|
2026-10-27
| ||
| 15:13 | Fix [9f4feaaaa1]: a "-" between name characters is part of a constraint name, not a subtraction leaf check-in: 1d88292331 user: jan.nijtmans tags: tcltest-dash-constraints | |
|
2026-10-22
| ||
| 11:24 | • Ticket [9f4feaaaa1] tcltest test contraints with a dash (-) in the name doesn... artifact: 4a62e4dee4 user: jan.nijtmans | |
| 11:23 | Revert change for [9f4feaaaa1], so this one can be decided upon separately check-in: e5335d304d user: jan.nijtmans tags: trunk, main | |
| 08:25 | Many tcltest improvements, [773924ffff], [9f4feaaaa1], [811659ffff], [572400ffff], [8413e79232] and ... check-in: 891b7e7921 user: jan.nijtmans tags: trunk, main | |
|
2026-10-20
| ||
| 19:14 | • Ticket [9f4feaaaa1] tcltest test contraints with a dash (-) in the name doesn... artifact: abc708b102 user: dgp | |
| 14:02 | • Ticket [9f4feaaaa1]: 4 changes artifact: d2a4bee0f0 user: serhiy.storchaka | |
| 14:01 | Fix [9f4feaaaa1]: a "-" between name characters is part of a constraint name, not a subtraction, so ... closed check-in: 98752639ed user: serhiy.storchaka tags: tcltest-dash-constraints | |
|
2023-04-24
| ||
| 21:26 | • Ticket [9f4feaaaa1] tcltest test contraints with a dash (-) in the name doesn... artifact: bf127799b0 user: pointsman | |
| 20:38 | • Ticket [9f4feaaaa1]: 3 changes artifact: 85fb29d1cd user: dgp | |
| 19:12 | • New ticket [9f4feaaaa1]. artifact: 9223de29a8 user: pointsman | |
| Ticket UUID: | 9f4feaaaa11646ec51d345a2697de898850c16f9 | ||
| Title: | tcltest test contraints with a dash (-) in the name doesn't work | ||
| Type: | Bug | Created on: | 2023-04-24 19:12:35 |
| Submitter: | pointsman | Assigned to: | dgp |
| Subsystem: | 34. tcltest Package | Severity: | Severe |
| Priority: | 5 Medium | Last modified: | 2026-10-22 11:24:36 |
| Status: | Open | Closed by: | nobody |
| Resolution: | None | Closed on: | |
| Version: | 8.6.18, 9.0.5, 9.1.0 | ||
| Description: | ||||
Put the following stripped down example in a new file under tests, say rde.test
if {"::tcltest" ni [namespace children]} {
package require tcltest
namespace import -force ::tcltest::*
}
test sometest {some test description} -constraints {
bug-0306a5563
} -body {
set var a
set var
} -result a
cleanupTests
Now try to run this (say from the unix directory of the same soure tree) with
./tcltest ../tests/all.tcl -file rde.test
(You may even use a Tcl 8.6 tclsh or an installed Tcl 9 tclsh; they all show the same behaviour.)
Here on debian I get:
Test files run in separate interpreters
Running tests that match: *
Skipping test files that match: l.*.test
Only running test files that match: rde.test
Tests began at Mon Apr 24 20:57:21 CEST 2023
rde.test
1
bug-0306a5563
Tests ended at Mon Apr 24 20:57:21 CEST 2023
all.tcl: Total 1 Passed 0 Skipped 1 Failed 0
Sourced 1 Test Files.
Notice the strange out put and the missing skip reason.
To set the constraint on the cmd line to true with
./tcltest ../tests/all.tcl -file rde.test -constraints bug-0306a5563
doesn't change anything on the tests run output (other then the start/end time of
course.)
If I rename the constraint to bug0306a5563 in the example above everything works as expected. The run with
./tcltest ../tests/all.tcl -file rde.test
returns
[...]
Test files run in separate interpreters
Running tests that match: *
Skipping test files that match: l.*.test
Only running test files that match: rde.test
Tests began at Mon Apr 24 21:03:39 CEST 2023
rde.test
Tests ended at Mon Apr 24 21:03:39 CEST 2023
all.tcl: Total 1 Passed 0 Skipped 1 Failed 0
Sourced 1 Test Files.
Number of tests skipped for each constraint:
1 bug0306a5563
so no strange output inbetween and a table with the skip reasons.
Running with
./tcltest ../tests/all.tcl -file rde.test -constraints bug0306a5563
returns
[...]
Only running test files that match: rde.test
Tests began at Mon Apr 24 21:05:32 CEST 2023
rde.test
Tests ended at Mon Apr 24 21:05:32 CEST 2023
all.tcl: Total 1 Passed 1 Skipped 0 Failed 0
Sourced 1 Test Files.
The tcltest man page say: "Each constraint has a name, which may be any string".
The example above shows something else.
| ||||
| User Comments: | ||||
dgp added on 2023-04-24 20:38:51:
You call the character a dash, but it is also a minus sign. The presence of an operator causes tcltest to interpret the -constraints argument as a constraint expression, a combination of the constraint "bug" and the constraint "0306a5563". I'm happy to bring this to an end, either completely, or converting to expressions only on logic operators instead of all operators. Just need to bump to tcltest 3. pointsman added on 2023-04-24 21:26:44:
Well, I quoted what I found in the man page about that already in the ticket: "Each constraint has a name, which may be any string" I suspected that it may be something with the dichotomy of the -constraints attribute value "keywordList|expression". But the documentation is mostly silent what "expression" means, perhaps because it recommends to not use it. Seems, even if it should be not used it comes easily into the way of using tcltest. That's unfortunate. I have no clue how often and for which cases contraints expressions are used. So I can easily say remove that completely. It may be important to some. But an expression which _substract_ a boolen value from another boolean doesn't seem a good solution for no matter what problem. serhiy.storchaka added on 2026-10-20 14:02:40:
The stray output and the missing skip table were fixed by [98f03d1dc3] ([0aaeb9c380]), but on main, 9.0 and 8.6 such names are still parsed as Proposed fix in branch [98752639ed|tcltest-dash-constraints] (stacked on [e055230591|tcltest-nested-match], for [773924ffff]): a "-" between name characters is part of the name, so dgp added on 2026-10-20 19:14:31:
This is not a new feature, but a compatibility breaking change, so best introduced in a new major release tcltest 3. Given the freedom to break compat, I would be happier breaking it harder and eliminating constraint expressions completely. We can sort out the enthusiasts for the different approaches in a TIP. jan.nijtmans added on 2026-10-22 11:24:36:
I already merged the proposed fix to trunk, but - thinking about it - better discuss about this first seperately. | ||||
