LJ braintrust time

this from my mom:

Sybil and a friend want to go to a nice restaurant in the Raleigh area where they can have a glass of wine and perhaps listen to some live music.

So, any recommendations out there?


my weight set / smith machine was sticking, so I sprayed on some X-1R (or whatever that stuff is called that Gary bought for the door) and it’s nice and smooth now. It allowed me to increase my squats weight yesterday, which I’m feeling today 🙂

Oh, and my weekend was spent in Charlotte with Chad/Danny/Leroy/etc. – a few hours of basketball, a couple hours doing biceps (which I never do in my home routine)and back at the Wachovia gym (nice place!) – lots of video games (NBA Street vol. 2, various wrestling games, Super Tennis, MarioKart, etc.) and lots of movies (Vampire Hunter D: Blood Lust was excellent, I only saw about 1/3rd of Blood: The Last Vampire, oh, and I finally saw The Usual Suspects, although they foreshadow the ending way too much for it to have been a surprise IMHO)

since LJ keeps eating my comments, i'll make it a post

well, i had a massive explanation here of the whole situation, but LJ went and ate it – that’s what I get for not doing big comments in a separate editor … *sigh*

so, short(er) answer: yes, that’s it, and 99% of the time that’s fine, but IIS4/NT4 has some corner cases where it needs to obey the spec and almost always does so (cases where it doesn’t know the size of the content when it sends the http headers, so it has to Connection: close and then rely on the client to accept data til the server closes the socket). Unfortunately, there are corner cases (ASP output too large, overflows the ASP buffer, looks to be the case) where IIS “forgets” that the current socket is one of the ones that needs to be fully closed and ends up at the end still just half-closing.

Wouldn’t be nearly as much of an issue if it supported things like chunked encoding (IIS5 does, thankfully). I hate http transactions with “read til I close the socket” – they’re just asking for trouble, and in this particular case, it’s found it 🙂

tcp question

is it valid to have the last data packet from the server include the FIN bit set to double as the first phase of the FIN/FIN-ACK/ACK closing procedure?

I’m doing ethereal traces of IE6 and IIS4 talking to each other.

I need to go check the spec, methinks 🙂


slept 10 hours monday night, 5.5 hours last night. How very odd.

I’m buying into this “bell curve” theory has been explaining (the whole long-distance, 65%-75% maxHR, etc). I’ve been putting my target as 70%-75%.

Last night because of doing legs (squats, leg extensions, etc) I didn’t feel like I could really “run” so I ended up (relative to me) speed-walking around 4.0 mph with enough arm activity to maintain the 73% level for about 100 minutes or so.

Unrelated, but my laptop is doing the “constant loud beep” thing, and turning it over last night it seems like the CPU fan’s not starting up, so I’m guessing that’s the problem. It did this before, maybe 2 years ago, but just fixed itself that time. Looks like I’ll have to break out a screwdriver tonight, no idea what I’m going to be able to actually get to inside a mobile PII module, though. Looks like the fan attempts to start at first, it kind of jerks, so I’m not sure what’s up… fun fun fun

oh, and filed a bug against eclipse 2.1 yesterday. I like reproducible bugs. Why? Because I encounter them so rarely at work these days 😛

it's a good thing the IT director guy doesn't take me too seriously

jdownhower3 (8:17:40 AM): how about 1:30?
flavorjames (8:18:18 AM): i’ve gotta leave by 2 to pick her up, but half an hour may be enough
jdownhower3 (8:20:20 AM): For you, 1PM it is.
flavorjames (8:39:57 AM): I will build a temple and sacrifice many a goat in your honor, sir!
jdownhower3 (8:40:58 AM): lovely, but please do not sacrifice any of Busa’s goats….
flavorjames (8:41:16 AM): their lives will be spared… for now!