Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Make panic in TclParseNumber() work when IEEE_FLOATING_POINT is not defined. | jan.nijtmans | 2017-06-15 | 1 | -1/+1 |
| | |||||
* | Make read and write operations on Tcl_Var handles available in tclInt API. | dkf | 2017-06-08 | 7 | -58/+349 |
|\ | | | | | (This will be useful for tclquadcode, and it addresses a basic omission in Tcl's API that made working with Tcl_Vars rather more specialized than it was ever really intended to be. However, this also closes off the part of the API that was not reasonably usable by ordinary external code; LVT indices require too deep an entanglement into Tcl's implementation.) | ||||
| * | Expose some of the core variable access APIs. dkf_expose_ptrgetvar_8_6 | dkf | 2017-06-06 | 7 | -58/+349 |
| | | | | | | (Cherrypick from [b4dfc30083]) | ||||
* | | Revert part of [95d096e0378b460c6c5168bb55bb2ca8b2fd799e|95d096e037]: Missed ↵ | jan.nijtmans | 2017-06-08 | 1 | -12/+35 |
| | | | | | | | | the fact that tolower() was optimized for the ASCII case as well, so this was a mistake! | ||||
* | | Fix [2738427]: Tcl_NumUtfChars(...) no overflow check. | jan.nijtmans | 2017-06-08 | 1 | -13/+14 |
|\ \ | |||||
| * | | Fix [2738427]: Tcl_NumUtfChars(...) no overflow check. | jan.nijtmans | 2017-06-08 | 1 | -13/+14 |
| | | | |||||
* | | | tclUtil.c: Use TclUtfToUniChar() in stead of handling ASCII characters ↵ | jan.nijtmans | 2017-06-08 | 2 | -36/+13 |
| | | | | | | | | | | | | | | | separately: This macro already does that. Add new test-case for Tcl_NumUtfChars(), for a knownBug still to be fixed. | ||||
* | | | Tcl_GetWideIntFromObj() -> TclGetWideIntFromObj(), and minor other ↵ | jan.nijtmans | 2017-06-07 | 3 | -22/+19 |
| |/ |/| | | | | | simplifications/optimizations. No functional change. | ||||
* | | Follow-up to [67aa9a2070]: Use uppercase consistantly, slight optimization ↵ | jan.nijtmans | 2017-06-06 | 1 | -18/+18 |
|\ \ | |/ | | | | | in character tests, comment fixes. No change in functionality. | ||||
| * | [67aa9a2070] Tcl_UtfToUniChar returns single byte for invalid UTF-8 input as ↵ | jan.nijtmans | 2017-06-06 | 1 | -75/+52 |
| | | | | | | | | documented. | ||||
* | | merge core-8-6-branchbug_67aa9a2070 | jan.nijtmans | 2017-06-02 | 1 | -10/+8 |
|\ \ | |||||
| * | | Change refCount field in DictObj from int to size_t. Cherry-picked from ↵ | jan.nijtmans | 2017-06-02 | 1 | -10/+8 |
| | | | | | | | | | | | | "sebres-8-6-clock-speedup-cr1" branch. | ||||
* | | | Fix [67aa9a207037ae67f9014b544c3db34fa732f2dc|67aa9a2070]: Security: Invalid ↵ | jan.nijtmans | 2017-06-02 | 1 | -3/+9 |
|/ / | | | | | | | UTF-8 can inject unexpected characters | ||||
* | | fixed [a3fb3356b76ec4a853d1b86aadc08675f8bef359]: segfault by sorting of the ↵ | sebres | 2017-05-29 | 1 | -19/+18 |
| | | | | | | | | large lists (firstly mistakenly introduced in [af40c6fb6940bab7]), additionally simplify done-points in Tcl_LsortObjCmd. | ||||
* | | Tcl_UtfToUniChar() -> TclUtfToUniChar() in various places: No change in ↵ | jan.nijtmans | 2017-05-29 | 8 | -58/+58 |
| | | | | | | | | functionality, just faster if ASCII only strings are involved. | ||||
