<?xml version="1.0" encoding="ISO-8859-1"?><!-- generator="wordpress/2.0" -->
<rss version="2.0" 
	xmlns:content="http://purl.org/rss/1.0/modules/content/">
<channel>
	<title>Comments on: feh work</title>
	<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/</link>
	<description>Tom Gilbert's LinuxBrit. Software releases and personal blog.</description>
	<pubDate>Sun, 07 Sep 2008 21:01:49 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.0</generator>

	<item>
		<title>by: giblet</title>
		<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-288</link>
		<pubDate>Mon, 06 Sep 2004 15:49:32 +0000</pubDate>
		<guid>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-288</guid>
					<description>Hehe, okay I'm with you there. There was one comment I almost added to my post - which was &quot;I'd be a lot happier if the repo wasn't stored in a berkeley DB, or maybe just the metadata..&quot;. But then I figured I was being fussy, and I know how much cool stuff like transactions and atomicity you get for free with DB, so I left it out.

That kind of recovery would definitely be a bitch though. I think my backup policy for subversion is going to be twofold. A hot backup (cp -a on the tree) and an endian-agnostic database dump :)</description>
		<content:encoded><![CDATA[<p>Hehe, okay I&#8217;m with you there. There was one comment I almost added to my post - which was &#8220;I&#8217;d be a lot happier if the repo wasn&#8217;t stored in a berkeley DB, or maybe just the metadata..&#8221;. But then I figured I was being fussy, and I know how much cool stuff like transactions and atomicity you get for free with DB, so I left it out.</p>
<p>That kind of recovery would definitely be a bitch though. I think my backup policy for subversion is going to be twofold. A hot backup (cp -a on the tree) and an endian-agnostic database dump :)
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: richlowe</title>
		<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-287</link>
		<pubDate>Mon, 06 Sep 2004 15:27:55 +0000</pubDate>
		<guid>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-287</guid>
					<description>You are entirely correct on all counts, I was going for brevity :)

The machine that the subversion server was on died.
That machine was MSB, I had to move to an LSB machine.

but to run the dump, I needed an MSB host, no other way around it.

Thus I spent 24 hours trying to remember who I knew with an MSB machine, getting an account on there, and waiting (a considerable amount of time) for subversion, apr, neon, and berkelydb to run on it.  rsyncing my entire repo to the host, running the dump, and rsyncing the data back.

My point is that I tend to think that using an endian agnostic format for the repo would have been a damn good idea, and know first hand just how much of a pain in the ass it can be in disaster recovery if you have to move between ends for reasons beyond your control :)</description>
		<content:encoded><![CDATA[<p>You are entirely correct on all counts, I was going for brevity :)</p>
<p>The machine that the subversion server was on died.<br />
That machine was MSB, I had to move to an LSB machine.</p>
<p>but to run the dump, I needed an MSB host, no other way around it.</p>
<p>Thus I spent 24 hours trying to remember who I knew with an MSB machine, getting an account on there, and waiting (a considerable amount of time) for subversion, apr, neon, and berkelydb to run on it.  rsyncing my entire repo to the host, running the dump, and rsyncing the data back.</p>
<p>My point is that I tend to think that using an endian agnostic format for the repo would have been a damn good idea, and know first hand just how much of a pain in the ass it can be in disaster recovery if you have to move between ends for reasons beyond your control :)
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: giblet</title>
		<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-279</link>
		<pubDate>Sun, 05 Sep 2004 10:30:50 +0000</pubDate>
		<guid>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-279</guid>
					<description>Hrm, according to the docs there's an svnadmin dump command which spits out the repository in a neutral format - I gather you're supposed to use that if you're going to move between endiannesses.. Sounds like you had a standard Berkeley DB portability problem to me, rather than anything specific to subversion..</description>
		<content:encoded><![CDATA[<p>Hrm, according to the docs there&#8217;s an svnadmin dump command which spits out the repository in a neutral format - I gather you&#8217;re supposed to use that if you&#8217;re going to move between endiannesses.. Sounds like you had a standard Berkeley DB portability problem to me, rather than anything specific to subversion..
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: richlowe</title>
		<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-277</link>
		<pubDate>Sun, 05 Sep 2004 07:11:09 +0000</pubDate>
		<guid>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-277</guid>
					<description>Ooooooh, I just realized if feh CVS is there now I can commit again!

Not sure &lt;b&gt;what&lt;/b&gt; I'd commit, but still, it's theoretically possible!

