Author Topic: Feature Request: First Start on Linux Notification (explained inside)  (Read 3183 times)

Zorchenhimer

  • Autococker
  • Posts: 2614
I think it would help those who use linux to have a "Loading" dialog pop up when the game loads for the first time.  Or at least some notification when launched from the command line when the random number generation thing is taking place.  That way, you would be able to see if the game is in fact doing something and not just hanging there.

Cameron

  • Global Moderator
  • Autococker
  • Posts: 2686
That's a good Idea.  I remember loading about 10 paintball's until 1 loaded and got stuck and didn't know which window i was in :P.

jitspoe

  • Administrator
  • Autococker
  • Posts: 18802
I think this might be the better solution:

http://dplogin.com/dplogin/featurevote/feature.php?id=10105

Zorchenhimer

  • Autococker
  • Posts: 2614
Haha, forgot about that one.

Krizdo4

  • PGP
  • Posts: 43
I think this might be the better solution:

http://dplogin.com/dplogin/featurevote/feature.php?id=10105

On linux, the /dev/urandom provides non-blocking pseudo-random data. If the quality of that is better than the windows random source being used or at least good enough then I don't see any need for using the high-quality random random generator /dev/random.
It would seem to speed up generating the key saving you the trouble of setting up it's own thread.

jitspoe

  • Administrator
  • Autococker
  • Posts: 18802
On linux, the /dev/urandom provides non-blocking pseudo-random data. If the quality of that is better than the windows random source being used or at least good enough then I don't see any need for using the high-quality random random generator /dev/random.
It would seem to speed up generating the key saving you the trouble of setting up it's own thread.
Problem is, gcrypt uses the super-incredibly-painfully-slow random number generation, even when you tell it to use the fast.  I'm tempted to hardcode the source to use the fast and compile it into the binary.

Krizdo4

  • PGP
  • Posts: 43
Problem is, gcrypt uses the super-incredibly-painfully-slow random number generation, even when you tell it to use the fast.  I'm tempted to hardcode the source to use the fast and compile it into the binary.

Is this what you're refering to when you say you're telling it to use fast.
GCRYCTL_ENABLE_QUICK_RANDOM;

jitspoe

  • Administrator
  • Autococker
  • Posts: 18802
I can't remember if that was it specifically, but I traced through the source and it used the slow random every time.