Tk Source Code

Backport of [880f8c8d59] from 9.0 and 9.1. A pane in a hidden notebook tab is unmapped, so its geometry request was taken as the initial request of a pane which is not shown yet, and it overwrote the sash position. The new Ttk_ContentIsMapped() tells whether the pane has been placed, which does not change while the panedwindow itself is unmapped. Test paned-unmapped-1. check-in: 0b59d9cd user: serhiy.storchaka tags: core-8-6-branch

Backport of [87fa518cee] from 9.0 and 9.1: the SIGINT, SIGHUP and SIGTERM handlers are only installed if the application which embeds Tk has not installed its own (e.g. Python raises KeyboardInterrupt from its SIGINT handler). check-in: c505fe41 user: serhiy.storchaka tags: core-8-6-branch

Backport of the fixes from 9.0 and 9.1.

A WM_KEYDOWN message with VK_PACKET, sent for a character entered via an input method or the touch keyboard, was re-posted with the character as the virtual key code, so the KeyPress event had the keysym of a function key. Now no events are generated for the VK_PACKET key itself; the WM_CHAR message produced by TranslateMessage() generates the key events.

GetTranslatedKey() retrieved the WM_CHAR messages with GetMessageA(), so Windows converted the characters to the code page of the keyboard layout, and TkpGetString() decoded them with a code page guessed from the input language. When the two differed, the typed characters were mangled. Now the messages are retrieved as Unicode and the characters are stored as UTF-8 in the key event, and the code page machinery is removed.

Tests event-8.4, event-8.5 and event-8.6, and the missing testsendinput constraint. check-in: 4a7d6f6b user: serhiy.storchaka tags: core-8-6-branch

The entry of a cloned menu points to a clone of the cascade menu, not to the cascade menu itself. If the cascade entry of a menubar menu was configured with the menu which it already had, ConfigureMenuCloneEntries() passed the -menu option to the clones anyway, so their entries pointed to the cascade menu itself, and the next change of the entry deleted it as if it were a clone. Now the -menu option is not passed to the clones if the cascade menu is not changed. Test menu-42.1. leaf check-in: ef4c3587 user: serhiy.storchaka tags: menu-clone-cascade-destroy

Switching the keyboard layout on Windows releases the modifiers of the key which switches it, so a KeyRelease event for Control arrives after the switch. It was delivered only when the test ran the event loop, which happens after the bindings are set, so it was recorded by tests event-8.7 and event-8.8 and they failed. Now the event loop is run after switching the layout and before setting the bindings. closed check-in: 62fc0288 user: serhiy.storchaka tags: win-keysyms

A menu which is used as a menubar or is torn off is not posted itself: its clone is posted, and on Windows and macOS the menus are drawn by the system, so no Tk window is mapped at all. winfo and wm therefore report nothing useful for such a menu and for the menus of its cascade entries, which is what the ticket describes. leaf check-in: 1742a9e8 user: serhiy.storchaka tags: doc-menu-clone-geometry

Every other widget draws into off-screen memory and copies the result on the screen in a single operation, but the menu drew its entries directly into its window, so they blinked when the pointer moved over them, which is noticeable on a slow display. Each entry is now drawn into a pixmap of its size and copied. The rendering is not changed. leaf check-in: 771fe02c user: serhiy.storchaka tags: x11-menu-double-buffer, core-x11-menu-double-buffer

On X11 the window manager sizes the wrapper window, which is higher than the toplevel by the height of the menubar, and all wm commands are in terms of the toplevel. UpdateSizeHints() passed the minimum and the maximum size to the window manager unchanged, so a window with a menubar could be resized by the user to less than "wm minsize" and could not be resized to "wm maxsize", in both cases by the height of the menubar. The gridded case and a window which is not resizable already accounted for it. Test unixWm-63.1. leaf check-in: 2de1c036 user: serhiy.storchaka tags: x11-menubar-size-hints

On X11 a posted menu is a window of the application: tk_popup and the post widget command return as soon as the menu is posted, and the menu generates <Map> and <Unmap> events. On Windows and macOS the menu is displayed by the system, which runs its own event loop while the menu is posted, so the commands do not return before the menu is unposted again and no <Map> or <Unmap> events are generated for the menu. Neither manual page said so. leaf check-in: a2b23a51 user: serhiy.storchaka tags: doc-popup-platform

