Daniel Kasak writes:
>
> $self->{drawing_area}->set_events(
Gtk 2 is ->add_events() usually from INIT_INSTANCE to have extra events
selected when realized, if that's still so.
___
gtk-perl-list mailing list
Apparently perl 23.something pre-releases are threatening to make xs
EXTEND() of negative amounts an error. I have this from cpantesters of
some of my code per https://rt.cpan.org/Ticket/Display.html?id=108274
but I don't have a bleeding perl to actually try.
If I'm right it occurs when
Daniel Kasak d.j.kasak...@gmail.com writes:
However, the other 2 treeviews never scroll. Why not?
If the target adjustments you change are the right ones then you'd think
it ought to work.
As a bit of shameless self-promotion, I made Glib::Ex::ConnectProperties
for tying property values
Torsten Schoenfeld kaffeeti...@gmx.de writes:
If you do, however, call $label-parent, then the marshalling code sees
that the underlying GObject has a floating reference and takes over
ownership of it (without adding a new reference). The GObject now has
two references to it, one from $self
I wrote:
Yes, undoing that fixes it. Dunno why.
How has this bit gone? It definitely seems to do something unexpectedly
bad. How would you feel about reverting until the reason can be found?
(Debian has updated to 1.302 and so I've been seeing it ... :-)
--
No, eees hamster.
Torsten Schoenfeld kaffeeti...@gmx.de writes:
Can you verify that (perhaps with git bisect)?
Yes, undoing that fixes it. Dunno why.
___
gtk-perl-list mailing list
gtk-perl-list@gnome.org
https://mail.gnome.org/mailman/listinfo/gtk-perl-list
With recent debian i386 glib 2.36.3 and gtk 2.24.20 the foo.pl below
dies with
label Gtk2::Label=HASH(0x9fe6cf8)
label now packed
parent MyContainer=HASH(0x9e5d770)
return from INIT_INSTANCE
GLib-GObject-CRITICAL **: g_object_notify_queue_thaw: assertion
Brian Manning c...@xaoc.org writes:
+I18N::Langinfo-import(qw(langinfo CODESET));
You could call it without importing if desired. I've used
eval {
$charset = I18N::Langinfo::langinfo(I18N::Langinfo::CODESET());
};
___
gtk-perl-list
I started a ticket for passing args through unchanged.
I might have a go at making register_object() more flexible too.
https://bugzilla.gnome.org/show_bug.cgi?id=690459
___
gtk-perl-list mailing list
gtk-perl-list@gnome.org
I wrote:
Maybe Glib::Object::New if a new() is it's only purpose.
I started a ticket for this. I think it's a reasonably sound idea.
Maybe the docs could be polished a bit to justify it or make it clearer
though.
https://bugzilla.gnome.org/show_bug.cgi?id=690461
--
I'm waiting for my
I wrote:
Oh, that might be one of mine. I see Glib::Log pod is
Glib-set_default_handler() where it should be
Glib::Log-set_default_handler().
It took a long time, but
https://bugzilla.gnome.org/show_bug.cgi?id=690462
--
Even the white bits were black.
Torsten Schoenfeld kaffeeti...@gmx.de writes:
Note the roundtrip from @_ to %arg to a list.
Ah yes. Perhaps it would save a couple of nanoseconds to pass them
straight through (untested),
shift; # our own name 'Glib::Subclass::Object'
my $superclass = shift;
my $class = caller;
I've had some cpantesters test failures from my Gtk2-Ex-WidgetBits
nonsense
http://www.cpantesters.org/cpan/report/7e29e62e-3156-11e2-b1c8-a5d7a290f8f5
with something about adding Gtk2::Buildable to a subclass
GLib-GObject-WARNING **: cannot add interface type `GtkBuildable' to
type
Domovoy domo...@errlock.org writes:
Glib::Log-set_default_handler
Oh, that might be one of mine. I see Glib::Log pod is
Glib-set_default_handler() where it should be
Glib::Log-set_default_handler().
(The code and .t tests look right, just the =for signature which becomes
the docs ...)
