[FFmpeg-devel-irc] IRC log for 2010-08-14

[irc at mansr.com]

[02:35:31] * saintdev wonders where peloverde is :/
[04:33:44] <saintdev> P9wU6_eb
[04:33:53] <Dark_Shikari> is that a password?
[04:34:21] <saintdev> oops
[04:34:22] <Dark_Shikari> saintd3v: did you just type your password? =p
[04:34:24] <Dark_Shikari> you should change it.
[04:34:43] <saintdev> good thing that's only for local logins ;)
[04:34:53] <saintdev> either way already changed
[04:37:34] <saintdev> and that's why i use pubkey auth for any remote connections :P
[07:02:54] <andoma> mru: merbanan: still on for lunch today?
[07:08:54] <merbanan> andoma: yes
[07:09:13] <kshishkov> mru tried to ping you yesterday, so probably yes
[07:09:17] <merbanan> mru: is probably in the air or going to the airport
[07:09:27] <merbanan> andoma: what time ?
[07:10:14] <andoma> 12?
[07:10:26] <andoma> so MÃ¥ns arrives here today?
[07:10:47] <merbanan> I hope so :)
[07:10:49] <kshishkov> he was in Sweden this week
[07:11:30] <merbanan> ok, then I guess he is on his way to stockholm from dalarna
[07:13:37] <merbanan> andoma: he was on the train
[07:13:59] <merbanan> which would arrive around 1200
[07:14:13] <merbanan> so he said 1300 would be a suitable time
[07:14:17] <andoma> okok
[07:14:30] <merbanan> @ cafe nero ?
[07:16:48] <andoma> yes
[07:21:43] <merbanan> andoma: what is the address for cafe nero ?
[07:22:23] <kshishkov> I suppose it's Roslagsgatan 4
[07:22:34] <kshishkov> (if Google Maps doesn't lie)
[07:24:32] <andoma> yes: http://gulasidorna.eniro.se/f%C3%B6retag/caf%C3%A9-buco-nero:66106387
[07:28:52] <merbanan> andoma: ok, message sent, this time he should be there
[07:29:00] <andoma> :)
[07:37:56] <_av500_> gm
[08:21:37] <elenril> so how is gsoc looking so far?
[14:43:14] <kierank> someone review libavsequencer ;)
[14:48:28] <lu_zero> aargh
[14:48:33] <lu_zero> too much code =P
[16:44:59] <kierank> any idea what this code is trying to achieve? http://pastebin.org/483277
[16:46:34] <twice11> Did you get that from decompiling?
[16:47:17] <twice11> The second line sets foo to "-1U" if foo is != 0.
[16:47:26] <kierank> it comes from a large block of asm
[16:47:34] <twice11> The third line translates -1U to 3.
[16:47:39] <kierank> foo is in this case always positive
[16:47:54] <twice11> The first line is unimportant.
[16:48:13] <twice11> Lines 1-3 boil down to foo = foo ? 3 : 0;
[16:48:32] <twice11> So Lines 1-4 are "foo = foo ? 6 : 3";
[16:48:48] <twice11> The sign of foo doesn't matter for testing nullness.
[16:49:32] <twice11> Lines 2-4 are typical compiler output for the expression I quoted. No idea why line 1 is negating first.
[16:50:28] <kierank> thanks, that makes sense
[16:51:08] <twice11> Hmm. I think I know what line 1 is up to.
[16:51:41] <twice11> its the C version of: neg eax; sbb eax, eax; and eax, 3; add eax, 3
[16:52:12] <twice11> neg sets the carry flag if eax is non-zero, so CARRY = !!eax
[16:52:24] <twice11> It's only executed to set the carry flag.
[16:53:26] <kierank> yes, that is what the asm was
[16:53:29] <twice11> So I am *quite* sure that the lines you quoted are decompiler output. (Although maybe manual decompiling)
[16:53:41] <kierank> yes, "manual" decompiling
[16:54:16] <twice11> Quote it as assembly next time. It's uncommon C code, but a usual x86 assembly idiom.
[20:35:47] <CIA-93> ffmpeg: rbultje * r24796 /trunk/libavformat/tcp.c:
[20:35:48] <CIA-93> ffmpeg: Print error messages in case of connection failure or name resolution failure
[20:35:48] <CIA-93> ffmpeg: in tcp.c.