[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: memory management in ocf-files
From: |
Jose |
Subject: |
Re: memory management in ocf-files |
Date: |
Mon, 1 Oct 2012 17:03:44 +0300 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:15.0) Gecko/20120827 Thunderbird/15.0 |
Hello Jordi, and thanks for your quick answer.
On 10/01/2012 04:08 PM, Jordi Gutiérrez Hermoso wrote:
By the way, would you be interested in trying the new instrument
control package?
http://wiki.octave.org/Instrument_control_package#Example_3:_Logic_analyzer
I used the word "sensors" to make the explanations simpler, but in
reality the "sensors" are a mixed system that contains a robot and other
devices, for which the Instrument control package is not suitable. But
thanks for the suggestion.
Octave uses COWs:
http://en.wikipedia.org/wiki/Copy_on_write
If you pass data into an Octave class (e.g. Array, Matrix, any member
of the octave_value hierarchy), then it expects to own the data and do
its own copying and cleanup.
... and this cleanup also cleans the original octave_value objects that
I created in C++ and passed to octave through an oct-file, right? Just
to be sure that I understood you 100%.
What happens with those objects that I create in C++ and are not passed
to octave?
You can also use the mex interface in Octave, but it's just a wrapper
over the Octave classes. Yes, perhaps a wrapper could be slightly
slower, but try using your existing mex code before you decide it's
too slow and you need to rewrite it for Octave.
Do note your GPL obligations regarding Octave:
http://wiki.octave.org/FAQ#Licensing_issues
Just to be sure about your point: in short, you are reminding me that
whatever oct-file we create, it must be released under a GPL-compatible
license, right?
If so, thanks for the reminder. In any case, we have no problems with that.
BR
Jose