[Openmcl-devel] new 0.14 binaries (finally!)

[Ram Krishnan]

On Tuesday, July 8, 2003, at 07:36 PM, Gary Byers wrote:

> On Tue, 8 Jul 2003, John DeSoi wrote:
>
>>
>>> I'm not sure how mature or well-tested those native-thread changes  
>>> are;
>>> perhaps (if the current CVS paserve is indeed running on multiple
>>> platforms again) the OpenMCL-0.14 testing-and-example-of-how-to-use-
>>> native-threads version should be derived from the current CVS paserve
>>> instead of that year+ old release.
>>
>> Yes, I was not aware of this example and I have not seen any patches
>> for it in CVS. I knew Eric was not running the CVS version because of
>> the directory paths. If some one will provide me a pointer to the
>> paserve/native threads example, I'll incorporate them in the next  
>> round
>> of changes.
>>
>
> Diffs relative to portableaserve-1.2.12c are in:
>
> <ftp://clozure.com/pub/testing/portableaserve-1.2.12c-openmcl-0.14- 
> 030415.diff>
>
> A distribution with those patches applied is in:
>
> <ftp://clozure.com/pub/testing/portableaserve-openmcl-0.14- 
> 030415.tar.gz>
>
> The bizarre error that Eric reported was new as of yesterday; the  
> problem
> with the redefinition of IP-ADDR-TO-DOTTED (IIRC) has been there for a
> while, so I'm not sure how widely tested this has been.  (I pounded on
> it fairly hard for a few days last spring, but I was doing a limited
> set of things.)

I've been loading portableaserve with the following expression for a  
while now:

(let ((*warn-if-redefine-kernel* nil))
   (setf (logical-pathname-translations "aserve")
         `(("**;*.*.*"  
"/src/cl-lib/portableaserve-openmcl-0.14/aserve/**/"))
         (logical-pathname-translations "acl-compat")
         `(("**;*.*.*"  
"/src/cl-lib/portableaserve-openmcl-0.14/acl-compat/**/")))
   (load "acl-compat:acl-compat-mcl.system")
   (load "aserve:aserve-mcl.system")
   (mk:load-system :acl-compat)
   (mk:load-system :aserve))

Does your check-in remove the need for the *warn-if-redefine-kernel*  
binding?

> In version 1.2.12c, there are changes in the portability layer and in
> the mainline code: the main server listen/accept function uses  
> semaphores
> and other constructs (instead of PROCESS-ADD-RUN-REASON, etc.) to
> synchronize with the client threads.  I -think- that it's fairly  
> reasonable
> and understandable; I confess that I was more interested in making an
> example of how to do thread synchronization in 0.14 than in keeping  
> paserve
> maintainable ...

Just so you know your effort, above and beyond duty, is greatly  
appreciated, Thank you.

-ram


_______________________________________________
Openmcl-devel mailing list
Openmcl-devel at clozure.com
http://clozure.com/cgi-bin/mailman/listinfo/openmcl-devel