From bugzilla-daemon at maemo.org Tue May 1 15:40:11 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Tue, 1 May 2007 15:40:11 +0300 (EEST) Subject: [hafqa] [Bug 1276] New: wrong allocators used in hildon_window_get_borders Message-ID: <20070501124011.E7409E0279@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1276 Summary: wrong allocators used in hildon_window_get_borders Product: haf Version: unspecified Platform: All OS/Version: Linux Status: UNCONFIRMED Severity: critical Priority: P2 Component: hildon-libs AssignedTo: michael.kostrzewa at nokia.com ReportedBy: chpe at gnome.org QAContact: hafqa at maemo.org hildon_window_get_borders : g_free (priv->borders); g_free (priv->toolbar_borders); gtk_widget_style_get (GTK_WIDGET (window), "borders",&priv->borders, "toolbar-borders", &priv->toolbar_borders, NULL); if (! priv->borders) priv->borders = (GtkBorder *) g_malloc0 (sizeof (GtkBorder)); if (! priv->toolbar_borders) priv->toolbar_borders = (GtkBorder *) g_malloc0 (sizeof (GtkBorder)); In gtk 2.11, GtkBorder is allocated with GSlice [http://bugzilla.gnome.org/show_bug.cgi?id=383557]. So if the pointer was from gtk_widget_style_get, it'll be allocated with GSlice, but in the failing case you use g_malloc0, and you use g_free to free it. IMHO storing a GtkBorder instead of a GtkBorder* in HildonWindowPrivate, and doing GtkBorder *border = NULL; gtk_widget_style_get (widget, "borders", &border, ...) if (border != NULL) { priv->border = *border; gtk_border_free (border); } else memset (&priv->border, 0, sizeof (GtkBorder)); would be the simplest solution. I haven't checked whether the same problem occurs in other parts of the code. -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Wed May 2 07:44:58 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Wed, 2 May 2007 07:44:58 +0300 (EEST) Subject: [hafqa] [Bug 1279] New: gtk.Range style property "slider-width" freaks out when set too large Message-ID: <20070502044458.218DCE01D2@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1279 Summary: gtk.Range style property "slider-width" freaks out when set too large Product: haf Version: unspecified Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: gtk AssignedTo: tommi.komulainen at nokia.com ReportedBy: sean at cs.gmu.edu QAContact: hafqa at maemo.org The following python code demonstrates a curious effect on VScales and HScales when we change the slider width style property beyond about 25: the slider starts "repeating" itself in a bizarre fashion. import gtk gtk.rc_parse_string('''style "ToolkitScale" { GtkRange::slider-width = 30 } class "GtkRange" style "ToolkitScale" ''') w = gtk.Window() w.add(gtk.HScale()) w.show_all() gtk.main() -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Wed May 2 07:57:00 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Wed, 2 May 2007 07:57:00 +0300 (EEST) Subject: [hafqa] [Bug 1279] gtk.Range style property "slider-width" freaks out when set too large Message-ID: <20070502045700.07F9EE00E3@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1279 ------- Additional Comments From sean at cs.gmu.edu 2007-05-02 07:56 ------- Yeah, yeah, I know I'm supposed to file a separate bug report, but this is related. If you disable the slider (set_sensitive(False)), the trough area is displaced by 1 pixel where the thumb used to be, a weird effect. Look at the far left of the slider region in this code: import gtk gtk.rc_parse_string('''style "ToolkitScale" { GtkRange::slider-width = 30 } class "GtkRange" style "ToolkitScale" ''') w = gtk.Window() h = gtk.HScale() w.add(h) h.set_sensitive(False) w.show_all() gtk.main() -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Wed May 2 10:32:54 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Wed, 2 May 2007 10:32:54 +0300 (EEST) Subject: [hafqa] [Bug 1279] gtk.Range style property "slider-width" freaks out when set too large Message-ID: <20070502073254.DC0ECE03F6@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1279 tommi.komulainen at nokia.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID ------- Additional Comments From tommi.komulainen at nokia.com 2007-05-02 10:32 ------- For graphical bugs such as this one, please include a screenshot. Style properties are, as the name implies, for the theme designers to maintain. And in case of a bitmap based theme such as ours, the pixel values must go hand in hand with the actual bitmaps. Changing the dimensions of one without the other is very likely to cause graphical glitches like this one. Just don't do it. There's a reason why style properties are read-only. -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Wed May 2 18:10:58 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Wed, 2 May 2007 18:10:58 +0300 (EEST) Subject: [hafqa] [Bug 1279] gtk.Range style property "slider-width" freaks out when set too large Message-ID: <20070502151058.8C8F2E0524@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1279 ------- Additional Comments From sean at cs.gmu.edu 2007-05-02 18:10 ------- (In reply to comment #2) > Changing the dimensions of one without the > other is very likely to cause graphical glitches like this one. Just don't do > it. There's a reason why style properties are read-only. Actually, similar style glitches exist in slider without me changing anything. So I have a problem: Nokia's GTK styles are buggy. I have identified several major style bugs in various widgets where a mistake was made by the designer, typically a mismatch between parameters X and Y. I can change X to match Y in the style string and this fixes things nicely