tcl-core — Tcl Core Team discussions about Tcl development

[TCLCORE] Tcl 9 `expr` bug, int.rep typing

[Andreas K. <and...@gm...>]

Hi all.

I believe that I have found a bug in the `expr` command of

	Tcl 9 (commit <1f19a433ec>).

This issue is, IMHO, __not__ a release blocker.

It should not affect the Tcl level at all.

Note however that it can trip/break scripts using extensions
performing type-checks while not doing or able to do shimmering.

IOW extensions expecting expect specific int-reps. Among the affected
will be extensions providing objtypes do not have a string-rep, making
shimmering impossible.

Other extensions will simply shimmer the value back to the int-rep
they need. This is why the script level is not affected, except for
a performance loss I did not bother to try and measure.


Test code:
```
proc show-type {x} { puts <>=[::tcl::unsupported::representation $x] }

set i xx
show-type $i
show-type [expr {0 ? $i : $i}]
show-type [expr {$i}]
```

For Tcl 8.6 the results are:

```
<>=value is a pure string with a refcount of 4, object pointer at 0x1ec87b0, 
string representation "xx"
<>=value is a pure string with a refcount of 3, object pointer at 0x1ec9440, 
string representation "xx"
<>=value is a pure string with a refcount of 3, object pointer at 0x1ecacd0, 
string representation "xx"
```

For Tcl 9 <commit 1f19a433ec> the results for the `expr` command change:

```
<>=value is a pure string with a refcount of 4, object pointer at 
0x55dc2b5776d0, string representation "xx"
<>=value is a list with a refcount of 3, object pointer at 0x55dc2b52aab0, 
internal representation 0x55dc2b5a30a0:0x0, string representation "xx"
<>=value is a list with a refcount of 3, object pointer at 0x55dc2b578c00, 
internal representation 0x55dc2b5a30a0:0x0, string representation "xx"
```

It looks as if the $x going into `expr` is wrapped into a single-element list.
Something I truly see no reason for.

-- 
Happy Tcling,
	Andreas Kupries <and...@gm...>
			<https://core.tcl-lang.org/akupries/>
			<https://akupries.tclers.tk/>
	Developer @	SUSE Software Solutions Germany GmbH
-------------------------------------------------------------------------------

Re: [TCLCORE] TK9.0 rc1 on Windows/arm64

[Jan M. <0x...@gm...>]

Hi Jan.

On Tue, Sep 24, 2024 at 10:38 AM Jan Nijtmans <jan...@gm...> wrote:

> Hi Jan. I looked at this further. You should use
>       --enable-64bit=arm64
> The  'aarch64' should have been an
> alias to 'arm64' (that's how Windows names it), but
> that part doesn't work as intended.
>
> I will have a further look at it, but that - most likely -
> won't make it in the 9.0.0 release.

That was very helpful, thank you so much! I can confirm that using
your suggested changes[0], a Linux/amd64 box can now produce working
Windows/arm64 binaries[1].

Best regards and thanks again.

-j

----

  [0]: https://gitlab.com/cznic/tk9.0/-/commit/ba2cba4d074faa3347a150c18eece40a42e918e6#836efb6e25a091dcb4ff8e1dbb2f0be6a5cbf14c_137_137
  [1]: https://opu.peklo.biz/p/24/09/24/1727170677-e799c.png

Re: [TCLCORE] &mdash; in changes file

[Torsten B. <be...@ty...>]

In Markdown, you typically use two dashes '--' for the ndash and three '---' for the mdash. But ... it depends on the Markdown rendering engine whether they support this syntax. Pandoc does. And you can of course, as Jan said, just use the correct utf-8 character if your engine supports unicode.

(you can find out which implementation supports by trying here: https://babelmark.github.io <https://babelmark.github.io/>)

Regards, Torsten


> Am 24.09.2024 um 10:32 schrieb Jan Nijtmans <jan...@gm...>:
> 
> Op ma 23 sep 2024 om 18:33 schreef Donald G Porter via Tcl-Core <tcl...@li... <mailto:tcl...@li...>>:
> 
> Several lines in changes.md look like this:
> 
>   - `file home` &mdash; Get the user home directory.
> 
> Is there really no more readable source text in Markdown for this?
> 
> Yes, there is:
>     <https://core.tcl-lang.org/tcl/info/fc13eb89d80bad96 <https://core.tcl-lang.org/tcl/info/fc13eb89d80bad96>>
> 
> Since markdown is in UTF-8, any unicode character
> can be used just by inserting the correct byte-sequence.
> 
> Hope this helps,
>      Jan Nijtmans
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core

Re: [TCLCORE] TK9.0 rc1 on Windows/arm64

[Jan N. <jan...@gm...>]

Op ma 23 sep 2024 om 12:04 schreef Jan Mercl <0x...@gm...>:

> On lines 137 and 142 the '--enable-64bit=aarch64' works around
> producing x86_64 dlls othewrwise.
> However libtommath.dll and zlib1.dll produced in the win/ directory
> are still x86_64 dlls.
>

Hi Jan. I looked at this further. You should use
      --enable-64bit=arm64
The  'aarch64' should have been an
alias to 'arm64' (that's how Windows names it), but
that part doesn't work as intended.

I will have a further look at it, but that - most likely -
won't make it in the 9.0.0 release.

