comp.lang.ada
 help / color / mirror / Atom feed
From: Adrian Hoe <adrianhoe@nowhere.com>
Subject: Re: gcc/gnat 3.3
Date: Fri, 07 Nov 2003 15:15:25 +0800
Date: 2003-11-07T15:15:25+08:00	[thread overview]
Message-ID: <3FAB468D.7030509@nowhere.com> (raw)
In-Reply-To: boffui$1eau5s$1@ID-198985.news.uni-berlin.de

Andreas Almroth wrote:

> Adrian Hoe wrote:
> 
>>
>> When I tried to run a simple tasking program, I got "Segmentation 
>> Fault (code dumped)". Could gcc not build properly? I believe the 
>> threading was not build properly. Or did I miss some parameters?
>>
> 
> Right, as I mentioned earlier in the thread, tasking is not working on 
> Solaris 9 (SPARC). The sort of workaround (hack in my opinion) that 
> worked up to Solaris 8, does not work, as Sun apparently changed the 
> behaviour on how in the memory keep track on threads.
> 
> Although I fairly sure it won't work, you could experiment with using 
> alternative thread libraries, and hope that the ada code is designed to 
> change its behaviour. Check in the configure doc for supported threading 
> models by gcc, and hopefully those works with ada rts as well.
> 
> In the gnat 5.0 sources there is a new implementation, which I have 
> tried to use in 3.3 (not 3.3.2), but it did not work. Fault on my side 
> I'm sure.
> The upcoming version of gcc is rumored to have a newly synced ada tree 
> from the latest? gnat, so we will have to wait and see.
> 
> I'm currently on Solaris 8, so I have no problem using tasking in the 
> application I'm developing, but migration to Solaris 9 is "on wait" 
> until tasking is proven to work.



So, the tasking problem is caused by Solaris 9. I'm sure ACT or someone 
is working on this issue? Does anyone know of any progress?

Too bad I don't have the Solaris 8 CDs. It came pre-installed with 
Solaris 9 CDs bundled in SB150. :( I guess I will have to wait till a 
gnat workaround has been released. :(

How about Solaris 10? I read there is a major work out on Solaris 10 
architecture. Any news?
-- 
Adrian Hoe
m a i l b o x AT a d r i a n h o e . c o m




  reply	other threads:[~2003-11-07  7:15 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-10-17 14:45 gcc/gnat 3.3 Adrian Hoe
2003-10-17 21:12 ` Jeff C,
2003-10-17 21:48   ` GCC 5.01p (was: Re: gcc/gnat 3.3) Ludovic Brenta
2003-10-17 22:01     ` Jeff C,
2003-10-20 19:04 ` gcc/gnat 3.3 Andreas Almroth
2003-10-24  9:13   ` Adrian Hoe
2003-10-28  4:19     ` Adrian Hoe
2003-10-28  4:26       ` Jeff C,
2003-10-28 15:36         ` Adrian Hoe
2003-10-29  0:24           ` Jeff C,
2003-11-01  5:51             ` Adrian Hoe
2003-11-01 14:22               ` Jeff C,
2003-11-01 14:28                 ` Jeff C,
2003-11-01 14:52                 ` Ludovic Brenta
2003-11-03  9:36                   ` Adrian Hoe
2003-11-05 18:35                     ` Andreas Almroth
2003-11-06  5:06                       ` Adrian Hoe
2003-11-06  5:20                         ` Adrian Hoe
2003-11-06  6:19                           ` Andreas Almroth
2003-11-06  6:52                             ` Adrian Hoe
2003-11-06  8:54                               ` Andreas Almroth
2003-11-06 11:17                                 ` Adrian Hoe
2003-11-06 13:55                                   ` Andreas Almroth
2003-11-07  4:09                                     ` Adrian Hoe
2003-11-07  6:50                                       ` Andreas Almroth
2003-11-19  8:10                                         ` Adrian Hoe
2003-11-07  4:15                                     ` Adrian Hoe
2003-11-07  7:00                                       ` Andreas Almroth
2003-11-07  7:15                                         ` Adrian Hoe [this message]
2003-11-07 17:05                                           ` Andreas Almroth
2003-11-06  6:33                           ` Adrian Hoe
2003-10-28 14:48       ` Steve
2003-11-01 16:29         ` Andreas Almroth
2003-11-03  3:49           ` Adrian Hoe
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox