summaryrefslogtreecommitdiffstats
path: root/generic
Commit message (Collapse)AuthorAgeFilesLines
* Minor change in InspectRetainedUndoItems().gcramer2017-06-101-2/+1
|
* Guarded version of Tcl_DecrRefCount included, this makes debugging easier.gcramer2017-06-104-41/+124
|
* Coorection and change of defaults (-all) for command "inspect".gcramer2017-06-101-7/+7
|
* Fix in tkBool.h for Windows.gcramer2017-06-091-6/+3
|
* Finally (hopefully) changed to stdbool.h, this is C99 conform, and should ↵gcramer2017-06-0821-135/+47
| | | | compile with ancient MSVC, because for the latter I'm still using the type definition to int. With this change I had to replace some of the struct attributes to int.
* Work-around for Mac: "bool" is replaced with "tkbool_t", and the (quasi) ↵gcramer2017-06-0821-12/+94
| | | | keyword "bool" now will only be used internally.
* Chnage in tkBool: Apple's clang compiler is converting the typedef to an int ↵gcramer2017-06-086-16/+24
| | | | into a different type with one byte, and this caused a crash.
* Safety handling in MarkDeleteProc(), delete mark anyway, but print a warning ↵gcramer2017-06-081-11/+26
| | | | if reference count is not zero.
* Last fix has been reverted, because here it's only a temporary state, while ↵gcramer2017-06-082-3/+1
| | | | clearing the widget.
* Fix when releasing all marks (TkTextFreeMarks): only retain private marks if ↵gcramer2017-06-082-1/+5
| | | | option -steadymarks is enabled.
* Fix in MarkCheckProc(): if -steadymarks is not enabled, then we should not ↵gcramer2017-06-081-2/+8
| | | | have preserved marks.
* Fix in SetMark(): only steady marks are not changing the position relative ↵gcramer2017-06-081-1/+1
| | | | to other marks.
* Fix in MarkCheckProc().gcramer2017-06-081-8/+10
|
* This snippet corresponds to [b8d773c77a], it's propagation of the latter fix ↵fvogel2017-06-051-1/+5
| | | | inside [82c5df4ae4]. This also makes the revised_text branch compile again on Windows.
* merge trunkfvogel2017-06-024-27/+49
|\
| * Fix [bc43fd20cf]: paneconfigure not working as expectedfvogel2017-06-021-4/+20
| |\
| | * Fix [bc43fd20cf]: paneconfigure not working as expectedbug_bc43fd20cffvogel2017-05-311-4/+20
| | |
| * | merge core-8-6-branchjan.nijtmans2017-05-201-0/+1
| |\ \ | | |/
| | * .... missing else ... in previous commitjan.nijtmans2017-05-201-0/+1
| | |
| * | merge core-8-6-branchjan.nijtmans2017-05-201-1/+1
| |\ \ | | |/
| | * Bug-fix [434d294df] continuation: fix logic error detected by Christian ↵jan.nijtmans2017-05-201-1/+1
| | | | | | | | | | | | Werner. Thanks!
| | * Eliminate empty if() statement, if DEBUG_LAYOUT_WITH_BASE_CHUNKS is not ↵jan.nijtmans2017-05-191-2/+2
| | | | | | | | | | | | defined. (backported from androwish. Thanks, Christian!)
| * | merge core-8-6-branchjan.nijtmans2017-05-193-23/+28
| |\ \ | | |/
| | * Slightly use of more "int" in stead of double, for 100% compatibility at ↵bug_434d294dfjan.nijtmans2017-05-191-15/+14
| | |\ | | | | | | | | | | | | script level.
| | * \ merge trunk. jan.nijtmans2017-05-181-1/+5
| | |\ \ | | | |/ | | | | Make "font actual" return integers, for compatibility.
| | * | Another attempt to fix ↵jan.nijtmans2017-05-153-21/+23
| | |/ | | | | | | | | | [434d294df8b053246ee86e7898d06bc3a6d1d771|434d294df8], this time (hopefully) suitable for 8.6. (less changes than the original attempt)
* | | Fixed errors of commit [b2f64dc3df] that prevented tests textDisp-2.23, ↵fvogel2017-05-281-2/+2
| | | | | | | | | | | | textDisp-5.1 and textDisp-22.9 from passing
* | | Bugfix [df89872c94]:gcramer2017-05-281-32/+27
| | | | | | | | | | | | | | | (1) Case "lines" now is based on character position, not byte position as before (2) Test case textIndex-19.14 added (3) Update of manual, with refinment of the concerned descriptions.