Domovoy domo...@errlock.org writes:
Glib::Type-register_object
That's the fundamental part. Glib::Object::Subclass is only a front end
to it. You can register_object() yourself if desired.
our %signals = (
our %properties = (
That info just disappears into glib normally, no
Domovoy domo...@errlock.org writes:
So, i finally took a look at Glib::Object::Subclass.pm, and saw that it
was doing its job by defining an import sub.
And putting itself at the start of the caller's @ISA ...
use Glib::Object::Subclass
Gtk2::ScrolledWindow::;
Jeffrey Ratcliffe jeffrey.ratcli...@gmail.com writes:
It seems that in subclassed a widget, any properties assigned in
-new() are not available in INIT_INSTANCE.
They reach you in a SET_PROPERTY call, after the init. But perhaps you
already knew that. It's quite desirable to allow property
Aaron Lewis alewis1...@yahoo.com writes:
I don't know how to detect the combination of mouse click +
SHIFT/CTRL.
$event-state gives the GdkModifierType keys (per Gtk2::Gdk::Event).
3. I also want the user to be able to select multiple rectangles by
dragging the mouse across a portion of
Jeffrey Ratcliffe jeffrey.ratcli...@gmail.com writes:
Trapping the delete and destroy events to hide the window works, as
does allowing destroy to destroy, but I can't get the delete event to
destroy.
It seems a little better for me to $widget-destroy in the delete
instead of chaining up, but
Torsten Schoenfeld kaffeeti...@gmx.de writes:
• Do we continue to use SvPV/newSVpv but also store the UTF8 flag, and
if it was set, restore it?
Sounds likely, in as much as it's round-trip clean. What does posix say
about the encoding of argv? Text would be in the locale charset
wouldn't it?
Torsten Schoenfeld kaffeeti...@gmx.de writes:
Yeah, just Gtk3::RadioMenuItem.
Do GtkRadioButton and GtkRadioToolButton have something similar?
Maybe gtk_radio_action_set_group() too, or perhaps that was never
documented. Undef makes some sense if the C func takes NULL ...
Torsten Schoenfeld kaffeeti...@gmx.de writes:
On 25.06.2012 14:43, Dave M wrote:
my $group = undef;
Try '$group = []'. I think this is also how Gtk2 works.
Both undef and [] in the gtk2, I believe.
___
gtk-perl-list mailing list
Torsten Schoenfeld kaffeeti...@gmx.de writes:
So I suspect it's a glib bug.
Yep.
The only possibly related difference I can think of might be whether
gobject uses the generic libffi-based signal marshaller.
Hmm. The debian 2.30.2 has a depend on its libffi5, and shows up in ldd
zentara zzmilosc...@gmail.com writes:
$label-signal_connect (populate_popup = sub {
my (undef, $menu) = @_;
my $item = Gtk2::MenuItem-new ('Whee');
$menu-append ($item);
$item-show;
Yep. For one of mine I put a new entry at the start instead of the end.
Depends how
The foo.pl below has trouble in a notify closure reached from a
container add closure.
The notify closure prints its invocation_hint and the second one
suggests glib thinks it's still in signal add, not signal notify, so
the marshalling for signal_chain_from_overridden() gets an arg type
Torsten Schoenfeld kaffeeti...@gmx.de writes:
This smells a lot like some circular reference issue,
I spot one at
$r-signal_connect
(edited = \text_cell_edited, $view);
under $column_info[$i]{attr} eq 'text'. Is that a hard ref from the
renderer back to its containing
Steve Wills swi...@freebsd.org writes:
Basically, make test hangs at t/9.t when it reaches test 26.
Within the first Glib::Child-watch_add is it?
After
looking at it, it seems to be due to the call to g_child_watch_add_full
eventually calling g_thread_create_posix_impl which uses pthreads
Martin Schlemmer martin.schlem...@nwu.ac.za writes:
On 2011/11/20 at 01:34 AM, Torsten Schoenfeld kaffeeti...@gmx.de wrote:
Or you could just always go through get()/set() instead of direct hash
access.
I do not mind either way, just wanted to verify that it was due to
some reason, and
YangXi jianding...@msn.com writes:
Even if we have to bless the referenCE, does the blessing actually
bind package subs on the referenT?
