Category computers

[Cross posted from for archival purposes…]

Well, I guess I’ll be the first to complain about 10.4.3 here.

#1 — The Mail.app problem I complained about before (and submitted a bug report to Apple on) has not been fixed.

#2 — Despite adding more RAW support, the Coolpix 5400’s RAW (NEF) files remain unsupported. This is supported by dcraw so Apple has no excuse for not making it work. This too has been submitted as a bug back in April or May.

#3 — While Safari has gained the ability to render the ACID2 browser test I have been able to make it break. I visit the ACID2 test page itself, then after the rendering is complete I spin the ball on my mouse (a Mighty Mouse, in this case) and the face breaks and scrolls even though there is no scroll bar. Here is a screenshot of the problem.

#4 — The out-of-the-blue failing of my attempt to rsync lots of data to an external drive continues to fail as follows:

Documents/iChats/Erika Alpert on 2005-09-09 at 23.57.ichat
        3500 100%   18.99kB/s    0:00:00  (6610, 10.0% of 185201)
rsync: fstat failed: No such file or directory (2)
❚

Then after being sent a break:

^Crsync error: received SIGUSR1 or SIGINT (code 20) at /SourceCache/rsync/rsync-20/rsync/rsync.c(231)
rsync: writefd_unbuffered failed to write 103 bytes: phase "unknown" [generator]: Broken pipe (32)
rsync error: error in rsync protocol data stream (code 12) at /SourceCache/rsync/rsync-20/rsync/io.c(909)
mercury:~ c0nsumer$ 

This hasn’t yet been reported to Apple as an official bug report yet because I don’t have a standard test case for it yet. However, it worked under 10.3.9 just fine. Unfortunately this is the bug I can’t work around, because even copying data to the external drives with Finder fails in the same manner.

So, Apple… Drop the ball much?

Okay, a moment for me to bitch about the Mozilla people:

Who the hell makes an application which supports SOCKS as a proxy type, but doesn’t provide the option for name lookups to be done via the proxy? Both Thunderbird and Firefox have been this way forever, and only a Mozilla nightly properly presents this option.

Maybe the only people who use SOCKS these days are kiddies bouncing through open proxies? Or maybe the only users of SOCKS are on weirdly configured networks which will resolve any name, but only provide connections out through a single box?

Dumb dumb dumb dumb dumb.

You heard it here first, kiddies… MS05-044 through MS05-052 have been released. Go patch your computers.

[Cross posted to and for archival purposes…]

Last night I ran into an interesting issue with regards to printing from an XP machine to an iMac running OS X 10.4, and I thought people here would want to know about it.

First, the Mac has an HP Deskjet 952c connected to it, using the queue name which was automatically created when the printer was plugged in.

Second, the client machine is running XP SP2 and has simply mapped to the printer over the network and has the appropriate driver loaded.

The problem I was observing is that the XP machine would send a print job, and the Mac would report it as completed, but nothing would have actually been printed.

The solution to this problem is to access the CUPS web-based management stuffs (http://localhost:631) and edit the printer so that it’s queue name has no spaces in it. In my case I just removed the spaces and replaced them with underscores. Next I deleted the mapped printer on the XP machine and reconnected it, and everthing was fine.

I’ve been told by a friend that OS X Server actually warns a user about such potential issues when creating a print queue, but I can say that the client version of the OS doesn’t present the user with any sort of notification.

So, yeah. I hope that helps someone.

[Cross posting this from because of smarter people here. Please delete if inappropriate.]

Does anyone here have any pointers to good info on OS X’s pagefile? It appears that it is all handled by dynamic pager (8), but I’m having a hard time finding info beyond simply moving the pagefile to another disk.

The reason I’m asking is because, in my experience, one can greatly improve paging performance by allocating pagefile space on multiple spindles. As I’ve got a pair of disks in my G5, that’s exactly what I’d like to do.

As it is, it looks like dynamic_pager can only take one argument for the base filename for the pagefile. Hopefully this isn’t the case…