summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
| | | | |
| \ \ \ \
*-. \ \ \ \ [5adc350683] Stop Tcl forcing EOF condition on channels to be permanent. It ↵dgp2014-11-066-56/+487
|\ \ \ \ \ \ | | | |_|_|/ | | |/| | | | | | | | | may be fleeting, and all parts of Tcl channel ecosystem have to deal with that. New assertions and tests to keep us on track.
| | * | | | [5adc350683] Stop Tcl forcing EOF condition on channels to be permanent.dgp2014-11-064-52/+297
| | |\ \ \ \ | | | | | | | | | | | | | | | | | | | | | It may be fleeting, and all parts of Tcl channel ecosystem have to deal with that. New assertions and tests to keep us on track.
| * | \ \ \ \ merge iogt fixes.bug_5adc350683_86dgp2014-11-062-4/+85
| |\ \ \ \ \ \ | | | |/ / / / | | |/| | | |
| | * | | | | Also test transfroms that delay.bug_5adc350683dgp2014-11-061-0/+30
| | | | | | |
| | * | | | | fix failing testdgp2014-11-061-4/+11
| | | | | | |
| | * | | | | New test iogt-7.0 demos bug in [testchannel transform].dgp2014-11-061-0/+42
| | | | | | |
| * | | | | | Another test checking that handling when transform returns nothing is right.dgp2014-11-061-1/+35
| | | | | | |
| * | | | | | New tests iortrans-4.11* demonstrate what was wrong with the "leaky EOF flag"dgp2014-11-061-4/+85
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | approach in 8.6.1 and earlier. If each level of the channel stack is to have control over its EOF independently, we have to provide for that, even though the Filesystem read APIs make it a big pain. Also test robustness against varing buffer sizes.
| * | | | | | Reflected Transform channel fix. Be sure each EOF on the base channel getsdgp2014-11-051-3/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | passed up to become an eof of the transform before continuing on to additional ReadRaw() from the base channel. This way we don't miss fleeting EOFs.
| * | | | | | New test iortrans-4.10 to demo failure of channel transformation to handledgp2014-11-051-0/+56
| | | | | | | | | | | | | | | | | | | | | | | | | | | | fleeting EOF in the base channel. Falls into infinite block. Regression compared with Tcl 8.6.1.
| * | | | | | merge trunkdgp2014-11-051-1/+1
| |\ \ \ \ \ \ | |/ / / / / / |/| | | | | |
* | | | | | | [214cc0eb22] Restore [lappend $var] return value to the 8.6.1- behavior.dgp2014-11-051-1/+1
| | | | | | | | | | | | | | | | | | | | | If this is going to change, lets not do it by accident.
| * | | | | | Make sure reflected channels do not make a double call to Tcl_ReadRaw(),dgp2014-11-031-4/+5
| | | | | | | | | | | | | | | | | | | | | with the unwarranted assumption that EOF is a permanent condition.
| * | | | | | Same patch re-enabling read after EOF, but here applied to 8.6.dgp2014-11-033-50/+216
| |\ \ \ \ \ \ |/ / / / / / / | | _ / / / / | | / / / / Likely additional changes needed in the other channel transforms new in 8.6.
| * | | | | merge 8.5dgp2014-11-034-3/+19
| |\ \ \ \ \ | | |/ / / /
| * | | | | [5adc350683] Reworked the management of the EOF states to re-enable thedgp2014-11-031-43/+139
| | | | | | | | | | | | | | | | | | | | | | | | ability to read beyond EOF. Plenty of assert()s to keep thing from going off track again.
| * | | | | Base test for [5adc350683].dgp2014-10-291-0/+20
| | | | | |
| * | | | | merge 8.5dgp2014-10-291-0/+2
| |\ \ \ \ \
| * | | | | | Work in progress restoring ability to [read] after [eof] and get non-emptydgp2014-10-282-29/+79
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | strings back in those cases where the channel has them to offer. Also working through all the implications of this possibility on Tcl's more exotic channel features, like stacking.
* | | | | | | Add test-case for previous commit, which shows that when trying to open a ↵jan.nijtmans2014-11-032-0/+8
|\ \ \ \ \ \ \ | | |_|/ / / / | |/| | | | | | | | | | | | filename with invalid characters gives the right error-message. (same bug existed on UNIX too, which is now fixed)
| * | | | | | Add test-case for previous commit, which shows that when trying to open a ↵jan.nijtmans2014-11-032-0/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | filename with invalid characters gives the right error-message. (same bug existed on UNIX too, which is now fixed)
* | | | | | | Better errormessage when file path contains invalid characters.jan.nijtmans2014-11-031-0/+5
|\ \ \ \ \ \ \ | |/ / / / / /
| * | | | | | Better errormessage when file path contains invalid characters. See: ↵jan.nijtmans2014-11-031-0/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | [03414f517b7a74].
