<?xml version="1.0"?>
<rss version="2.0.">
  <channel>
    <title>Advogato blog for kevinvv</title>
    <link>http://www.advogato.org/person/kevinvv/</link>
    <description>Advogato blog for kevinvv</description>
    <language>en-us</language>
    <generator>mod_virgule</generator>
    <pubDate>Mon, 7 Jul 2008 01:44:13 GMT</pubDate>
    <item>
      <pubDate>Mon, 28 Jan 2002 06:30:38 GMT</pubDate>
      <title>28 Jan 2002</title>
      <link>http://www.advogato.org/person/kevinvv/diary.html?start=6</link>
      <guid>http://www.advogato.org/person/kevinvv/diary.html?start=6</guid>
      <description>&lt;p&gt;Well, It's been about a month since I last posted.  
For the first two weeks I was cruising in Central and 
South America, then worked for a week, and now back 
at school for a week.  So overall it was a great break, 
going to Mexico, Costa Rica, Panama, Colombia, 
Aruba, and Jamaica with the family.  Fun times, without 
email or osama bin laden.  Met some cool people on 
the cruise too, and been keeping in touch.
&lt;p&gt;
CoreFoundation is working well on Win32 now.  
CFSockets are fully functional, and CFRunLoops work, 
though I still need to go back and add in support for 
CFRunLoopTimers.  Last night I decided to try CF out 
on a Linux box for fun, and after a little bit of work it was 
mostly working.  I put in some CFRunLoop support 
using pthreads condition variables.  So it's essentially 
working, but I havent' yet found a suitable Linux API for 
implementing CFRunLoopTimers.

&lt;p&gt;
Also been working quite a bit on SystemStarter, and 
making some good progress.  Now items can display 
messages onscreen using an IPC call.  Items can also 
signal whether they succeed or fail.  Also they can get 
the value of configuration setting from SystemStarter 
(i.e. ask if it's a verbose or safe boot.)  Lots of nifty 
features.
</description>
    </item>
    <item>
      <pubDate>Fri, 28 Dec 2001 08:39:51 GMT</pubDate>
      <title>28 Dec 2001</title>
      <link>http://www.advogato.org/person/kevinvv/diary.html?start=5</link>
      <guid>http://www.advogato.org/person/kevinvv/diary.html?start=5</guid>
      <description>&lt;p&gt;
Been working on a project which intersects with a lot of 
the functionality of Darwin's CoreFoundation.  Namely, 
a TCP/IP server application that does a lot with XML.  
The interesting part is that the market for this program 
is windows machines, and eventually some unix 
machines.

&lt;p&gt;
CoreFoundation works remarkably well on Windows, it 
took a little patching to get the RunLoop and CFSocket 
code up to speed, but now it seems to be working well, 
and makes for some highly portable application code.  
Already I've run into a number of areas where my life is 
significantly easier because I have some 
CoreFoundation data-types to back me up.

&lt;p&gt;
I'm leaving civilization for a couple weeks, but when I 
return I have to organize my patches, make sure they're 
sane, and then send them back for integration.  Most of 
the stuff outside of CFSocket will be good fixes for 
Darwin (mostly code laziness that gcc doesn't freak out 
over), and the CFSocket stuff is mostly windows 
specific fixes.
</description>
    </item>
    <item>
      <pubDate>Fri, 14 Dec 2001 02:00:44 GMT</pubDate>
      <title>14 Dec 2001</title>
      <link>http://www.advogato.org/person/kevinvv/diary.html?start=4</link>
      <guid>http://www.advogato.org/person/kevinvv/diary.html?start=4</guid>
      <description>&lt;p&gt;
I packaged up the
&lt;a href="http://cory.eecs.berkeley.edu/~kevinvv/
SSHAskPassword.html" &gt;SSHAskPassword.app&lt;/a&gt; 
and posted it on my site.  It's a neat little addition to the 
SSHAgentServices I'm working on--it lets the various 
ssh utilities prompt for a password in Mac OS X when 
not attached to a terminal.  This should make things 
like scp droplets easier.  (If only the Finder would 
cooperate.)
</description>
    </item>
    <item>
      <pubDate>Wed, 12 Dec 2001 12:22:39 GMT</pubDate>
      <title>12 Dec 2001</title>
      <link>http://www.advogato.org/person/kevinvv/diary.html?start=3</link>
      <guid>http://www.advogato.org/person/kevinvv/diary.html?start=3</guid>
      <description>&lt;p&gt;
I fired off my &lt;a href="http://www.advogato.org/proj/SSHAgentServices/" &gt;
SSHAgentServices&lt;/a&gt; plugin to 
Rick since he's as 
much of an ssh nut as I am.  He noticed something I 
hadn't... apps launched by the Finder didnt' get the 
environment variables.  I guess I'm enough of a Dock 
and Terminal weenie that I never use the Finder and 
didn't catch that one.  Anyway, I noted the restriction in 
my readme and started trying to figure out how to get 
around it.

&lt;p&gt;
I found 
&lt;a href="http://developer.apple.com/qa/qa2001/qa1067.html" &gt;
QA1067&lt;/a&gt; which indicated the Finder would pass 
along environment variables.  So what happened?

