[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <51102D65.70709@gmail.com>
Date: Mon, 04 Feb 2013 19:51:33 -0200
From: Claudio André <claudioandre.br@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: OpenCL problems
Em 04-02-2013 19:31, magnum escreveu:
> What happens without that show_usage cludge? Does it fail to give proper usage output? Ah no, I think I see now: this way it wont even list the formats unless we have a proper device, right?
>
> I think we should think about this before doing too much. Maybe we should list the formats anyway. Or maybe not. Argh.
>
> The simplest might be to drop this idea and just do like unstable: As long as you don't use an OpenCL format, it's fine not having any working platform and/or device. Only when you do, you'll get a device/platform error.
>
> magnum
The best solution would be move (at least part of) opt_init() routine
and make it follow status_init();
If we don't (this is unstable way):
- use bleeding as is now (compiled with OpenCL), will always show
*-opencl formats in --help.
- use bleeding as is now (compiled with OpenCL) in a machine without any
devices will show an error if you try any *-opencl format. Hey, luser
saw this *-format using --help.
- ordinary formats will work as expected.
Claudio
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
