[PD-dev] [info complete $cmd_from_pd] profiling

Hans-Christoph Steiner hans at at.or.at
Thu Aug 20 21:40:26 CEST 2009


I added the [info complete ...] code to the pd-devel/0.41.4 branch.  I  
think that means that I synced all your code changes into this branch,  
and I've added a bunch of new stuff, including translation support,  
fixed up the menubar to support dynamic stuff better (like the window  
menu).

There are some minor remaining differences in code organization, which  
I don't have a strong opinion on.  Basically they seem unresolved, so  
for now I am leaving them as they are in devel.

- devel still has pd_post.tcl with pdtk_post in it
- devel has ::pd_bindings::sendkey instead of pdsend_key

Then I haven't added the audio/midi/startup dialogs or pdwindow yet,  
just a matter of time.  I think those should be dialog_*.tcl like the  
rest of the dialogs.

.hc

On Aug 18, 2009, at 1:06 AM, Miller Puckette wrote:

> Hi Hans -
>
> I didn't profile it -- I had to add that because otherwise my usual  
> patch
> doesn't open.  The trouble is that, when there are LOTS of GUI  
> updates from
> Pd, sometimes the socket 'reblocks' them and something comes in that  
> isn't
> complete.  It's easy to get an incomplete message like "f a b" cut  
> into
> "f a " and then "b", which makes tk puke if "f" needs 2 args. The only
> check I was able to design was to set a convention that any message  
> that
> was broken by a newline had to be enclosed in braces -- so if  
> there's a
> trailing newline and no brace is open, it's OK to interpret it.
>
> I monkeyed with 3 or 4 other things and didn't keep a good log I'm  
> afraid...
> the last thing I hit was for some reason the line:
>
> #    if { $::windowingsystem eq "win32" } {create_system_menu  
> $mymenubar}
>
> in pd_menus.tcl got up my PC's nose.
>
> cheers
> M
>
>
> On Tue, Aug 18, 2009 at 12:00:02AM -0400, Hans-Christoph Steiner  
> wrote:
>>
>> Hey Miller,
>>
>> Just going thru the code you checked in.  I noticed this
>> in ::pd_connect::pd_readsocket,  I am guessing this is the re-add any
>> "\n" that are missing.  I am curious, did you profile it at all?  I
>> was hesitant to do that because I thought it might have a big
>> performance hit.
>>
>> Here's the snippet in question:
>>
>>    while {![info complete $cmd_from_pd] || \
>>        [string index $cmd_from_pd end] != "\n"} {
>>        append cmd_from_pd [read $pd_socket]
>>        if {[eof $pd_socket]} {
>>        close $pd_socket
>>        exit
>>        }
>>    }
>>
>>
>> .hc
>>
>>
>> ----------------------------------------------------------------------------
>>
>> "[T]he greatest purveyor of violence in the world today [is] my own
>> government." - Martin Luther King, Jr.
>>
>>
>>
>>
>> _______________________________________________
>> Pd-dev mailing list
>> Pd-dev at iem.at
>> http://lists.puredata.info/listinfo/pd-dev





----------------------------------------------------------------------------

"[T]he greatest purveyor of violence in the world today [is] my own  
government." - Martin Luther King, Jr.







More information about the Pd-dev mailing list