Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Use Tcl_NewWideIntObj() for values that might be bigger than 32-bit | jan.nijtmans | 2022-02-10 | 1 | -2/+2 |
| | |||||
* | Merge 8.6 | jan.nijtmans | 2022-02-10 | 1 | -5/+5 |
|\ | |||||
| * | Use TclListObjLength/TclListObjGetElements in stead of ↵ | jan.nijtmans | 2022-02-10 | 1 | -5/+5 |
| | | | | | | | | Tcl_ListObjLength/Tcl_ListObjGetElements everywhere. This is slightly more efficient if the refered list already has the correct type | ||||
* | | Merge 8.6 | jan.nijtmans | 2022-01-12 | 1 | -35/+35 |
|\ \ | |/ | |||||
| * | Fix [da6f155ca4]: STRING_SIZE() macro: parenthesize numChars usage. Fix more ↵ | jan.nijtmans | 2022-01-12 | 1 | -1/+1 |
| | | | | | | | | macro's like this. | ||||
* | | Make everything compile warning-free when TCL_THREADS=0 | jan.nijtmans | 2021-08-23 | 1 | -0/+2 |
|\ \ | |/ | |||||
| * | Remove incorrect/useless comment | jan.nijtmans | 2021-03-18 | 1 | -2/+1 |
| | | |||||
* | | TCL_THREADS=0 means no threads too | jan.nijtmans | 2021-06-29 | 1 | -1/+1 |
| | | |||||
* | | Re-apply fix for [67a5eabbd3d1]: refchan, coroutine, and postevent from the ↵ | pooryorick | 2021-06-16 | 1 | -6/+60 |
| | | | | | | | | "watch" proc | ||||
* | | Fix [18f4a94d03] by backout out [9bcec7cd880540c3] | jan.nijtmans | 2021-06-15 | 1 | -60/+6 |
| | | |||||
* | | Fix for io-28.6. | pooryorick | 2021-04-26 | 1 | -4/+4 |
| | | |||||
* | | Proposed implementation of RFE [4b4830eb54] | jan.nijtmans | 2021-03-18 | 1 | -5/+98 |
| | | |||||
* | | Omit -Wdeclaration-after-statement from CFLAGS: No longer needed since we ↵ | jan.nijtmans | 2021-01-08 | 1 | -4/+4 |
| | | | | | | | | | | stopped support for MSVC 6++ Change Tcl_WideInt -> long long in various places: Tcl_WideInt could be a 128-bit type, we don't want that everywhere. | ||||
* | | Add -finput-charset=UTF-8 and -fextended-identifiers to gcc (and clang). All ↵ | jan.nijtmans | 2020-12-08 | 1 | -1/+1 |
| | | | | | | | | C sources can now use UTF-8, as far as gcc/clang/msvc support it. Not used yet | ||||
* | | Fix MSVC++ 6.0 build | jan.nijtmans | 2020-10-14 | 1 | -6/+7 |
| | | |||||
* | | Merge 8.6 | jan.nijtmans | 2020-10-13 | 1 | -3/+5 |
|\ \ | |/ | |||||
| * | More usage of TclNewObj() in stead of Tcl_NewObj() and TclNewIntObj() in ↵ | jan.nijtmans | 2020-10-13 | 1 | -4/+6 |
| | | | | | | | | stead of Tcl_NewIntObj() | ||||
* | | More usage of TclNewIntObj() macro | jan.nijtmans | 2020-08-13 | 1 | -3/+5 |
| | | |||||
* | | More unused arguments through the IO code. | dgp | 2020-03-06 | 1 | -15/+8 |
| | | |||||
* | | Merge 8.6 | jan.nijtmans | 2020-03-04 | 1 | -9/+9 |
|\ \ | |/ | |||||
| * | Implement WideSeekProc() for all channels which have a SeekProc(). Implement ↵ | jan.nijtmans | 2020-03-04 | 1 | -57/+93 |
| | | | | | | | | | | Close2Proc() for all channels, as minimal wrapper around CloseProc(). Backported (with comments) and adapted from core-8-branch. | ||||
* | | Implement TIP #562: Deprecate channel types 1-4 | jan.nijtmans | 2020-02-28 | 1 | -5/+20 |
|\ \ | |||||
| * | | Make Tcl_ChannelCloseProc() deprecated too. Implement close2Proc and ↵ | jan.nijtmans | 2020-01-31 | 1 | -4/+9 |
| | | | | | | | | | | | | wideSeekProc for all internal channel types. | ||||
| * | | Merge 8.7 | jan.nijtmans | 2020-01-30 | 1 | -1/+9 |
| |\ \ | |||||
| * | | | Deprecate channel types 1-4 | jan.nijtmans | 2020-01-29 | 1 | -0/+2 |
| |/ / | |||||
* | | | Let's try a real C++ build with Travis on Linux | jan.nijtmans | 2019-08-30 | 1 | -3/+8 |
|\ \ \ | |/ / | |||||
* | | | Merge tip-548 | jan.nijtmans | 2019-08-15 | 1 | -1/+1 |
|\ \ \ | |/ / | |||||
| * | | Eliminate "register" keyword _everywhere_ in Tcl. This keyword is deprecated ↵ | jan.nijtmans | 2019-07-17 | 1 | -1/+1 |
