the book or some best practces that I can code to.  Most of what i have googled on UpStart has been difficult to piece together and code to.</description>
		<content:encoded><![CDATA[<p>Question:<br />
Where can I find documentation and examples for this new world?  Namely, I need and example or explaination of how to write a the ideal daemon for UpStart - should it fork, contact dbus or create a pid file?</p>
<p>Next how do I write a Upstart init.d file to replace a SYSV formatted file I used under Fedora, or a Debian based init.d for Ubuntu?  Again, an example would be nice. I can supply an autotools package with debian|redhat.initd templates, and two versions of the same C template daemon; libc &amp; glib based with only &#8211;userid username &#8211;version &#8211;force (pid create) &#8211;help, if that would help. Its on my sourceforge page with gapcmon.  gdaemons-0.1.1.tar.bz2 <a href="https://sourceforge.net/project/showfiles.php?group_id=157888&amp;package_id=280067&amp;release_id=606274" rel="nofollow">https://sourceforge.net/project/showfiles.php?group_id=157888&amp;package_id=280067&amp;release_id=606274</a></p>
<p>I ask because I am interested in what your doing and there is a benefit for me; one init.d script for all linux platforms (WooHoo).  Thus, I want to understand how the two or three daemons I have released will be advantaged by upstart.  I have no problem adding dbus, although I already create a pid file.  Also I always fork once, but that two can be changed.  I&#8217;m looking for the book or some best practces that I can code to.  Most of what i have googled on UpStart has been difficult to piece together and code to.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Scott James Remnant &#187; Blog Archive &#187; Upstart 0.5: Job Environment</title>
		<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/#comment-798</link>
		<dc:creator>Scott James Remnant &#187; Blog Archive &#187; Upstart 0.5: Job Environment</dc:creator>
		<pubDate>Wed, 16 Apr 2008 14:07:22 +0000</pubDate>
		<guid isPermaLink="false">http://www.netsplit.com/?p=144#comment-798</guid>
		<description>[...] Home      &#171; Upstart 0.5: Job Lifecycle [...]</description>
		<content:encoded><![CDATA[<p>[...] Home      &laquo; Upstart 0.5: Job Lifecycle [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Vlad</title>
		<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/#comment-797</link>
		<dc:creator>Vlad</dc:creator>
		<pubDate>Mon, 14 Apr 2008 22:53:02 +0000</pubDate>
		<guid isPermaLink="false">http://www.netsplit.com/?p=144#comment-797</guid>
		<description>&#62;&#62; up to the PlanetKDE admins I guess

If you've already contacted Chris Lee, that's good. He might take a while to update PlanetKDE -- I know it took about 2 weeks for my name to appear on the list.</description>
		<content:encoded><![CDATA[<p>&gt;&gt; up to the PlanetKDE admins I guess</p>
<p>If you&#8217;ve already contacted Chris Lee, that&#8217;s good. He might take a while to update PlanetKDE &#8212; I know it took about 2 weeks for my name to appear on the list.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jeff Bailey</title>
		<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/#comment-796</link>
		<dc:creator>Jeff Bailey</dc:creator>
		<pubDate>Mon, 14 Apr 2008 17:32:25 +0000</pubDate>
		<guid isPermaLink="false">http://www.netsplit.com/?p=144#comment-796</guid>
		<description>&#62; @jbailey: arguments to “script” are in the TODO, I’d planned to just default to shell if no arguments were given; is 
&#62; there a compelling reason to have a “shell” instead?

The biggest argument is clarity.  You've already got 'exec' to differentiate from 'script'.  But the popular scripting language in people's minds shifts from year to year.  A few years ago it would've been Perl.  These days it's Python.

Another weaker argument for having aliases is the possibility of defining them in a config file.  That way if someone wants "shell" to mean "/bin/posh" on their system, they can change it in one place instead of sed'ing all of the config files.</description>
		<content:encoded><![CDATA[<p>&gt; @jbailey: arguments to “script” are in the TODO, I’d planned to just default to shell if no arguments were given; is<br />
&gt; there a compelling reason to have a “shell” instead?</p>
<p>The biggest argument is clarity.  You&#8217;ve already got &#8216;exec&#8217; to differentiate from &#8217;script&#8217;.  But the popular scripting language in people&#8217;s minds shifts from year to year.  A few years ago it would&#8217;ve been Perl.  These days it&#8217;s Python.</p>
<p>Another weaker argument for having aliases is the possibility of defining them in a config file.  That way if someone wants &#8220;shell&#8221; to mean &#8220;/bin/posh&#8221; on their system, they can change it in one place instead of sed&#8217;ing all of the config files.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Scott James Remnant</title>
		<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/#comment-795</link>
		<dc:creator>Scott James Remnant</dc:creator>
		<pubDate>Mon, 14 Apr 2008 17:24:42 +0000</pubDate>
		<guid isPermaLink="false">http://www.netsplit.com/?p=144#comment-795</guid>
		<description>@jbailey: arguments to "script" are in the TODO, I'd planned to just default to shell if no arguments were given; is there a compelling reason to have a "shell" instead?

