XS Stash DOES support tied objects... apparently

Well I'm not quite sure what's going on here, but it appears
that the 
XS Stash does indeed have support for tied hashes and
arrays, and from the 
look of it, always has.

The problem is that the hash assignment code was broken in a
subtle way
so as to make it look like it didn't work.

I think I've fixed the bug, so it looks like we now have
full support
for tied objects in the XS Stash.

Yay!

A


_______________________________________________
templates mailing list
templatestemplate-toolkit.org
http://lists.template-toolkit.org/mailman/listinfo/t
emplates

Andy Wardley wrote:
> Well I'm not quite sure what's going on here, but it
appears that the 
> XS Stash does indeed have support for tied hashes and
arrays, and from the 
> look of it, always has.
> 
> The problem is that the hash assignment code was broken
in a subtle way
> so as to make it look like it didn't work.
> 
> I think I've fixed the bug, so it looks like we now
have full support
> for tied objects in the XS Stash.
> 
> Yay!

That's great! Maybe I'll give it a try in the next version.
Don't forget to update the documentation for the module:

http://www.template-toolkit.org/docs/pl
ain/Modules/Template/Stash/XS.html

Was there ever any word on that second "bug" on
that page?

-- Josh

_______________________________________________
templates mailing list
templatestemplate-toolkit.org
http://lists.template-toolkit.org/mailman/listinfo/t
emplates

Andy Wardley wrote:
> I think I've fixed the bug, so it looks like we now
have full support
> for tied objects in the XS Stash.

It turns out my "fixes" introduced a memory leak. 
How do I know?
Because t/leak.t told me so.  

In my defence, I was doing what perlguts told me to do
(increment the
ref count before storing a hash/list item, and decrement it
if the 
hv_store() / av_store() method returns NULL).  But whatever
perlguts
said, the code was wrong.  

Removing the offending ref count updates made t/leak.t
happy, thereby
demonstrating that the XS Stash now does the same thing as
the Perl 
version, but faster.  Job done, I can now go out
skateboarding.  

Needless to say, I can't stress enough how important it is
to have a 
thorough test suite for your software.  Spend all day
debugging, or 
go out skating?  Easy choice  

All Hail the Test Suite!

A


_______________________________________________
templates mailing list
templatestemplate-toolkit.org
http://lists.template-toolkit.org/mailman/listinfo/t
emplates

That's great.  Be sure to let us know when there's a
developer release 
available to test.  I've been dying to have the XS stash
available for 
my work for the longest time, since I was always under the
impression 
that tied variables were broken there.

-Stephen

Andy Wardley wrote:
> Andy Wardley wrote:
> 
>>I think I've fixed the bug, so it looks like we now
have full support
>>for tied objects in the XS Stash.
> 
> 
> It turns out my "fixes" introduced a memory
leak.  How do I know?
> Because t/leak.t told me so.  
> 
> In my defence, I was doing what perlguts told me to do
(increment the
> ref count before storing a hash/list item, and
decrement it if the 
> hv_store() / av_store() method returns NULL).  But
whatever perlguts
> said, the code was wrong.  
> 
> Removing the offending ref count updates made t/leak.t
happy, thereby
> demonstrating that the XS Stash now does the same thing
as the Perl 
> version, but faster.  Job done, I can now go out
skateboarding.  
> 
> Needless to say, I can't stress enough how important it
is to have a 
> thorough test suite for your software.  Spend all day
debugging, or 
> go out skating?  Easy choice  
> 
> All Hail the Test Suite!
> 
> A
> 
> 
> _______________________________________________
> templates mailing list
> templatestemplate-toolkit.org
> http://lists.template-toolkit.org/mailman/listinfo/t
emplates


_______________________________________________
templates mailing list
templatestemplate-toolkit.org
http://lists.template-toolkit.org/mailman/listinfo/t
emplates

Andy,

> Well I'm not quite sure what's going on here, but it
appears that the 
> XS Stash does indeed have support for tied hashes and
arrays, and from the 
> look of it, always has.
> 
> The problem is that the hash assignment code was broken
in a subtle way
> so as to make it look like it didn't work.
> 
> I think I've fixed the bug, so it looks like we now
have full support
> for tied objects in the XS Stash.
> 
> Yay!

Yay indeed!  I'll have to give it another shot whenever you
release the next version.

Thanx for looking into it!


		-- Buddy

_______________________________________________
templates mailing list
templatestemplate-toolkit.org
http://lists.template-toolkit.org/mailman/listinfo/t
emplates