I think the blessing is a property of the underlying hash (or whatever).
Eg. couple of lines below.
It seems I should create
an SV reference for my HV,
Torsten Schoenfeld kaffeeti...@gmx.de writes:
I wanted to link to some XS tutorial covering all this, but the
example at http://perldoc.perl.org/perlxstut.html#EXAMPLE-6 actually
uses a different pattern to achieve the same thing: it mortalizes both
the referent and the reference,
I suppose
eexpress eex...@gmail.com writes:
i write a very simple gtk2-perl script, want draw on root window. i
test a lot of methods, but still can not get expose event
handler.
I'm not sure if expose events are available for non-widget windows.
They might disappear down the
Torsten Schoenfeld kaffeeti...@gmx.de writes:
Gtk2::Action::get_proxies returns widgets which have only a single,
floating ref on them.
I could never tell if the connect-proxies signal was doing the right
thing in letting get_proxies() look like that. I suppose together with
uimanager it's
Terrence Brannon scheme...@gmail.com writes:
t/a.t .. 1/30
# Failed test 'default log handler: install undef, prev default'
# at t/a.t line 78.
# got: undef
# expected: 'CODE(0x27cdfdc)'
# Looks like you failed 1 test of 30.
Hmm. That might be one of
Roderich Schupp roderich.sch...@googlemail.com writes:
Maybe gperl_get_object and gperl_get_object_check need
SvGETMAGIC(sv);
*before* checking SvROK(sv)?
Yes, I think so. Always seems a bit fraught enquiring into the nature
of an sv with those OK or TYPE macros.
If I intentionally
Roderich Schupp roderich.sch...@googlemail.com writes:
That's why assigning the value to a dummy variable works: it clears
the tie magic.
Clears as in runs the magic I suppose, so as to reach the ixhash
FETCH(). Does that mean gperl_get_object() and friends ought to
SvGETMAGIC() or something
Torsten Schoenfeld kaffeeti...@gmx.de writes:
preloading of subs is a performance issue,
Maybe a class and its parents could be fully populated only when an
object of that type becomes visible at the perl level or is created from
perl. That might mean separate use Gtk2::Button etc for every
Duncan Ferguson duncan_j_fergu...@yahoo.co.uk writes:
- run a piece of perl code to get the window ID before exec'ing ssh
in the window with the correct params to connect to the remote host
xterm sets a WINDOWID environment variable, if that's not already what
you're doing.
Tk key event
Salvatore De Paolis iwkse...@gmx.com writes:
The only issue is, the image i'm going to set is an animated gif and the
animation is not visible but just the static picture.
That'll be doomed if you've gone away blocked doing something else.
Another way is to change the mouse to a watch cursor
Emmanuel Rodriguez emmanuel.rodrig...@gmail.com writes:
xvfb-run --server-args=-screen 0 1024x768x24 $@
That'd be about right.
Gtk3::OffscreenWindow instead of Gtk2::Window
Shouldn't need to change the program. The only thing that xvfb won't
have is a window manager, if some sloppy code is
Torsten Schoenfeld kaffeeti...@gmx.de writes:
I'll try to roll a release on Thursday.
There can't be much wrong with the cvs head either can there?
Except one or two patches waiting to be applied ...
___
gtk-perl-list mailing list
Dave M dave.n...@gmail.com writes:
I'll leave it alone then.
I think the rc priority level can let user or theme override,
supposedly. Level :application is probably right for applications, or
level :gtk for implementing a widget subclass.
Gtk2::Rc-parse_string ('HERE');
style
François Rappaz francois.rap...@unifr.ch writes:
Gtk2::Widget::reparent($vbox, $sfctrl);
Would that be better as a method $vbox-reparent($sfctrl) ?
___
gtk-perl-list mailing list
gtk-perl-list@gnome.org
François Rappaz francois.rap...@unifr.ch writes:
Is it possible to include the child window on the fly ?
Probably not easily. A given widget can only appear in one place at the
one time, though you can reparent to shift it about.
Is this possible without having to copy all the fields and
Konovalov, Vadim (Vadim)** CTR ** vadim.konova...@alcatel-lucent.com writes:
Is this behaviour introduced for Gtk3, or what version of perl-Gtk this was
introduced?