* | | | | | | Disable assertion until tls bug it detects is fixed.dgp2014-11-011-1/+7
| | | | | | |
* | | | | | | Extend WinIsReserved() to recognize COM[5-9]: as valid com ports as well. ↵jan.nijtmans2014-10-311-3/+3
|\ \ \ \ \ \ \ | |/ / / / / / | | | | | | | | | | | | | | Prevent possible buffer-overflow if COM name ends with ':'.
| * | | | | | Extend WinIsReserved() to recognize COM[5-9]: as valid com ports as well.jan.nijtmans2014-10-311-2/+2
| | | | | | |
* | | | | | | When translating a reserved devicename to native pathname, strip ':' ↵jan.nijtmans2014-10-311-5/+9
|\ \ \ \ \ \ \ | |/ / / / / / | | | | | | / | |_|_|_|_|/ |/| | | | | postfix. Possible fix for [dcc03414f5], but anyway a good idea.
| * | | | | When translating a reserved devicename to native pathname, strip ':' ↵jan.nijtmans2014-10-311-1/+4
| | |/ / / | |/| | | | | | | | | | | | | postfix. Possible fix for [dcc03414f5], but anyway a good idea.
* | | | | Support for Windows 10jan.nijtmans2014-10-261-0/+2
|\ \ \ \ \ | |/ / / /
| * | | | Support for Windows 10jan.nijtmans2014-10-261-0/+2
| |/ / /
* | | | merge markdgp2014-10-240-0/+0
|\ \ \ \ | |/ / / | | | / | |_|/ |/| |
| * | merge releasedgp2014-10-249-9/+21
| |\ \
| | * | update changescore_8_5_17core_8_5_17_rcdgp2014-10-181-0/+2
| | | |
| | * | merge 8.5dgp2014-10-182-6/+112
| | |\ \ | | |/ / | |/| |
| | * | missed a bumpdgp2014-10-131-1/+1
| | | |
| | * | update changes filedgp2014-10-131-0/+10
| | | |
| | * | Bump to 8.5.17dgp2014-10-108-8/+8
| | | |
* | | | [10dc6daa37] New fix for [gets] on non-blocking channel. This time properly ↵dgp2014-10-182-6/+112
|\ \ \ \ | |/ / / | | | | | | | | accounts for the effects of ENCODING_LINESIZE.
| * | | [10dc6daa37] New fix for [gets] on non-blocking channel. This timedgp2014-10-182-6/+112
| |\ \ \ | | |/ / | |/| | properly accounts for the effects of ENCODING_LINESIZE.
| | * | [10dc6daa37] [gets] on a non-blocking channel must take care so thatdgp2014-10-172-0/+74
| |/ / | | | | | | | | | | | | | | | | | | 1) At least one call to the channel driver input proc gets made. Failure to do this locks up the channel - catastrophic FAIL. 2) After any driver call reports BLOCKED, don't call again. This is less serious, but FAILs to respect the non-blocking setting. Code corrections and tests included, to restore 8.5.15 compat.
* | | fconfigure -peername and -sockname return empty string while async connect ↵oehhar2014-10-173-43/+104
|\ \ \ | |_|/ |/| | | | | running.
| * | New tests: 14.16: -peername empty while async connect running, 14.17: -socknameoehhar2014-10-171-0/+22
| | |
| * | Merge trunkoehhar2014-10-1713-61/+214
| |\ \ | |/ / |/| |
* | | Resolve test conflicts over global varsdgp2014-10-102-4/+11
| | |
* | | merge markdgp2014-10-100-0/+0
|\ \ \ | | |/ | |/|
| * | backport those tests that can bedgp2014-10-101-0/+62
| | |
* | | [bf7135428c] Restore the Tcl_Write() return value logic.dgp2014-10-101-1/+4
|\ \ \ | |/ /
| * | [bf7135428c] Restore the Tcl_Write() return value logic.dgp2014-10-101-1/+4
| | |
* | | [ed29c4da21] Completed fix for [chan copy] handling [chan blocked].dgp2014-10-101-1/+5
| | |
* | | Add Colin's test for coro floor above street level [Bug #3008307]ferrieux2014-10-101-0/+14
| | |
generated by cgit v0.12 at 2024-07-15 10:38:51 (GMT)