&lt;p&gt;
Found it.  The Finder is launched before the plugins are 
run, the Dock afterward.  That explains why the Dock 
gets the variables and the Finder doesn't.</description>
    </item>
    <item>
      <pubDate>Mon, 10 Dec 2001 14:52:25 GMT</pubDate>
      <title>10 Dec 2001</title>
      <link>http://www.advogato.org/person/kevinvv/diary.html?start=2</link>
      <guid>http://www.advogato.org/person/kevinvv/diary.html?start=2</guid>
      <description>&lt;p&gt;Woohoo!  Well less than 24 hours ago I decided to 
bite the bullet and make a real ssh-agent solution for 
Mac OS X.  I'm pleased to announce its completion.&amp;lt;/
p&amp;gt;
&lt;p&gt;&lt;a href="http://cory.eecs.berkeley.edu/~kevinvv/
SSHAgentServices.html" &gt;SSHAgentServices&lt;/a&gt;
&lt;p&gt;I probably should make up a little project page on 
here... anyway, the problem was that ssh-agents could 
be started in individual terminal windows, but when a 
new window is created, the information wouldn't carry 
over.  There have been a few scripts to deal with this, I 
invented my own and used it for a while, but this is 
what's really been needed all along.
&lt;p&gt;The problem is detailed in &lt;a href="http://
www.opensource.apple.com/bugs/X/LoginWindow/
2575100.html" &gt;Radar 2575100&lt;/a&gt;
&lt;p&gt;So it took a lot longer to make than it should have, 
mostly from silly mistakes on my part.  For example 
calling kill() with the arguments transposed, and 
figuring out to call kill() in the first place... but luckily I 
had iPod to keep me company...and a chorus of 
crickets...
&lt;p&gt;Pretty soon I'll see if i can get a GUIfied ssh-add 
going.  Perhaps one that reads passphrases out of the 
keychain which is unlocked right after login.
&lt;p&gt;Now that I've reverse-engineered the loginPlugin 
API, I've got some other ideas that I'm going to work 
on... mostly to do with setting environment variables; 
probably I'll read them out of a plist that can be 
modified via the SystemPreferences...and I think there 
are some new docs out on that API, so I'm all set.</description>
    </item>
    <item>
      <pubDate>Mon, 10 Dec 2001 09:12:51 GMT</pubDate>
      <title>10 Dec 2001</title>
      <link>http://www.advogato.org/person/kevinvv/diary.html?start=1</link>
      <guid>http://www.advogato.org/person/kevinvv/diary.html?start=1</guid>
      <description>&lt;p&gt;
For a while now I've been working with Fred 
S&#xE1;nchez on Darwin's SystemStarter project.

&lt;p&gt;The initial version of SystemStarter was basically a 
replacement of the &lt;tt&gt;/etc/rc&lt;/tt&gt; mechanism, with 
some neat forward-looking design decisions on Fred's 
part.  I've found this area of operating systems to be 
traditionally weak.  The infancy (and malleability) of 
Darwin, combined with the neat work Fred started, got 
me interested in the project.  Since then, Fred 
and I have been working on more advanced features, 
some of which are unique to Darwin.
&lt;p&gt;Overall I think it's pretty exciting stuff.  SystemStarter 
is breaking away from an entrenched part of UNIX, 
and providing something more robust.
&lt;p&gt;The SystemStarter work is discussed in the paper 
"&lt;a href="http://www.opendarwin.org/~kevin/
SystemStarter.html" &gt;SystemStarter and the Mac OS X 
Startup Process&lt;/a&gt;" which I had the opportunity to co-
author along with Fred, and will be presented at &lt;a href="http://www.usenix.org/events/bsdcon02/" &gt;
BSDCon 2002&lt;/a&gt;.  I'm really excited about BSDCon; 
it's the first time I've been published, and it 
helps to fuel that typical Apple-
related delusion of working on software that makes a 
difference :-)

&lt;p&gt;
The ultimate goal is to allow services to be 
started, stopped, and restarted dynamically; and to 
have them communicate with SystemStarter enough 
that SystemStarter can make informed decisions.  
</description>
    </item>
    <item>
      <pubDate>Mon, 10 Dec 2001 07:42:02 GMT</pubDate>
      <title>10 Dec 2001</title>
      <link>http://www.advogato.org/person/kevinvv/diary.html?start=0</link>
      <guid>http://www.advogato.org/person/kevinvv/diary.html?start=0</guid>
      <description>&lt;p&gt;Well, here's my first entry on advogato.  I ran across 
the site during a google search, and noticed a few 
people I know, so I thought I'd give it a try.  Looks like 
everyone else has been at this a while, sometimes I'm 
really out of the loop.
&lt;p&gt;
I wanted to step to the plate swinging, so I also created 
a project for my &lt;a href="http://www.advogato.org/proj/PerlService/" &gt;
PerlService&lt;/a&gt; which is a 
useful little hack I put together but haven't publicly 
released yet.  Everyone I've showed it to has liked it 
though, (provided they don't turn a deaf ear when the 
word "Macintosh" is uttered), so now it's out for all to 
enjoy.
&lt;p&gt;
Currently I'm battling an infestation of crickets, thanks to 
a prank that someone pulled on the guy living upstairs.  
Unfortunately crickets know no political boundaries, 
and I've been atrributed to collateral damage.

&lt;p&gt; &lt;p&gt;
I'm also battling a wretched latency to the outside world 
thanks to some Windows machine in the network I'm 
on.  At first I thought it was a chatty NIC, but I've been 
able to rule that out, it's just the wonderful world of 
Win32.

&lt;p&gt; &lt;p&gt;If my entry seems a little frantic, it's probably 
because I'm dealing with 3000ms latency between 
ethernet packets, and 50ms latency between cricket 
packets.  Unfortunatley the underwhelming and 
overwhelming performances don't interefere 
destructively.
</description>
    </item>
  </channel>
</rss>
