Thread safeness of cint?

View: New views
2 Messages — Rating Filter:   Alert me  

Thread safeness of cint?

by Joost Kraaijeveld :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi,

The source of cint explicitly say that cint is not MT safe. Is that
still the case for cint7? If so, does anyone know if it is a lot of work
(going to impossible without a complete rewrite)?

TIA

--
Groeten,

Joost Kraaijeveld
Askesis B.V.
Molukkenstraat 14
6524NB Nijmegen
tel: 024-3888063 / 06-51855277
fax: 024-3608416
web: www.askesis.nl



Re: Thread safeness of cint?

by Axel Naumann :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi Joost,

cint5 is easier to use in a MT environment than cint7, because it uses
less STL containers. I believe Cint5 was once sort of thread safe. Sort
of, because the trick was to load a separate libCint.so per thread,
having separate functions and most importantly separate globals. You can
see in the lower half of G__ci.h how this is done. We have not tried
this for a long time, so it will most probably not work. But it should
be the easiest way for you, if the "one libCint.so per thread" approach
is acceptable for you.

Cheers, Axel.


On 2009-09-25 10:23, Joost Kraaijeveld wrote:
> Hi,
>
> The source of cint explicitly say that cint is not MT safe. Is that
> still the case for cint7? If so, does anyone know if it is a lot of work
> (going to impossible without a complete rewrite)?
>
> TIA
>