* | | fix build failure with TCL_MEM_DEBUG introduced by [8b717dc06a3e3d49] | aspect | 2017-05-19 | 1 | -3/+3 |
| | | |||||
* | | Let local variables declared from within macro's always start with ↵ | jan.nijtmans | 2017-05-09 | 2 | -41/+41 |
| | | | | | | | | underscore, this fixes some gcc warnings with -Wshadow. | ||||
* | | [6015221f59] Segfault after overflow of [binary] field specifier numeric count. | dgp | 2017-05-05 | 1 | -1/+9 |
|\ \ | |/ | |||||
| * | [6015221f59] Segfault after overflow of [binary] field specifier numeric count. | dgp | 2017-05-05 | 1 | -1/+9 |
| | | |||||
* | | [f9fe90d0fa] [file join] normalization. See filesystem-1.52* | dgp | 2017-05-01 | 1 | -1/+11 |
|\ \ | |/ | |||||
| * | [f9fe90d0fa] [file join] normalization. See filesystem-1.52* | dgp | 2017-05-01 | 1 | -1/+11 |
| | | |||||
* | | [f34cf83dd0] An optimization was being taken in a case where it produced the ↵ | dgp | 2017-04-28 | 1 | -0/+7 |
|\ \ | |/ | | | | | wrong result, failing to collapse multiple /// into /. Testing on Windows where path expectations may vary would be a good idea, but since this is just an optimization avoidance, I suspect we're ok. | ||||
| * | [f34cf83dd0] An optimization was being taken in a case where it produced | dgp | 2017-04-28 | 1 | -0/+7 |
| | | | | | | | | | | the wrong result, failing to collapse multiple /// into /. Testing on Windows where path expectations may vary would be a good idea, but since this is just an optimization avoidance, I suspect we're ok. | ||||
| * | [04e26c02c0] Remove useless condition that raises warnings. | dgp | 2017-04-27 | 1 | -6/+2 |
| | | |||||
* | | silence uninit variable warnings | dgp | 2017-04-28 | 1 | -1/+1 |
| | | |||||
* | | [04e26c02c0] Remove useless condition that raises warnings. | dgp | 2017-04-27 | 1 | -4/+2 |
| | | |||||
* | | [50750c735a] Fix for uninit memory handling issue in zlib transforms.bug_50750c735a | dkf | 2017-04-25 | 1 | -18/+16 |
| | | |||||
* | | A better way of getting source file location information when disassembling. | dkf | 2017-04-22 | 3 | -20/+54 |
| | | |||||
* | | Don't use "0%o" format in test-cases, as it suggest's it's the normal way to ↵ | jan.nijtmans | 2017-04-14 | 1 | -1/+1 |
| | | | | | | | | | | format octal numbers: it isn't. Better use "%#o". Add tests for "format" and "scan" corner-cases which weren't documented (except in TIP's) neither had tests before. | ||||
* | | Revert recent commit [80252e0aed]. TIP 237 is clear that %llu is invalid. | dgp | 2017-04-12 | 2 | -22/+13 |
| | | |||||
* | | merge fix-42202ba1e5ff566e | sebres | 2017-04-12 | 1 | -1/+31 |
|\ \ | |||||
| * | | bug fix for [42202ba1e5ff566e0f9abb9f890e460fbc6c1c5c]: segfault by coro injectfix_42202ba1e5ff566e | sebres | 2017-04-04 | 1 | -1/+31 |
| | | | | | | | | | | | | rewritten callback for ::tcl::unsupported::inject, without leave the interpreter in unusable state (inaccurate environment exchange by adding eval callback), test covered now. | ||||
* | | | Fix [4b12ccb3363e81b132e8dbe12aeec596102be1a8|4b12ccb336]: format/scan %llu ↵ | jan.nijtmans | 2017-04-05 | 2 | -13/+22 |
| | | | | | | | | | | | | doesn't work. Also added new test-cases showing the expected behavior. | ||||