Maybe I'll spend my time enhancing pheh back to the level of feh... it's a cooler app anyway ;-)</description>
		<content:encoded><![CDATA[<p>Ooooooh, I just realized if feh <acronym title="Concurrent Versions System">CVS</acronym> is there now I can commit again!</p>
<p>Not sure <b>what</b> I&#8217;d commit, but still, it&#8217;s theoretically possible!</p>
<p>Maybe I&#8217;ll spend my time enhancing pheh back to the level of feh&#8230; it&#8217;s a cooler app anyway ;-)
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: richlowe</title>
		<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-276</link>
		<pubDate>Sun, 05 Sep 2004 07:09:04 +0000</pubDate>
		<guid>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-276</guid>
					<description>I still have an intense hatred of subversion related to spending upwards of a solid 24 hour block fixing a repo after it moved between machines of diferent endiannesseses.

From what I read the FSFS stuff fixes that, at the expense of using more space, and not being as reliable.

What a fun choice.

Why oh why can't they finish OpenCM already :(
That looked nifty, then they rewrote it, then it looked nifty again, then they rewrote it.

Wait, I've seen this pattern before...</description>
		<content:encoded><![CDATA[<p>I still have an intense hatred of subversion related to spending upwards of a solid 24 hour block fixing a repo after it moved between machines of diferent endiannesseses.</p>
<p>From what I read the FSFS stuff fixes that, at the expense of using more space, and not being as reliable.</p>
<p>What a fun choice.</p>
<p>Why oh why can&#8217;t they finish OpenCM already :(<br />
That looked nifty, then they rewrote it, then it looked nifty again, then they rewrote it.</p>
<p>Wait, I&#8217;ve seen this pattern before&#8230;
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: pabs</title>
		<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-262</link>
		<pubDate>Sat, 04 Sep 2004 02:35:29 +0000</pubDate>
		<guid>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-262</guid>
					<description>Yup.  The article on your link bar mentioned something about ViewCVS support.  Wel, the comments did anyway</description>
		<content:encoded><![CDATA[<p>Yup.  The article on your link bar mentioned something about ViewCVS support.  Wel, the comments did anyway
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: giblet</title>
		<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-256</link>
		<pubDate>Fri, 03 Sep 2004 18:01:15 +0000</pubDate>
		<guid>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-256</guid>
					<description>Oh yeah, actually I'm told viewcvs supports subversion too, so maybe we can set that up now :)</description>
		<content:encoded><![CDATA[<p>Oh yeah, actually I&#8217;m told viewcvs supports subversion too, so maybe we can set that up now :)
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: pabs</title>
		<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-244</link>
		<pubDate>Fri, 03 Sep 2004 17:40:28 +0000</pubDate>
		<guid>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-244</guid>
					<description>/var/lib/cvs is for anon, /cvs is for spiffy ssh users like  myself.  And you forgot to tell them about the ViewCVS interface you.</description>
		<content:encoded><![CDATA[<p>/var/lib/cvs is for anon, /cvs is for spiffy ssh users like  myself.  And you forgot to tell them about the ViewCVS interface you.
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: giblet</title>
		<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-242</link>
		<pubDate>Fri, 03 Sep 2004 11:56:49 +0000</pubDate>
		<guid>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-242</guid>
					<description>Yeah - I feel the same, only I don't feel like bothering to set it up as I'm in the process of switching to subversion :D</description>
		<content:encoded><![CDATA[<p>Yeah - I feel the same, only I don&#8217;t feel like bothering to set it up as I&#8217;m in the process of switching to subversion :D
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: richlowe</title>
		<link>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-241</link>
		<pubDate>Fri, 03 Sep 2004 09:28:15 +0000</pubDate>
		<guid>http://linuxbrit.co.uk/blog/2004/08/31/feh-work/#comment-241</guid>
					<description>That explains a lot, I've never touched pserver.
pserver is evil, pserver should never be used, &lt;b&gt;ever&lt;/b&gt;.

Maybe it doesn't like it being a link, or (more likely), maybe nobody told pserver that /cvs is valid...

I'd much rather something like anoncvssh was used though.

(the thingy that acts a shell for a nonpriviledged account and does nothing other than allow anonymous cvs checkouts over ssh, I think the name is anoncvssh anyway, something like that :))</description>
		<content:encoded><![CDATA[<p>That explains a lot, I&#8217;ve never touched pserver.<br />
pserver is evil, pserver should never be used, <b>ever</b>.</p>
<p>Maybe it doesn&#8217;t like it being a link, or (more likely), maybe nobody told pserver that /cvs is valid&#8230;</p>
<p>I&#8217;d much rather something like anoncvssh was used though.</p>
<p>(the thingy that acts a shell for a nonpriviledged account and does nothing other than allow anonymous <acronym title="Concurrent Versions System">CVS</acronym> checkouts over ssh, I think the name is anoncvssh anyway, something like that :))
</p>
]]></content:encoded>
				</item>
</channel>
</rss>
