[vdr] Re: Version of 1.3 that works ?

[Rainer Zocholl <UseNet-Posting-Nospam-74308-@zocki.toppoint.de>]

luca@ventoso.org(Luca Olivetti)  14.12.04 11:12

Once upon a time "Luca Olivetti " shaped the electrons to say...

>Rainer Zocholl wrote:


>> try 1.3.17 with the thread save patches(you MUST apply them!)

>It's not necessary if your glibc is compiled with thread local storage
>enabled (as is the case with mandrake 10.0). 

That's the -broken- tls lib, or?

>In fact I have everything working (including streamdev) 
>with none of these patches

Yes not the slighest doubt that it will work.
As long as the CPU is very slow or very fast or only
one thread is really running.
The timing windows are very smal.

But here it was absoletely reproducible:
osdteletext plugin make vdr segfaulting.
patching osdteletex does not help
but applying thet thread safe patches to vdr helps.
It was very very convincing.


>(though I have some occasional vdr crash).

The timing windows are very small.


But i don't see how "tls" should be able to help with
reentrance problems.
I can imagine how a lib function can forced to used a 
thread specific data segment.
but I can't imagine how "thread local storage" can make a
non-reentrant function "reentrant", "recursive usable". And i will 
not swear that no functions that used non-"_r" functions will never 
be called in an other context...by the same thread.
It sound that there may two (very similar) things beiing mixed up: 
"thread safeness" and "reentrance"?

Rainer