The Gtk2 bindings create all xsubs of all classes at startup (that's
right isn't it?).
I am very in favour of light binding
Attila bardi attila.ba...@gmail.com writes:
Then how can I set the window size and remove the right from the UI user to
resize it?
Maybe Gtk2::Window set_geometry_hints() saying the minimum and maximum
sizes the same value.
Or Gtk2::Gdk::Window set_functions() to take away the resize handles,
Jeffrey Ratcliffe jeffrey.ratcli...@gmail.com writes:
How do you test a function that is not in a module?
All I meant was to run some private non-gui code like
is (Gtk2::Ex::MyWidget::_crunch_something('abc),
'def');
directly out of what's otherwise a widget module, if you don't
Torsten Schoenfeld kaffeeti...@gmx.de writes:
use Glib::Object::Introspection;
Glib::Object::Introspection-setup (
basename = 'Gtk',
version = '3.0',
package = 'Gtk3');
Does that load up pretty much all classes and function methods etc?
Seeing gtk2 grow and grow made me
Mario Kemper mario.kem...@googlemail.com writes:
This means that we don't have to change our code, right?
Just make a release double quick :-).
Or use the Gtk2::Gdk-keyval_from_name(). The full list of symbols from
keys %... may be the only thing Gtk2::Gdk::Keysyms does better than
Jouko Salo jouko.s...@gmail.com writes:
*** Can't locate object method get_path via package
Gtk2::TreeRowReference=SCALAR(0x975aa58) (perhaps you forgot to load
Gtk2::TreeRowReference=SCALAR(0x975aa58)?)
Sounds strange, but a rowreference is a boxed and strange things tend
to happen there.
Jeffrey Ratcliffe jeffrey.ratcli...@gmail.com writes:
Has anyone tried to use these (or a different tool) for testing a
Gtk2-Perl app?
I haven't gone beyond calling module methods or activating Gtk2::Action
or buttons etc. Getting enough state and waiting and whatnot for even
that tends to be
Jeffrey Ratcliffe jeffrey.ratcli...@gmail.com writes:
put everything that doesn't involve the
GUI into module(s)
Not necessarily split into modules, but in functions which can be
exercised without too much gui. Creating some widgets to support
shouldn't be a problem. Test .t scripts may have
Mario Kemper mario.kem...@googlemail.com writes:
how do I convert rgba information into a pixbuf? I have an array with
rgba values (0-255) for each pixel. How do I need to pack those
information in order to use Gtk2::Gdk::Pixbuf-new_from_data?
Bytes in a string, eg. \377\0\0 for a red pixel.
Jeremy Volkening volken...@mailshack.com writes:
If I first capture the pointer value for the page within the signal
handler and then use it in the signal_emit function, it works fine, but
this is a rather convoluted way to go about things.
I expect a class handler could pass it upwards to
Mario Kemper mario.kem...@googlemail.com writes:
What else do we - from a gtk2-perl perspective - have to take into
account?
Some sort of gdk_list_keysyms() func could have %Gtk2::Gdk::Keysyms
built from the library at runtime instead of grepped out of the sources.
It could even be a tied hash
Mario Kemper mario.kem...@googlemail.com writes:
if ($event-keyval == $Gtk2::Gdk::Keysyms{Escape}) {
I imagine that can stay. You could also do though
... == Gtk2::Gdk-keyval_from_name('Escape')
if that seemed like less loading, or a workaround, or whatever.
(It's also one of those
Mario Kemper mario.kem...@googlemail.com writes:
'KEY_VoidSymbol' = 0xff,
I suppose the latest incompatible change in gtk 2.22 for GDK_KEY_ has
leaked out. Torsten made a change in the cvs. (Has it been released as
such?)
___
gtk-perl-list
Jens Luedicke jens.luedi...@gmail.com writes:
$button-signal_connect(clicked = bind(\*MyWindow::callback, $self,
_1, _2, _3));
One thing something like that might do more than closures would be to
weaken the $self target held onto so as to avoid circular references.