* | | Bugfix [46d7a4d153]: erroneous reallocation with NULL instead of bitset.gcramer2017-05-271-1/+1
| | |
* | | Bugfix [cda289a8ea]: The old handling/implementation of the selection ↵gcramer2017-05-266-279/+339
| | | | | | | | | | | | options is tohubohu, so I used the opportunity to overwork it: (1) I added the tag options -inactivebackground and -inactiveforegound, and these options are tied to widget options -inactiveselectbackground and -inactiveselectforeground. (2) For symmetry reasons I added the tag options -inactiveselectbackground and -inactiveselectforeground, these options will overrule the options -inactivebackground and -inactivebackgound of the "sel" tag, provided that the actual tag has a higher priority. (3) The manual has been updated with new options. Furthermore section "THE SELECTION" has been refined. (4) In legacy widget tag option -selectbackground is tied to widget option -selectbackground if the tag option -selectbackground is not null, otherwise the widget option is tied to tag option "-background", this is very confusing, and not conform to documentation, this binding has been changed. Now the widget option -selectbackground is tied with tag option "-background" (of the "sel" tag), this is conform to (revised and legacy) documentation, it is a clear behavior, and allows more freedom in configuration. The tag options "-selectbackground" and "-selectforeground" now will overrule the options "-background" and "-foreground" of the "sel" tag, provided that the actual tag has a higher priority. (5) I changed test case textTag-5.23 according to (3). Moreover this test case has been extended for testing all bindings. BTW: Test case text-5.24 has been removed, it was a duplicate of prior textTag-5.23. (6) Complete rework of function MakeStyle(), the "sel" tag now will be handled separately, after all other tags have been processed, this makes it easier to follow the flow. (7) The old implementation has an erroneous resource management with the shared (tied) options of the "sel" tag and the selection options of the widget. This has been replaced with a proper implementation. Unfortunately the new implementation for resource management of shared options is a bit tricky, because the option table does not support shared options. (8). DEF_TEXT_INACTIVE_SELECT_BG_COLOR has been set to NULL for Windows, this should finally fix the issue of this bug report.
* | | "#ifdef MAC_OSX_TK" code replaced, the generic implementation should not ↵gcramer2017-05-252-29/+32
| | | | | | | | | | | | contain platform specific code (only debugging code is an exception).
* | | Correction of typo in comment.gcramer2017-05-241-1/+1
| | |
* | | (1) TkBTreeGetSegmentTags() now is sorting tags according to given sort ↵gcramer2017-05-247-159/+167
| | | | | | | | | | | | | | | method. This is simplifying function MakeStyle() significantly, and the tags in output of commands "dump" and "inspect" will also be sorted. (2) DEF_TEXT_INACTIVE_SELECT_FG_COLOR has been set to NULL for Windows.
* | | Avoid to trigger a wrong assertion, this is fixing bug [3cd0454c1d] (under ↵gcramer2017-05-201-0/+1
| | | | | | | | | | | | Linux).
* | | Some comments added/changed.gcramer2017-05-182-2/+8
| | |
* | | Bugfix [39e510f69e]: Memory corruption eliminated.gcramer2017-05-171-9/+18
| | |
* | | Some warnings (Mac) eliminated.gcramer2017-05-152-14/+16
| | |
* | | Bugfix [cda289a8ea]: Then I have chosen the wrong default foreground color ↵gcramer2017-05-141-1/+1
| | | | | | | | | | | | for DEF_TEXT_INACTIVE_SELECT_FG_COLOR (Windows), I've changed it to "SystemWindowText", this should show the same foreground color as legacy text widget when unfocused.
* | | Refactorization, and a correction of previous commit in ↵gcramer2017-05-141-8/+8
| | | | | | | | | | | | TkBTreeStartSearchBack().
* | | (1) Wrong quick test in FindTagEnd() removed.gcramer2017-05-142-36/+49
| | | | | | | | | | | | | | | (2) Correction of assertion in TkTextTagCmd:TAG_PREVRANGE. (3) Some comments added. (4) Some code beautyfied.
* | | merge trunkfvogel2017-05-134-9/+34
|\ \ \ | |/ /
| * | merge core-8-6-branchjan.nijtmans2017-05-121-1/+1
| |\ \ | | |/
| | * Revert [f38091d0]: If TK_LAYOUT_WITH_BASE_CHUNKS is set (MacOSX), this ↵jan.nijtmans2017-05-121-1/+1
| | | | | | | | | | | | change leads to crash, struct CharInfo is defined differently depending on TK_LAYOUT_WITH_BASE_CHUNKS (however strange that is ...)
| * | merge core-8-6-branchjan.nijtmans2017-05-103-7/+32
| |\ \ | | |/
| | * New internal function TkpCancelWarp(), which does a proper warp clean-up ↵jan.nijtmans2017-05-103-8/+33
| | | | | | | | | | | | when the display is closed. Follow-up to [db8c541b6bff91848c53006b0fe352136fbd4be9|db8c541b6b]. Backported from androwish [http://www.androwish.org/index.html/info/0b3392c9134c228f|0b3392c9134c228f] (which was a little more than just a "add tk upstream changes" ...). Thanks, Christian!
| | * Remove calls to Tk_FreeXId(), since it's just a NOP.jan.nijtmans2017-05-081-7/+0
| | |
| * | merge core-8-6-branchjan.nijtmans2017-05-082-3/+3
| |\ \ | | |/
| | * Don't use sizeof(struct) when the structure has a char array as last ↵jan.nijtmans2017-05-082-3/+3
| | | | | | | | | | | | element: If the size of this array changes, we'll be in trouble.
* | | Bug in TkTextPickCurrent eliminated.gcramer2017-05-134-3/+32
| | |
* | | Bugfix [283d52f41]: "mark unset" now works with undo, "unset" is removing ↵gcramer2017-05-106-144/+93
| | | | | | | | | | | | the mark from table, although it will be retained for undo.
generated by cgit v0.12 at 2025-03-04 23:13:16 (GMT)