Dear Yves,
You wrote:
> What is the size of this map ?
> Maybe RG is running out of stack space...
This should not be an issue. It is just an empty container. I have not put
anything in it...there is not code written in this test to add and contents to
the map.
That is why I'm so perplexed.
But in practicable use, it should stay relatively small, since a key / value
pair is only created if an incomplete ALSA MIDI message is sent (256 bytes is
the standard realtime MIDI chunk size if pulling from a USB or traditional MIDI
device).
System exclusive are pooled on a per device basis, so if somebody has 100 MIDI
devices hooked up, the map could possible have upto 100 key / value pairs in
it. I don't think that is too much for the stack to handle.
When a message is finally complete it no longer stay in the pull and is sent as
a message based on the alsa timestamp of the first message of the series.
Corner case:
If somebody sends a bunk dump --which I'm tests right now with my working code.
The number of elements will not expand but string will grow quite large as the
sys exc. events are pooled.
That could be an issue, but fixing that corner case is beyond the scope of this
bug fix.
...
My testing so seem to indicate success for my patch - but I created the map on
the heap not the stack.
I'm going to try some system dumps next and see how it and KMidiMon handle it.
....
I still have to test it further, and I have to flush the map (even if they are
incomplete) if recording is stopped abruptly. That will go in the punchout()
code. That is simple stuff.
Thanks for your interest.
Sincerely,
Julie S.
------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Rosegarden-devel mailing list
[email protected] - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-devel