Weakening ends up being
Torsten Schoenfeld kaffeeti...@gmx.de writes:
GdkCursor seems to simply assume that the GdkDisplay is valid
throughout the cursor's lifetime, but it doesn't take a reference on
it.
Ah yeah.
I don't think it's the bindings' responsibility to do something
about this.
What about a few words
I wrote:
wrapping callback funcs
Maybe CellLayout and TreeSortable could use the CvXSUBANY() thing which
is the ix / INTERFACE_MACRO: stuff. That's designed for xsub code to
have a fixed C func and extra info to use when it runs, is it? Which
would take those two places out of the realm of
Florian Ragwitz r...@debian.org writes:
data_sv = newSViv (PTR2IV (wrapper));
- sv_magic ((SV *) dummy_cv, 0, PERL_MAGIC_ext, (const char *) data_sv,
0);
+ gperl_attach_mg ((SV *) dummy_cv, data_sv);
Incidentally, that data_sv looks like it's leaked on DESTROY -- with or
The program below gets a segv for me with i386 debian experimental gtk
2.22 and glib 2.27.3.
It's a GdkDisplay with a cursor created on that display, and the display
destroyed before the cursor. Looks like the segv is at the undef $c
stage, with gdb per below.
I wonder if gdk_cursor_unref() is
Florian Ragwitz r...@debian.org writes:
more robust and improve interoperability
Yes.
extensions also using the PERL_MAGIC_ext mechanism.
The magic is only the underlying hash is it, what sort of things might
add further magic to that? Something fiddling around with what hash
keys are
Florian Ragwitz r...@debian.org writes:
+void
+gperl_attach_mg (SV * sv, void * ptr)
+{
+ sv_magicext (sv, NULL, PERL_MAGIC_ext, gperl_mg_vtbl,
+ (const char *)ptr, 0);
I suppose the alternative to magic would be a pointer-pointer hash
table would it? Would that be
Florian Ragwitz r...@debian.org writes:
If anyone other than us wants to use this functionality, he's free
to use XS::Object::Magic. The goal here is not to provide a general
purpose API - that already exists.
I suppose friends or subclasses of gobject/gboxed might use it
Florian Ragwitz r...@debian.org writes:
I wondered that too. I briefly looked over how it's used and it turned
out to be used as a return-by-reference value in some glib callback.
Looks like it's only an IV passed out to the perl, and the newSVsv
copies to pass there anyway.
newCONSTSUB
On
Florian Ragwitz r...@debian.org writes:
do something special when $this gets freed ...
The ones I think I've seen seem to be mainly for scalars rather than
hashes like the glib::object thingie.
Does a tied hash go through magic? If someone was crazy enough to want
to tie their gobject hash
Florian Ragwitz r...@debian.org writes:
Most of the time glib related modules want to use the gobject/gboxed
interfaces anyway and don't really care about associating plain pointers
with scalars.
Except those whacky bits like wrapping callback funcs for CellLayout or
TreeSortable to expose
Torsten Schoenfeld kaffeeti...@gmx.de writes:
It's now wrapped in git master.
I didn't realize set_option() won't replace existing option settings.
Puts a dampener on it for general storage or amending options after
loading :(
Might like to know the option keys that have been loaded too, but I
Dave M dave.n...@gmail.com writes:
[ 'FileMenu', undef, '_File' ],
I've lately liked the keyword style
{ name = 'FileMenu', label = '_File' }
making it harder to get the columns wrong :-)
(There's a bug ticket for some pod of both styles for radio actions,
others similar.)
Marius Feraru altb...@gmail.com writes:
On Wed, Nov 24, 2010 at 13:35, haratron harat...@gmail.com wrote:
They don't show up, despite having their visible property set to 1.
Just by manipulating this property won't get you too far.
Ie. that -visible() one is just the widget flags bits
haratron harat...@gmail.com writes:
$window-set_border_width(0);
$window-set_position('center_always');
They could be property settings in the builder xml I expect.
$window-signal_connect( destroy = sub {
Gtk2-main_quit;
});
And that too as
signal name=destroy
Marius Feraru altb...@gmail.com writes:
My apologies for the off topic message, but Kevin's reply made me curious:
Weirdness in the bindings is on topic isn't it? :-)
Is „visibile” a special case or is this true for all properties?
