| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
for [scale]'s design that it deals with its value as a formatted
string, and not as a double. [Bug 891141].
|
| | |
|
| |
|
|
| |
Also a few spots where code has been cleaned up more completely.
|
| | |
|
| | |
|
| |
|
|
|
|
| |
* generic/tkScale.c: header file. All tk*Port.h files have long
had a #include <math.h>, and other parts of Tk routinely make use
of libm-supplied math routines.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
updating Tk to use Tcl 8.4's fully CONST-ified
interface, and fully CONSTifying Tk at the
same time.
This patch includes purging Tk of its last
direct access to interp->result. [Bug 589853]
The substantial changes include copying
event sequence strings into Tcl_DStrings
in tkBind.c, and copying [text] indices into
Tcl_DStrings because parsing them involved
overwriting them. If this causes performance
trouble, that can be resolved by further
converting them to Tcl_Obj's.
The #defines USE_NON_CONST and USE_COMPAT_CONST
have the same effect for Tk as they do for Tcl.
(They actually change tcl.h)
|
| |
|
|
| |
Tcl_VarTraceProc typedef. [TIP 27] [Patch 532644].
|
| |
|
|
|
| |
argument of Tcl_GetIndexFromObj(Struct) from (char **) to
(CONST char **). [TIP 27] [Patch 504705]
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
evaluation of widgets.
* generic/tkButton.c (ButtonCreate):
* generic/tkFrame.c (CreateFrame):
* generic/tkMenubutton.c (Tk_MenubuttonObjCmd):
* generic/tkPlace.c (Tk_PlaceObjCmd):
* generic/tkScale.c (Tk_ScaleObjCmd):
* generic/tkMessage.c (Tk_MessageObjCmd):
* generic/tkEntry.c (Tk_EntryObjCmd, Tk_SpinboxObjCmd):
* generic/tkSquare.c (SquareObjCmd): redid the handling of
optionTables in widgets to allow them to be imported into other
namespaces. [Bug #456632]
|
| |
|
|
|
|
|
|
|
| |
TkClassProcs/TkSetClassProcs as Tk_ClassProcs/Tk_SetClassProcs,
adding a size field to Tk_ClassProcs to allow for future
expansion, and renaming the geometryProc to worldChangedProc,
which is more in keeping with the actual use of the callback.
See ChangeLog for details.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* generic/tkEntry.c (Tk_EntryObjCmd): adjusted finishing error
cases and changed TK_CONFIG_NULL_OK to TK_OPTION_NULL_OK
* tests/scale.test:
* generic/tkScale.c:
* generic/tkScale.h:
* unix/tkUnixScale.c:
* mac/tkMacScale.c: moved (PixelToValue|ValueToPixel|SetScaleValue)
to tkScale.c. Caused an associated variable to be immediately set
[Bug: 4833]
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* generic/tkCanvLine.c (ConfigureLine):
* generic/tkCanvPoly.c (ConfigurePoly):
* generic/tkCanvArc.c (Configure/DisplayArc): fixed handling for
negative dash values [Bug: 4104]
* generic/tkScale.c (TkRoundToResolution): fixed incorrect
assumption that (N+1)*delta = N*delta + delta with floating point
math [Bug: 3689, 4099]
(DestroyScale) Fixed check for cancelling TkpDisplayScale (was
REDRAW_ALL, is now REDRAW_PENDING)
|
| |
|
|
|
|
| |
* generic/tkScale.h:
* mac/tkMacScale.c:
* unix/tkUnixScale.c: fixed potential segv from patch in 3897
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* README: updated for patch level 8.3b1
* generic/tkScale.c:
* generic/tkScale.h: fixed possible core when freeing options
(cursor) associated with scale widget [Bug: 3897]
* generic/tk3d.c: added extra calculations to ensure that thin
frames get refreshed too [Bug: 3596]
* generic/tkCanvText.c:
* generic/tkEntry.c:
* generic/tkFont.c:
* generic/tkImgPPM.c: removed extranneous vars that were set but
never used.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