@tomasz, @jef: the entire point of the new version is to fix the problems that prevented us from taking full advantage of it.  Fedora won't have much luck without it either, and it's in their roadmap as well

@liquidat: yes, we're talking quite a lot

@vlad: up to the PlanetKDE admins I guess</description>
		<content:encoded><![CDATA[<p>@jbailey: arguments to &#8220;script&#8221; are in the TODO, I&#8217;d planned to just default to shell if no arguments were given; is there a compelling reason to have a &#8220;shell&#8221; instead?</p>
<p>@tomasz, @jef: the entire point of the new version is to fix the problems that prevented us from taking full advantage of it.  Fedora won&#8217;t have much luck without it either, and it&#8217;s in their roadmap as well</p>
<p>@liquidat: yes, we&#8217;re talking quite a lot</p>
<p>@vlad: up to the PlanetKDE admins I guess</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Vlad</title>
		<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/#comment-794</link>
		<dc:creator>Vlad</dc:creator>
		<pubDate>Sun, 13 Apr 2008 20:56:39 +0000</pubDate>
		<guid isPermaLink="false">http://www.netsplit.com/?p=144#comment-794</guid>
		<description>Why don't you syndicate your blog to PlanetKDE? A lot of KDE devs would be interested in your work, as it will affect them in the future.</description>
		<content:encoded><![CDATA[<p>Why don&#8217;t you syndicate your blog to PlanetKDE? A lot of KDE devs would be interested in your work, as it will affect them in the future.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: liquidat</title>
		<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/#comment-793</link>
		<dc:creator>liquidat</dc:creator>
		<pubDate>Sun, 13 Apr 2008 12:16:13 +0000</pubDate>
		<guid isPermaLink="false">http://www.netsplit.com/?p=144#comment-793</guid>
		<description>Speaking about upstart, is there any cooperation going on with the Fedora people? They are about to &lt;a href="http://fedoraproject.org/wiki/Features/Upstart" rel="nofollow"&gt;include upstart with Fedora 9&lt;/a&gt;, although in version 0.3.9.</description>
		<content:encoded><![CDATA[<p>Speaking about upstart, is there any cooperation going on with the Fedora people? They are about to <a href="http://fedoraproject.org/wiki/Features/Upstart" rel="nofollow">include upstart with Fedora 9</a>, although in version 0.3.9.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: James</title>
		<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/#comment-790</link>
		<dc:creator>James</dc:creator>
		<pubDate>Sun, 13 Apr 2008 09:09:16 +0000</pubDate>
		<guid isPermaLink="false">http://www.netsplit.com/?p=144#comment-790</guid>
		<description>Wow, {pre,post}-{start,stop}, it's like Debian package scripts.</description>
		<content:encoded><![CDATA[<p>Wow, {pre,post}-{start,stop}, it&#8217;s like Debian package scripts.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: jef</title>
		<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/#comment-783</link>
		<dc:creator>jef</dc:creator>
		<pubDate>Sun, 13 Apr 2008 01:54:13 +0000</pubDate>
		<guid isPermaLink="false">http://www.netsplit.com/?p=144#comment-783</guid>
		<description>Tomasz,

Fedora is actually planning on really taking advantage of upstart unlike Ubuntu in their next release.

http://fedoraproject.org/wiki/Features/30SecondStartup

They have already posted some benchmarks and progressing fast.</description>
		<content:encoded><![CDATA[<p>Tomasz,</p>
<p>Fedora is actually planning on really taking advantage of upstart unlike Ubuntu in their next release.</p>
<p><a href="http://fedoraproject.org/wiki/Features/30SecondStartup" rel="nofollow">http://fedoraproject.org/wiki/Features/30SecondStartup</a></p>
<p>They have already posted some benchmarks and progressing fast.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: ulrik</title>
		<link>http://www.netsplit.com/2008/04/12/upstart-05-job-lifecycle/#comment-782</link>
		<dc:creator>ulrik</dc:creator>
		<pubDate>Sun, 13 Apr 2008 01:03:17 +0000</pubDate>
		<guid isPermaLink="false">http://www.netsplit.com/?p=144#comment-782</guid>
		<description>Exciting that Upstart 0.5 is really coming along! I am hoping for great times for more distributions than just Ubuntu. Let's finally get rid of sysvrc for good.. Thanks for all your work so far SJR!</description>
		<content:encoded><![CDATA[<p>Exciting that Upstart 0.5 is really coming along! I am hoping for great times for more distributions than just Ubuntu. Let&#8217;s finally get rid of sysvrc for good.. Thanks for all your work so far SJR!</p>
]]></content:encoded>
	</item>
</channel>
</rss>
