On 10 May 2013 12:18, Richard Quadling <rquadling@xxxxxxxxx> wrote: > > On 10 May 2013 12:08, Nick Whiting <nwhiting@xxxxxxxxxxxxxxx> wrote: > >> >> Do you have a backtace for this? >> >> What is the gender class doing? >> >> Have u done a global search for keyword "clone"? >> >> >> >> On Friday, May 10, 2013, Richard Quadling wrote: >> >>> Hi. >>> >>> I'm having an issue where I get ... >>> >>> Fatal error: Trying to clone an uncloneable object of class >>> Smarty_Variable >>> in >>> >>> xxxxxx/trunk.newbuild/includes/frontend/site_includes/classes/smarty-3.10/sysplugins/smarty_internal_template.php >>> on line 269 >>> >>> This issue happens consistently on our live server and on our test >>> server, >>> but not on my dev setup. >>> >>> The issue doesn't happen if I comment out 1 line of completely unrelated >>> code ... >>> >>> $o_Gender = new Gender\Gender; >>> >>> If I immediately follow that with ... >>> >>> unset($o_Gender); >>> >>> and have no other access to $o_Gender, I still get the error. >>> >>> Comment out the code, no problems. >>> >>> The extension is used in other parts of the system with seemingly no >>> problem. That is, the code behaves as expected and I get no errors, but >>> those elements don't use Smarty. The error being reported is clearly >>> wrong. >>> And the extension (as far as I can see) has no interaction with global >>> elements in any way (I have to use the Gender namespace to access >>> anything >>> in it - which I think is correct). I've var_dump()'d a debug_backtrace() >>> at >>> the point of failure in the Smarty code, with and without the $o_Gender >>> variable being defined (it isn't used in the Smarty template - so Smarty >>> is >>> never touching it). When I compare the 2 dumps, the only differences is >>> in >>> the datetime stamp elements and the object count values (there's 1 more >>> when $o_Gender exists). >>> >>> My setup is on a CentOS VM running PHP V5.4.14 >>> >>> The live setup is on a remote CentOS server running PHP V5.3.21 >>> The test server is on CentOS server running PHP V5.3.3 >>> >>> I don't know CentOS well enough to just swap out a new version of PHP. >>> But >>> I will be getting some help on that. >>> >>> >>> Where do I start to find the problem? >>> >>> I have full root access to the command line test server, so I can, within >>> reason, follow instructions to run/wrap the code in any way needed. >>> >>> Any help would be GREATLY appreciated!!! >>> >>> Thanks in advance, >>> >>> Richard. >> >> > The Gender class is from the pecl/Gender extension. > > As for a backtrace, the code is in Smarty and works fine if I don't have > pecl/Gender instantiated. > > I'm guessing this is really an internals issue. The code operates on V5.3 > and V5.4 without the extension with no issue. Put the extension in and only > on one page (so far) do we get the issue. > The clone error is incorrect as the object can be cloned, and is, very successfully, just not when I've created an instance of gender. I'm thinking there's a memory issue, but I don't know how to diagnose it. -- Richard Quadling Twitter : @RQuadling EE : http://e-e.com/M_248814.html Zend : http://bit.ly/9O8vFY