All the flags bits ones, but not otherwise I don't think.
Thomas Funk t.f...@web.de writes:
So I guess the light higlighted row is an old display state.
It's not just the cursor thingie is it? Maybe set_cursor() and
scroll_to_cell() to start on the initial choice?
___
gtk-perl-list mailing list
Thomas Funk t.f...@web.de writes:
but it ends in an endless loop with a seg fault :-(
You'll have to stick it under the debugger, or post a complete
program ...
In which module can I find scroll_to_cell() ?
Another Gtk2::TreeView method.
--
Nobody knows nothing.
-- all you need to
Thomas Funk t.f...@web.de writes:
my $treeselection = set_treeview_row($treeview_object, Theme Manager);
$start_init = FALSE;
}
Oh, a set_cursor() under a cursor-changed might immediately recurse to
run cursor-changed again. Perhaps clear the $start_init before doing
the
Zettai Muri zettaim...@gmail.com writes:
I am currently trying to
change the mouse cursor to a hand on a 'button_press_event'.
As a bit of shameless self-promotion, my Gtk2::Ex::WidgetCursor is a bit
of oop-ery which can help making temporary cursor settings. It may be
more than you need
Mike Martin redt...@gmail.com writes:
Gtk-CRITICAL **: gtk_list_store_get_value: assertion `VALID_ITER
(iter, list_store)' failed at /usr/lib/perl5/Gtk2/SimpleList.pm line
248.
Oh, a TiedRow hangs onto an iter, so if the row it refers to has been
removed then that may be the error you get
Peter Juhasz peter.juhas...@gmail.com writes:
gtk-dialog-get-widget-for-response() function,
Yep, new in gtk 2.20, not wrapped yet.
Is there an other way, perhaps with $dialog-get_action_area?
I've had some joy from $dialog-get_action_area-get_children and
looking at
Mike Martin redt...@gmail.com writes:
if ( $cell-get(model)-get( $combo_iter, 1 ) eq $val ) {
If you know what columns are what then that's about it. There's also
$model-foreach (per the Gtk2::TreeModel docs) if you want someone else
to walk the iters for you (good for sub-rows if a
Torsten Schoenfeld kaffeeti...@gmx.de writes:
gdk_pixbuf_set_option
I though of using that too, so as to hold xpm x_hot and y_hot values
in the same place they'd come back from the loader. Couple of lines
below.
not part of the public API, but rather only available to backends
Hard to tell
Torsten Schoenfeld kaffeeti...@gmx.de writes:
built-in introspection mechanisms to construct the thing.
Ah yes, the g_module lookup thingie loading up classes out of the shared
libraries. It's not a security hole, it's a feature. :-)
Implementing GtkBuildable is only necessary if you want to
Torsten Schoenfeld kaffeeti...@gmx.de writes:
What if gtk_text_view_new_with_buffer some day changes such that it is
not equivalent to this anymore?
That'd be the danger, if new_with_buffer doesn't document what it
actually does. The gtk_menu_item_new_with_label is probably an example
of
Mike Martin redt...@gmail.com writes:
-{table}-child_get_property($w,'left-attach');
Yes.
I cant find any way to get the X11 name from GDK,
$win-property_get, but it's a little icky :). I forget what you're
supposed to pass for the length parameter. I limit on how huge a chunk
you want to
Dave Howorth dhowo...@mrc-lmb.cam.ac.uk writes:
# Process the events so the show actually does something
Gtk2-main_iteration while Gtk2-events_pending;
Beware of a race condition there. There won't be any drawing until
events get back from the server, and that won't be immediate, especially
Nicolas Courtel cour...@cena.fr writes:
the popup shows up empty, and
I can't figure out what I've done wrong.
It's only drawn when a map/expose/etc comes back from the server, by
which time you're off doing other things.