* | | | Code optimization/reduction: If TCL_WIDE_INT_IS_LONG is defined, the ↵ | jan.nijtmans | 2017-04-03 | 1 | -5/+12 |
| | | | | | | | | | | | | variable useWide is always 0, so related code can be eliminated without loss of functionality. | ||||
* | | | [900cb0284bc] Ensure that masking pseudo-methods used for export management ↵ | dkf | 2017-03-29 | 1 | -0/+1 |
| | | | | | | | | | | | | are not introspectable directly. [cherrypick] | ||||
* | | | In comments and documentation, distinguish between __int64 (a Microsoft ↵ | jan.nijtmans | 2017-03-29 | 2 | -3/+3 |
| | | | | | | | | | | | | 64-bit integer type) and int64_t (C99 equivalent) | ||||
* | | | Improve comment on TCL_WIDE_INT_IS_LONG, making it less misleading. | jan.nijtmans | 2017-03-27 | 2 | -3/+3 |
|/ / | | | | | In tclThreadTest.c, allow mainThreadId to be >32 bits. | ||||
* | | redundant end-of-line spacing | jan.nijtmans | 2017-03-15 | 4 | -16/+16 |
| | | |||||
* | | Compile [clock clicks], [clock microseconds], [clock milliseconds] and ↵ | Kevin B Kenny | 2017-03-15 | 7 | -13/+174 |
| | | | | | | | | [clock seconds]. | ||||
* | | Make 'clock' and 'encoding' into compilable ensembles that play with safe ↵kbk_clock_encoding_ensembles | Kevin B Kenny | 2017-03-14 | 3 | -1/+113 |
| | | | | | | | | interps | ||||
* | | Make 'clock' and 'encoding' into proper compilable ensembles | Kevin B Kenny | 2017-03-14 | 4 | -75/+258 |
| | | |||||
* | | Remove "makefile.bc". Not updated for many-many years, most likely it ↵ | jan.nijtmans | 2017-03-07 | 2 | -2/+1 |
| | | | | | | | | doesn't work any more and it doesn't seem to care anyone. Let's see if someone complains ;-) | ||||
* | | ParseTokens failed to fully respect its numBytes argument. | dgp | 2017-02-27 | 1 | -1/+4 |
|\ \ | |/ | |||||
| * | ParseTokens failed to fully respect its numBytes argument. | dgp | 2017-02-27 | 1 | -1/+4 |
| | | |||||
* | | zlib stream finalize calls deflate with no input - this case (and this case | aspect | 2017-02-11 | 1 | -2/+2 |
| | | | | | | | | | | only) must be allowed. Fixes 2nd issue in [25842c161f], introduced by [c1aff52ef3] | ||||
* | | Code cleanup (based on feedback from René Zaumseil): Only call ↵ | jan.nijtmans | 2017-02-07 | 1 | -118/+85 |
| | | | | | | | | GetInvalidIntFromObj() when Tcl_GetIntFromObj() fails. | ||||
* | | Fix [d0f7ba56f0e8f93b7efb5b09ebc30a824bdd577a|d0f7ba56f0]: INST_EQ ↵ | jan.nijtmans | 2017-02-01 | 1 | -18/+4 |
|\ \ | |/ | | | | | first-argument NaN shortcut is too aggressive | ||||
| * | Fix [d0f7ba56f0e8f93b7efb5b09ebc30a824bdd577a|d0f7ba56f0]: INST_EQ ↵ | jan.nijtmans | 2017-02-01 | 1 | -18/+4 |
| | | | | | | | | first-argument NaN shortcut is too aggressive | ||||
* | | Update documentation on recent changes in Tcl_LinkVar. | jan.nijtmans | 2017-01-31 | 3 | -37/+27 |
| | | | | | | | | Don't use TCL_NO_DEPRECATED for disabling tests-cases: Those were not deprecated in 8.6 yet. Minor code clean-up. No functional changes. | ||||
* | | Fix [39f6304c2e90549c209cd11a7920dc9921b9f48e|39f6304c2e] follow-up: Make ↵ | jan.nijtmans | 2017-01-20 | 1 | -3/+4 |
|\ \ | |/ | | | | | Tcl_LinkVar toleranto to the empty string as well |