Describe the current behaviour without saying what it is for: the option is not evaluated when the user selects the entry, and is evaluated only by the invoke widget command. leaf check-in: 6c701cbc user: serhiy.storchaka tags: doc-cascade-command

The manual said that the -command option of a cascade entry is evaluated whenever the entry is invoked and that this is not supported on Windows. In fact no platform evaluates it when the user selects the entry: on X11 tk::MenuInvoke posts the submenu instead, and on Windows and macOS the menus are handled by the system, which never invokes a cascade item. It is evaluated by the invoke widget command, on all platforms. The documentation is corrected accordingly and a test for a cascade entry is added to the tests of TkInvokeMenu. check-in: 311829db user: serhiy.storchaka tags: doc-cascade-command

Every configure of a canvas redrew all its items, even for options which do not affect their appearance, which is expensive for a canvas with many items. Now the items are redrawn only if an option other than -closeenough, -cursor, -takefocus, the scroll commands and the scroll increments was given. leaf check-in: 9ad5b07c user: serhiy.storchaka tags: canvas-configure-redraw, core-canvas-configure-redraw

On Windows the time of a mouse or key event was the time at which the message was processed, not the time at which it was posted, so if a binding script ran longer than the double click time, the events of the second click got too late times and no <Double-1> event was generated. A new TkpGetEventTime() returns GetMessageTime() on Windows and the current time on macOS. leaf check-in: 6498c685 user: serhiy.storchaka tags: win-event-time, core-win-event-time

When the scroll region is not larger than the visible area, the canvas could be scrolled into the empty area outside the region, although xview and yview reported that there is nothing to scroll. The origin is now kept at the top left corner of the scroll region in this case. leaf check-in: 6f39abf1 user: serhiy.storchaka tags: canvas-confine-small-region, core-canvas-confine-small-region

If the bindings for Enter and Leave moved the current item away from the pointer and back, the canvas chose the current item again and again in a loop when it was redisplayed, without returning to the event loop, so the application hung. Now the current item is chosen once per redisplay, and if it needs to be chosen again, this is done at the next redisplay.

Also remove the update commands scheduled by tests canvPs-3.1 and canvPs-3.2 for after the tests (part A of [180c47a672]): they were executed in later tests, and the new test canvas-30.1 hangs if one of them is executed in it. leaf check-in: 91ba3680 user: serhiy.storchaka tags: canvas-repick-loop, core-canvas-repick-loop

The glyphs are now drawn in batches of 512, and glyphs outside the 16-bit coordinate range of Xft are skipped, as in the Xft backend. This also fixes [770578] for this backend. leaf check-in: 4c1e52f0 user: serhiy.storchaka tags: x11-bidi-long-text

Tk_ConfigureInfo() reported a synonym option together with the database name of the option it refers to, not with its name as Tk_GetOptionInfo() does. The documentation of TK_CONFIG_SYNONYM said that the dbName field holds the name of the option, while it holds its database name, as in all option tables. leaf check-in: 03d1a43c user: serhiy.storchaka tags: oldconfig-synonym-info, core-oldconfig-synonym-info

Converting huge coordinates to the integer bounding box of a canvas item overflowed, either in the conversion itself or when the box was enlarged by the width of the item, so the item was never displayed or found. The coordinates are now clamped to half the int range before the conversion. leaf check-in: 3ebfb228 user: serhiy.storchaka tags: canvas-bbox-overflow, core-canvas-bbox-overflow

GDI does not draw the last point of a polyline drawn with a cosmetic pen, so which end of a line of width 1 was missing depended on the order of its points. Now the end points of such lines are drawn as on X11: an end point is drawn if and only if it is the left or top end of its segment, independently of the order of the points. leaf check-in: 38ab656c user: serhiy.storchaka tags: win-thin-line-endpoints

Text was drawn without the stipple, and angled text with the stipple was misplaced. Text with a stipple is now drawn into a memory DC and copied through the stipple pattern, in the same way as other stippled objects, for both the Uniscribe and the angled text path. leaf check-in: 99b9c92a user: serhiy.storchaka tags: win-text-stipple

The stipple of the GC was ignored when drawing arcs, chords, pie slices and rectangle outlines on Windows. The rendering through the stipple pattern used for polygons and lines is moved to a shared RenderStippled() and used for them too. This also fixes [220059] (-outlinestipple of rectangles). leaf check-in: 72e46db6 user: serhiy.storchaka tags: win-canvas-arc-stipple