skype version

[Gene Heskett]

Antoine wrote:
> Hi,
> Does anyone know the best version (download, ie, FC3 or binary) of
> skype to install on FC5?
> Cheers
> Antoine
> 
I believe they all reduce to the exact same code.  Which is in need of 
help on my machines (note plural) at least.  Booted to XP on this HP 
dv5320us lappy, the windows version works well, booted to linux and 
running that code with wine repeatedly crashes 54 seconds after a call 
is connected.  Running the linux code, I sound good on the other end, 
but it'll get into a stuttering state on my end and I lose reception.

This latter condition also exists on my FC2 box at home, using the 
nforce2 ac97 on board chipset there, with virtually identical and 
unusable  performance to the ATI IXP chipset in this lappy.  I don't 
believe that the oss emu support in alsa is up to such usage. FWIW, 
ekiga fails on my end even worse.  I've never hear a complete word in 
the headset when running ekiga, but am told I sound good on the other end.

None of these VoIP "sip:" apps, even 'twinkle' which is in active 
development with a fresh release just 3 weeks ago, seem to have made the 
transition from oss to alsa, or are willing to make the effort to make 
that transition.  I built it here yesterday, its prefs demand the 
connexant 30 oss emulation from alsa, and it crashes, requiring htop 
help to clean up the mess if you try to make it use the ATI-IXP alsa 
interface.

That said, I heard a rumor but can't substantiate it, that the linux 
version of skype is being re-written virtually from scratch.  The only 
thing I know for sure is that the yum repo for skype has been taken 
offline, causing me to have to remove that file from my /etc/yum.repos.d 
  directory else yumex hangs for a few minutes.

And that the common factor between all the failures is alsa.

Presumably if I was to install the shareware oss system, it would work, 
but while I do have it for the older FC2 box, its not been updated in a 
couple years during which time the motherboard on that machine has been 
replaced.  But I believe that test would be the end of the ultimate 
pissing match I've become embroiled in on alsa's bugzilla list over its 
lack of full duplex operation in anything but carefully preset 
laboratory experiments.  It does work nicely there, but not in the real 
world, which the alsa folks are apparently refusing to consider as a 
real problem because their response is 'fix the app, alsa works'.

-- 
Cheers, Gene