and makes things look better. But if I report these bugs, from your reply here I strongly suspect you will eventually modify *Y* to match *X*, which will make things look worse. So unreported bugs they shall remain. As to the style being discussed *here*: Nokia's standard sizes for their sliders and scroll bars are far too thin. They should be at twice as wide as they are presently, and hittable by the finger as well as the stylus. It's not nice to learn that not only are they not that way by default, but Nokia's style developers have seen fit to make _certain_ they _cannot_ be that way in any application. -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 09:53:51 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 09:53:51 +0300 (EEST) Subject: [hafqa] [Bug 719] Hildon Toolbar does not display overflow menu properly Message-ID: <20070503065351.CFF31E0687@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=719 tommi.komulainen at nokia.com changed: What |Removed |Added ---------------------------------------------------------------------------- Alias| |int-51737 -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 10:07:04 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 10:07:04 +0300 (EEST) Subject: [hafqa] [Bug 1196] gtkhtml needs to build-depend and check for gtk >= 2.10 Message-ID: <20070503070704.49BF7E05ED@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1196 tommi.komulainen at nokia.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From tommi.komulainen at nokia.com 2007-05-03 10:07 ------- Looks like FIXED, right? -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 10:35:24 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 10:35:24 +0300 (EEST) Subject: [hafqa] [Bug 1239] HPaned and VPaned set_position is broken Message-ID: <20070503073524.0EF4CE05FE@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1239 ------- Additional Comments From tommi.komulainen at nokia.com 2007-05-03 10:35 ------- What you may be experiencing is GtkPaned normal behavior related to resizing. Please attach a small testcase demonstrating this problem. Important aspect are in which order the widgets are created, shown and packed, as well as the packing parameters. -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 10:37:30 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 10:37:30 +0300 (EEST) Subject: [hafqa] [Bug 1240] label_yalign is broken in gtk.Frame, yet text is too close to frame box Message-ID: <20070503073730.CE5B4E0687@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1240 tommi.komulainen at nokia.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |tommi.komulainen at nokia.com AssignedTo|tommi.komulainen at nokia.com |xan.lopez at nokia.com ------- Additional Comments From tommi.komulainen at nokia.com 2007-05-03 10:37 ------- Xan, you did something with frames? -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 11:03:12 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 11:03:12 +0300 (EEST) Subject: [hafqa] [Bug 1242] gtk.Scale resolution is restricted to powers of 10 via set_digits Message-ID: <20070503080312.934C8E0518@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1242 tommi.komulainen at nokia.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME ------- Additional Comments From tommi.komulainen at nokia.com 2007-05-03 11:03 ------- I don't follow. The following gives me hscale with selection of (0,2,4,6,...) scale = gtk_hscale_new_with_range (0.0, 1024.0, 2.0); gtk_scale_set_digits (GTK_SCALE (scale), 0); Since scale is linear, it is only to be expected that you can't do 0,1,2,4,8,... -- directly that is, see http://developer.gnome.org/doc/API/2.0/gtk/GtkScale.html#GtkScale-format-value -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 11:08:40 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 11:08:40 +0300 (EEST) Subject: [hafqa] [Bug 1240] label_yalign is broken in gtk.Frame, yet text is too close to frame box Message-ID: <20070503080840.3F6A7E0518@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1240 ------- Additional Comments From xan.lopez at nokia.com 2007-05-03 11:08 ------- I've sent to patches upstream to fix drawing glitches with ythickness != 2 (the default value). If he's not touching that value I have no clue what could be going on, but a testcase would help. -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 11:09:43 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 11:09:43 +0300 (EEST) Subject: [hafqa] [Bug 1242] gtk.Scale resolution is restricted to powers of 10 via set_digits Message-ID: <20070503080943.32772E06A3@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1242 ------- Additional Comments From tommi.komulainen at nokia.com 2007-05-03 11:09 ------- Oh well, the scale above still allows selecting odd numbers with careful use of mouse (not keyboard) .. Anyway, you can control the output of that case too with the same signal. -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 11:14:08 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 11:14:08 +0300 (EEST) Subject: [hafqa] [Bug 1243] gtk.Scale's set_draw_value breaks tick marks Message-ID: <20070503081408.0DB1FE06A5@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1243 tommi.komulainen at nokia.