A window with a background pixmap can show without any action from the
The Gtk2::Buildable pod PLAIN OBJECTS reads like you can't build a
plain Glib::Object subclass at all unless you have the Gtk2::Buildable
interface on it. I suppose that's adapted from the gtk ref manual
GtkBuildable Description,
In order to allow construction from a GtkBuilder UI
If an object property is a Glib::Scalar, ie. Glib::ParamSpec-scalar,
can it be set with the property thingie in Gtk2::Builder ? The few
lines below get
Gtk-WARNING **: Failed to set property Foo.myscalar to hello: Could not
parse 'hello' as a GPerlSV at /tmp/scalar.pl line 21.
and
When a Gtk2::Action is activated, ie. the activate method which raises
the activate signal, is there anywhere to find the invoking menuitem or
toolitem, at least for the usual case of a uimanager generated thingie
activates it?
I had some actions which pop up a dialog thought to put the dialog on
I noticed various class methods like Gtk2::TextView-new_with_buffer()
ignore their class name arg, so if you call from a subclass
MyTextView-new_with_buffer() you only get a base Gtk2::TextView.
Sample code below.
Is it feasible to have some of those new_with_foo() methods follow the
class name?
Mike Martin redt...@gmail.com writes:
kill (15,$pid1,$pid); ## clears up running external proc ##
if (kill($pid1,$pid) !=0){
Don't forget a bit of time in there for the other process to actually
act on the sigterm (removing temp files or whatever). There's a
Glib::Child-watch_add
Mike Martin redt...@gmail.com writes:
Obviously there is a major issue with the need to have the list in the
right order. So I am wondering if anyone knows of a better and cleaner
way to do this
Gtk2::Builder can make widgets with properties etc. I tend to think it
merely replaces repetitive
muppet sc...@asofyet.org writes:
This behavior implies that there are cases in which the cell renderer
is never being asked to draw itself.
In some of my own tests I wait for map-event before checking certain
callbacks and settings, because a window manager can delay map and
expose for an
PS kali...@gmail.com writes:
one if for catching the time
Glib::Timeout-add will give you a call back when some time has elapsed
(documented in Glib::MainLoop), it's a good way to freshen a display if
you have to poll (rather than listening for events or inputs).
I mistakenly passed an unknown GdkGCValues key to Gtk2::Gdk::GC-new and
it was quietly ignored. Is that intentional, or just that hv_fetch is
easier than iterating in XS?
Croaking on an unknown key could be a good way to catch mistakes that
otherwise merely mysteriously not apply an apparent
Pelle Nilsson pellenils...@fastmail.fm writes:
$filter_rating-signal_connect(value_changed = sub {
$filter_rating_update_needed = TRUE;
Glib::Idle-add(sub {
Or don't do a second idle-add() if the flag is already set ...
The default idle priority is probably about right, as it should
Jeffrey Ratcliffe jeffrey.ratcli...@gmail.com writes:
GLib-GObject-WARNING **: invalid (NULL) pointer instance at
Gtk2-Ex-WYSIWYG/Gtk2/Ex/WYSIWYG.pm line 536.
Looks like creating the tooltips in the BEGIN needs gtk_init. Perhaps
defer that until the first INIT_INSTANCE, or wherever it's first
Matthew Braid ptkp...@mdb.id.au writes:
properties = [Glib::ParamSpec-uint('undo_stack',
'Undo Stack Size',
('The maximum size of the undo '.
'stack. Zero implies no
Marius Feraru altb...@gmail.com writes:
Or ~0,
Yep.
as Glib (quietly) truncates it
Actually I think it's the perl xsubs which truncate :-)
(The same as in a lot of perl interface stuff ...)
___
gtk-perl-list mailing list
gtk-perl-list@gnome.org
Xi Yang jianding...@msn.com writes:
But where are the documents of Glib::Idle?
Hiding in Glib::MainLoop. There's a new Glib::SOURCE_REMOVE there too
for a return value from an idle or timeout. (And similar but different
Gtk2::EVENT_PROPAGATE in Gtk2::Widget.)
Torsten Schoenfeld kaffeeti...@gmx.de writes:
* In syntax check mode, do not call gtk_init. gtk_init may abort the
program (e.g. if DISPLAY is not set), so only call gtk_init_check even
if asked to call gtk_init.
Might put something about that in the docs, since it'll be a bit subtle
1 - 100 of 398 matches
Mail list logo