| | | | | | | | | | | | | in C++ (removed in C++17, even), and essentially does nothing with most modern compilers. | ||||
* | | | Merge 8.7, and continue implementation. Almost done | jan.nijtmans | 2019-07-10 | 1 | -33/+33 |
|/ / | |||||
* | | Make more clear that TCL_INDEX_END|TCL_INDEX_NONE not necessary are int's ↵ | jan.nijtmans | 2019-05-08 | 1 | -4/+4 |
| | | | | | | | | | | (in Tcl 9 they are not). Eliminate use of (local) list_index_t type | ||||
* | | Add missed timer cleanup in tclIORChan.c/ReflectClose. | pooryorick | 2019-04-24 | 1 | -0/+6 |
| | | |||||
* | | Ensure that Tcl_CreateTimerHandler is not called if there is an existing timer | pooryorick | 2019-04-23 | 1 | -2/+6 |
| | | | | | | already scheduled. | ||||
* | | Fix for [67a5eabbd3d1], refchan, coroutine, and postevent from the "watch" | pooryorick | 2019-04-23 | 1 | -6/+50 |
| | | | | | | proc. | ||||
* | | More use of (efficient) TclHasIntRep() macro. Also eliminate many (size_t) ↵ | jan.nijtmans | 2019-02-27 | 1 | -3/+3 |
| | | | | | | | | and (unsigned) type-casts, which don't make sense any more. | ||||
* | | Eliminate the use of macro's like LLONG_MAX|MIN, since they assume that ↵ | jan.nijtmans | 2018-09-13 | 1 | -3/+3 |
| | | | | | | | | Tcl_WideInt equals "long long". Also eliminate uses of Tcl_WideAsLong() and friends, as - often - simple type cases make things more clear. | ||||
* | | Simplify usage of TCL_THREAD, along the lines of ↵ | jan.nijtmans | 2018-05-22 | 1 | -24/+24 |
| | | | | | | | | [eeddb0693a950be980a66de3811630a00c7bab54|eeddb0693a]. Suggested by DKF | ||||
* | | TIP #491 implementation: Threading Support: phasing out non-threaded builds | jan.nijtmans | 2018-05-17 | 1 | -24/+24 |
| | | |||||
* | | merge 8.6 | dgp | 2017-08-10 | 1 | -0/+4 |
|\ \ | |/ | |||||
| * | Silence compiler warning in --disable-threads build | dgp | 2017-08-10 | 1 | -0/+4 |
| | | |||||
* | | Remove some unnecessary "struct" definitions and some type casts no longer ↵ | jan.nijtmans | 2017-04-11 | 1 | -7/+7 |
| | | | | | | | | necessary. No functional changes. | ||||
* | | More minor style fixes. | dkf | 2017-04-10 | 1 | -20/+37 |
| | | |||||
* | | Use more "size_t" in stead of "int" internall. Also eliminate a lot of ↵ | jan.nijtmans | 2016-11-16 | 1 | -9/+9 |
| | | | | | | | | type-casts which are not necessary any more. | ||||
* | | Micro-optimization: Use TclGetStringFromObj in stead of Tcl_GetStringFromObj ↵ | jan.nijtmans | 2016-07-08 | 1 | -4/+4 |
|/ | | | | in many places where possible. | ||||
* | [bbc304f61a] Avoid event handling when reflected channel has a watch | dgp | 2016-03-08 | 1 | -2/+2 |
|\ | | | | | | | | | change half-completed. (First half in 1 thread, second in another). When this is allowed to happen, false alarm errors from [chan postevent] are the result when timing is unlucky. See test iocmd.tf-24.17.bug3522560 . | ||||
| * | [bbc304f61a] Proposed fix for reflected watch race condition. | dgp | 2016-03-02 | 1 | -1/+3 |
|/ | |||||
* | Backout failed attempt to fix [32ae34e63a]. | dgp | 2015-09-25 | 1 | -0/+8 |
|\ | |||||
* \ | [32ae34e63a] Prevent segfaults and data corruption when CopyData() is called ↵ | dgp | 2015-09-23 | 1 | -8/+0 |
|\ \ | |/ | | | | | | | recursively. Mark new test io-53.20 as "knownBug". It demos some unknown flaw in MoveBytes(). | ||||
| * | Protect CopyState buffer from conflicting uses when CopyData() is called | dgp | 2015-09-23 | 1 | -8/+0 |
| | | | | | | recursively. Also, have ReflectWatch() always give driver a chance to act. | ||||
| * | Backport the cheaper Preserve/Release on channels. | dgp | 2014-08-01 | 1 | -2/+2 |
| |\ | |/ |/| | |||||
* | | Reduce the cost of Preserve/Release on channels | dgp | 2014-07-30 | 1 | -2/+2 |
| | |