Hi.
I am working on a patch that should develop into a tool that will be used
by (hopefully many) instrumentalists/conductors, specially non-pd users.
While working on it, I thought that if a non-technical user runs into an
error, then he has always to look at the console, which is, for some kind
of people, kind of asking too much. (it might already be complicated for
some not to present a single-file standalone to work with)
Which reminded me: wouldn't it be useful to have an object that could
capture the console contents? The main purpose could be to get the error
messages (that only display in the console, I think no object sends a bang
or something on error). Then this information could be treated in a way to
make the patch more user-friendly (f.e. display a designed error message,
and consequently how to correct it).
Does this interest anyone? Does anyone want to have a go at it? Afaik
there's no object that can do this.
João Pais
João Pais wrote:
Hi.
Does this interest anyone? Does anyone want to have a go at it? Afaik
there's no object that can do this.
there have been numerous attempts and solutions to this problem. including using netsend to pipe the stderr back into Pd, adding hooks to register new post()-backends (which could be the message-output of an object),... the latter requires a patched Pd-version (the patch exists in the patchtracker, afaik) but miller never fully accepted it into the main source-tree (though part of it: the embedded pd-console is part of it; and the ability to choose between stderr and pd-console) the former does not require such, but is less efficient (but then: it also grabs errors that are not "post"ed but thrown e.g. by ioctl's)
just search the list archives for "capture console output" :-)
mfgasdr IOhannes
would it be possible with the current pd version to write an additional object like the max [error] that outputs all error messages on an outlet when turned on (message 1)? or does this also need the hook you were talking about. marius.
IOhannes m zmoelnig wrote:
João Pais wrote:
Hi.
Does this interest anyone? Does anyone want to have a go at it? Afaik
there's no object that can do this.there have been numerous attempts and solutions to this problem. including using netsend to pipe the stderr back into Pd, adding hooks to register new post()-backends (which could be the message-output of an object),... the latter requires a patched Pd-version (the patch exists in the patchtracker, afaik) but miller never fully accepted it into the main source-tree (though part of it: the embedded pd-console is part of it; and the ability to choose between stderr and pd-console) the former does not require such, but is less efficient (but then: it also grabs errors that are not "post"ed but thrown e.g. by ioctl's)
just search the list archives for "capture console output" :-)
mfgasdr IOhannes
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
You could probably hack some tcl code and send it to [sys_gui] and
have that send the .printout text to a [receive].
It would be a good project for getting to know the internals of Pd's
GUI. .printout is the console part of the Pd window. .printout.text
is the text widget that displays the text. Here's the Tcl/Tk docs on
the text widget:
http://tcl.tk/man/tcl8.4/TkCmd/text.htm
pdtk_post is the procedure that actually posts text to
the .printout.text. You could probably redefine pdtk_post to instead
send the data to a receive, something like this:
pdtk_post {stuff} { }
Hmm, I guess you can't send that to [sys_gui] since Pd can't do
{ and }. Another way would be to add "pd [concat pd error $stuff]"
to pdtk_post and then add a little bit of C code to get the data from
messages send to:
[route pd] | [route error]
.hc
On May 20, 2008, at 3:28 PM, marius schebella wrote:
would it be possible with the current pd version to write an
additional object like the max [error] that outputs all error messages on an
outlet when turned on (message 1)? or does this also need the hook you were talking about. marius.IOhannes m zmoelnig wrote:
João Pais wrote:
Hi.
Does this interest anyone? Does anyone want to have a go at it?
Afaik there's no object that can do this.there have been numerous attempts and solutions to this problem. including using netsend to pipe the stderr back into Pd, adding
hooks to register new post()-backends (which could be the message-output of an object),... the latter requires a patched Pd-version (the patch exists in the patchtracker, afaik) but miller never fully accepted it into the main source-tree (though part of it: the embedded pd-console is part of
it; and the ability to choose between stderr and pd-console) the former does not require such, but is less efficient (but then: it also grabs errors that are not "post"ed but thrown e.g. by ioctl's)just search the list archives for "capture console output" :-)
mfgasdr IOhannes
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/ listinfo/pd-list
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/ listinfo/pd-list
I spent 33 years and four months in active military service and
during that period I spent most of my time as a high class muscle man
for Big Business, for Wall Street and the bankers. - General
Smedley Butler
-
Hans-Christoph Steiner -
IOhannes m zmoelnig -
João Pais -
marius schebella