W3C home > Mailing lists > Public > xproc-dev@w3.org > September 2013

Re: Calabash, thread-local variables and memory leaks

From: Florent Georges <fgeorges@fgeorges.org>
Date: Fri, 20 Sep 2013 03:44:04 +0200
Message-ID: <CADyR_r08JWps1viOj5joDmNr6WyOr4_0vuMh503Ss18LAR4R-A@mail.gmail.com>
To: Norman Walsh <ndw@nwalsh.com>
Cc: XProc Dev <xproc-dev@w3.org>
  Norm,

  Just to confirm that after some test about what we discussed in
private, you were right actually.  I did fail to call close() on
XProcRuntime, and that indeed solved the problem.

  Sorry for the noise.  Thanks for your help!

-- 
Florent Georges
http://fgeorges.org/
http://h2oconsulting.be/


On 19 September 2013 16:55, Norman Walsh wrote:
> Florent Georges <fgeorges@fgeorges.org> writes:
>>   Because this is Calabash internals, I can't release them myself.
>> I am not sure how we can solve this one.  Maybe by adding a method
>> to release (or close) the runtime object itself, that throws away
>> everything it does not need anymore, recursively, including releasing
>> the thread-local variables?
>
> What version of XML Calabash are you using? There was a bug where the
> variable wasn't being freed, but I fixed that a while ago.
>
>>   Or maybe simply by removing the thread-local variable?  I am not
>> sure what it is used for.
>
> The thread local variable is used to communicate between the runtime
> object and the implementation of the function.
>
> The problem is:
>
> 1. The class initializers have no mechanism for passing in the XProcRuntime.
> 2. The instances that are evaluated are constructed by cloning the objects.
> 3. The implementation of the functions has to be able to access the XProcRuntime.
>
>                                         Be seeing you,
>                                           norm
>
> --
> Norman Walsh
> Lead Engineer
> MarkLogic Corporation
> Phone: +1 512 761 6676
> www.marklogic.com
Received on Friday, 20 September 2013 01:44:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:03:11 UTC