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WORKSFORME ------- Additional Comments From tommi.komulainen at nokia.com 2007-05-03 11:14 ------- Umm.. http://developer.gnome.org/doc/API/2.0/gtk/GtkScale.html#gtk-scale-set-digits "Sets the number of decimal places that are displayed in the value." Since after set_draw_value(False) the value is not displayed in the first place then by definition set_digits has no (visible) effect. Guess it's another case for the 'format-value' signal. -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 16:42:46 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 16:42:46 +0300 (EEST) Subject: [hafqa] [Bug 1243] gtk.Scale's set_draw_value breaks tick marks Message-ID: <20070503134246.6E60DE074E@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1243 sean at cs.gmu.edu changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|WORKSFORME | ------- Additional Comments From sean at cs.gmu.edu 2007-05-03 16:42 ------- Tommi, set_digits has TWO functions: 1. It sets the number of digits being displayed (through format-value). 2. It defines the number of ticks that the Scale is restricted to. The user is not permitted to drag the Scale to anything other than a tick mark position. (I agree that the displayed digit resolution has nothing to do with the number of tick marks desired, and so it is really *very* stupid of GTK to have tied these two together, but that's the way it is). When you call set_draw_value(False), it's supposed to turn of #1. Unfortunately it ALSO turns of #2, which is dead wrong. I am reopening this bug as it sounds like in your message that you only verified #1 (which isn't relevant) and figured, rightly, that #1 was appropriate. -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 16:52:55 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 16:52:55 +0300 (EEST) Subject: [hafqa] [Bug 1242] gtk.Scale resolution is restricted to powers of 10 via set_digits Message-ID: <20070503135255.422C2E072C@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1242 sean at cs.gmu.edu changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|WORKSFORME | ------- Additional Comments From sean at cs.gmu.edu 2007-05-03 16:52 ------- If you've verified the bug and have no resolution, why did you close it as resolved? I'm asking that it be reopened. As you have just discovered, the "step size" is another GTK bug: they don't actually *have* a step size. They have a *paging* size that's keyboard-only, as if users primarily used a keyboard. More importantly, N800 users practically NEVER use a keyboard to move a slider. As it stands, there is no way to change the tick resolution of a slider to something other than a power of 10. The workaround I've done in my code is to use powers of 10 internally but to totally change the slider format string to display something else. It's pretty brutally ugly. -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 17:00:00 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 17:00:00 +0300 (EEST) Subject: [hafqa] [Bug 1239] HPaned and VPaned set_position is broken Message-ID: <20070503140000.C97A7E074B@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1239 ------- Additional Comments From sean at cs.gmu.edu 2007-05-03 17:00 ------- Actually Tommi, this will happen even with NO widgets packed in the Paned at all. I'll put together a test case example for you. -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 17:19:17 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 17:19:17 +0300 (EEST) Subject: [hafqa] [Bug 1239] HPaned and VPaned set_position is broken Message-ID: <20070503141917.41904E0764@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1239 ------- Additional Comments From sean at cs.gmu.edu 2007-05-03 17:19 ------- Okay, here's a test case example. set_position is supposed to set the pane position in pixels. So if I set it to 70, it's supposed to move to be 70 pixels away from the left edge. But instead it goes 2/3 across the screen! Setting it to other values does similar things. Note also that the values don't have consistency -- they have some strange nonlinearity to them. And get_position still insists that the bar is located at 70 pixels! If I call set_position *after* gtk.main() is running and everything's been set up, then it sets to the right position. In the example below, this happens when you click on either of the buttons. See for yourself; and try other values (100, 50, etc.) for fun. import gtk w = gtk.Window() h = gtk.HPaned() b1 = gtk.Button("Bad") b1.connect("clicked", lambda widget, data=None: h.set_position(70)) b2 = gtk.Button("Bad") b2.connect("clicked", lambda widget, data=None: h.set_position(70)) h.pack1(b1, resize=True, shrink=True) h.pack2(b2, resize=True, shrink=True) h.set_position(70) w.add(h) w.show_all() gtk.main() -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact. From bugzilla-daemon at maemo.org Thu May 3 17:28:23 2007 From: bugzilla-daemon at maemo.org (bugzilla-daemon at maemo.org) Date: Thu, 3 May 2007 17:28:23 +0300 (EEST) Subject: [hafqa] [Bug 1240] label_yalign is broken in gtk.Frame, yet text is too close to frame box Message-ID: <20070503142823.BF764E076B@maemo.org> https://maemo.org/bugzilla/show_bug.cgi?id=1240 ------- Additional Comments From sean at cs.gmu.edu 2007-05-03 17:28 ------- Here's a test case example. Notice that the descenders of the letters are dropping into the Frame border (very ugly) on the N800. And there is no way to fix it because setting the yalign property has no effect (see the two attempts below). import gtk w = gtk.Window() f = gtk.Frame(label="ggjjhhyyppqq") f.set_label_align(0, 0.5) # This line doesn't have any effect f.set_property("label-yalign", 0.5) # This line doesn't have any effect b1 = gtk.Button("Yucky") f.add(b1) w.add(f) w.show_all() gtk.main() -- Configure bugmail: https://maemo.org/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact.