The Mighty Buzzard

Established Members
  • Content Count

  • Joined

  • Last visited

Everything posted by The Mighty Buzzard

  1. Just for clarification, the Windows binary is what runs under wine; the linux server is native code and doesn't use wine.
  2. @havox: You have to set that in utserver.conf if you want it to stay and you can ignore the checkbox.
  3. Ok, here's the version combinations I've tried and the result after leakiness or not became clear. Wine 1.0.1 comes from the official ubuntu repo (proposed) and the 1.1.x versions come from winehq's repo. I'll try and narrow down the highest version that doesn't have an issue. wine 1.1.18 + uT 1.8.3 (15104): leaked like hell, up to 800+M overnight wine 1.1.18 + uT 1.8.2: leaked like above but shut it down after it passed 200M wine 1.0.1 + uT 1.8.3 (15104): overnight and only 24M wine 1.1.15 + uT 1.8.3 (15104): leaked enough that i called fail at 110M wine 1.1.0 + uT 1.8.3 (15104): overnight 21M wine 1.1.10 + uT 1.8.3 (15104): called fail after 3 hours at 89M wine 1.1.5 + uT 1.8.3 (15104): five hours 20M wine 1.1.6 + uT 1.8.3 (15104): 13 hours 20M wine 1.1.7 + uT 1.8.3 (15104): 25 hours 21M wine 1.1.8 + uT 1.8.3 (15104): 4 hours 86M Guess 1.1.7 is as high as I'll be going until I find out what's causing the leak.
  4. @arvid: Yep, takes between a couple hours and overnight but it always happens. Also happens with uT 1.8.2 though, so I'm thinking it may very well be a wine bug or something else peculiar to this particular box. It could very well be a pain to track down though being as I noticed this the weekend after a wine update, uT update, kernel update, and migration to an encrypted ext4 /home partition. You'd think I'd know by now to let major changes sit for a while to see if they screwed anything up before making another but apparently I don't always. I'll try a few different wine versions this weekend and cross my fingers.
  5. PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 955 bob 20 0 349m 155m 25m S 0.0 7.7 143:32.88 firefox-bin 9421 root 20 0 126m 62m 23m R 5.3 3.1 348:55.82 Xorg 10058 bob 20 0 2607m 436m 18m S 1.0 21.5 14:49.58 uTorrent.exe Build 14809, wine-1.1.17. The memory leak is for sure solved in Windows?
  6. That'd be because even with the small wine overhead it's nowhere near java's level of bloat.
  7. Looks to be working on slack 11 with wine 0.9.31, aside from DHT which occasionally flakes out and requires a restart of uT to get back happy and the already mentioned minor issues.
  8. Damn yous, I almost spit beer all over the place!
  9. Come on now, Firon... it'd only double the size of uT to include pcre. As for Falcon4, I scoff at your puny regex skills but i'll agree that mod_rewrite implements them rather screwily. Give me good ole s/(.*?\.php)\/(.*?)\/(.*)/$1\?$2=$3/ any day.
  10. Yeah, I mean it's not like you even get to use all the fun words to tell someone what kind of git they're being.
  11. garnertr: Ultima just likes arguing more than Firon. Always has. Either that or he's just really easy to wind up, hard to tell.
  12. one50bpm: You're trying to use a non-routeable address outside of the intended lan is likely the problem. You can try till you're blue in the face and you'll never get a 192.168.*.* address to be routeable. If you don't have a static public IP have a look at or some such, do some reading, and start using your new cannonical name rather than an IP address (eg After that all you have to do is make sure port 8080, or whatever you chose, is properly forwarded in your router.
  13. You could always just block incoming BC handshakes. That mitigates a lot of BC's bad behavior by keeping it from being able to connect and disconnect to get itself unchoked but still lets you connect to BC clients. As Firon said though, this excludes BC 0.63. It's pretty easy to do with Kerio but you'll have to figure it out yourself for anything else.
  14. He was probably counting ludde's coffee mug too. I know that's my best coding partner.
  15. @Kraiser: You could have left out the "when I find out how" bit. Now all of us ancient (over 25) linux geeks have to do the Dance of Shame.
  16. @Hirvine+Abigail: Read at least the two most recent pages before you ask a question like that, please? In any case, 4-6 hashfails on a torrent isn't all that uncommon. All it takes is connecting to one or two peers who're sending bad data for whatever reason.
  17. You just completely misunderstood what that setting means, skynetman. It doesn't increase or decrease the number of slots, it increases the maximum active number of downloading or uploading torrents if one isn't seeing enough traffic to justify using up an active spot. The setting you're wanting is on the Torrent Options page. It's called Use additional upload slots if upload speed < 90%.
  18. You were probably distracted by your amusing avatar. Speaking of which....
  19. No, it won't. If an incoming peer has encryption disabled there is nothing whatsoever you can do to make the connection be encrypted. The only thing you can do is uncheck "Allow legacy incoming connections" which will simply not allow them to connect to you at all. You can still make outgoing connections to them that will be encrypted, assuming they're using one of the clients that support the new encryption, though.
  20. Yeah, that's how it's supposed to work. You told yours to never try to make encrypted connections so it isn't trying. He told his to only try to make encrypted connections so his does. What'd you expect?
  21. I'm seeing the same even though the DHT and PEX boxes are properly greyed out on the torrent's properties page.
  22. 64kbps * 0.8 = 51.2kbps (subtracting 20% for overhead and line noise) 51.2kbps / 8 = 6.4kBps (eight bits in a byte) Twenty percent is quite generous for overhead. Sixty two point five percent is being a leeching jackass. I got better than 3kBps upload on dialup. If you have enough line noise that you lose half of your signal to line noise then you need to go outside and get to running new lines or bitching daily to your telco if it's on their line.