Hope this helps,
     Jan Nijtmans

Re: [TCLCORE] &mdash; in changes file

[Jan N. <jan...@gm...>]

Op ma 23 sep 2024 om 18:33 schreef Donald G Porter via Tcl-Core <
tcl...@li...>:

>
> Several lines in changes.md look like this:
>
>   - `file home` &mdash; Get the user home directory.
>
> Is there really no more readable source text in Markdown for this?
>

Yes, there is:
    <https://core.tcl-lang.org/tcl/info/fc13eb89d80bad96>

Since markdown is in UTF-8, any unicode character
can be used just by inserting the correct byte-sequence.

Hope this helps,
     Jan Nijtmans

Re: [TCLCORE] TCL_PACKAGE_PATH 8.6.15 and 9.0

[Jan N. <jan...@gm...>]

Op ma 23 sep 2024 om 20:45 schreef Alexander Schöpe via Tcl-Core <
tcl...@li...>:

> Hello everyone,
>
> Previously, I was able to extend or adjust the `auto_path` in Tcl on my
> Mac using the environment variable `TCL_PACKAGE_PATH`. This no longer seems
> to work for either 8.6.15 or 9.0. What should I do?
>

>From the release notes:
  * TCL_PACKAGE_PATH change from Tcl list to platform path convention
         *** POTENTIAL INCOMPATIBILITY ***

This means, if TCL_PACKAGE_PATH contains multiple paths, they
should be separated with ':' now (or ':' on Windows).

Hope this helps,
      Jan Nijtmans

[TCLCORE] TCL_PACKAGE_PATH 8.6.15 and 9.0

[Alexander S. <a.s...@gm...>]

Hello everyone,

Previously, I was able to extend or adjust the `auto_path` in Tcl on my Mac using the environment variable `TCL_PACKAGE_PATH`. This no longer seems to work for either 8.6.15 or 9.0. What should I do?

So far on the Mac, I have been using the gnumakefile:
make -C tk8.6.15/macosx

Alex

Re: [TCLCORE] Verification request for Tcl 9 changes file

[Donald G P. <don...@ni...>]

On 9/23/24 12:57, apn...@ya... wrote:
> One more (unrelated to the below) possible delta to changes.md - the use of
> # comment syntax within an expr (TIP 582, new to Tcl 9) is not mentioned.
> Letting you know as I didn't want to modify changes.md myself this late in
> the game.

In the section "New command options" we have the line:

  - `expr` supports comments inside expressions

-- 
| Don Porter            Applied and Computational Mathematics Division |
| don...@ni...             Information Technology Laboratory |
| http://math.nist.gov/~DPorter/                                  NIST |
|______________________________________________________________________|

Re: [TCLCORE] Verification request for Tcl 9 changes file

[<apn...@ya...>]

Don,

One more (unrelated to the below) possible delta to changes.md - the use of
# comment syntax within an expr (TIP 582, new to Tcl 9) is not mentioned.
Letting you know as I didn't want to modify changes.md myself this late in
the game.

/Ashok

-----Original Message-----
From: Donald G Porter via Tcl-Core <tcl...@li...> 
Sent: Monday, September 23, 2024 9:55 PM
To: Tcl List Core <tcl...@li...>
Subject: [TCLCORE] Verification request for Tcl 9 changes file


The changes.md file for Tcl 9.0.0 says

"Windows platform needs Windows 7 or Windows Server 2008 R2 or later"

Can someone with knowledge either confirm or correct this claim?

Thanks!

-- 
| Don Porter            Applied and Computational Mathematics Division |
| don...@ni...             Information Technology Laboratory |
| http://math.nist.gov/~DPorter/                                  NIST |
|______________________________________________________________________|



_______________________________________________
Tcl-Core mailing list
Tcl...@li...
https://lists.sourceforge.net/lists/listinfo/tcl-core

Re: [TCLCORE] Verification request for Tcl 9 changes file

[<apn...@ya...>]

I think I added that line based on TIP 592. The purpose was to let people
know Vista and earlier are no longer supported in Tcl 9.

Perhaps it could be worded better, something along the lines of

Tcl 9 is supported on Windows 7, Windows Server 2008 R2 and newer.

As far as testing goes, I'm afraid as for Linux and Unix platforms, I don't
think anyone is actually testing all versions and service packs, the Server
editions in particular. I certainly do not have access.

/Ashok

-----Original Message-----
From: Donald G Porter via Tcl-Core <tcl...@li...> 
Sent: Monday, September 23, 2024 9:55 PM
To: Tcl List Core <tcl...@li...>
Subject: [TCLCORE] Verification request for Tcl 9 changes file


The changes.md file for Tcl 9.0.0 says

"Windows platform needs Windows 7 or Windows Server 2008 R2 or later"

Can someone with knowledge either confirm or correct this claim?

Thanks!

-- 
| Don Porter            Applied and Computational Mathematics Division |
| don...@ni...             Information Technology Laboratory |
| http://math.nist.gov/~DPorter/                                  NIST |
|______________________________________________________________________|



_______________________________________________
Tcl-Core mailing list
Tcl...@li...
https://lists.sourceforge.net/lists/listinfo/tcl-core

[TCLCORE] &mdash; in changes file

[Donald G P. <don...@ni...>]

Several lines in changes.md look like this:

  - `file home` — Get the user home directory.

Is there really no more readable source text in Markdown for this?

Whatever improved appearance this produces after Markdown rendering is
(IMO) not worth the considerable blow to readability for those reading
the original source file.

I'm hoping someone can tell me the Markdown source magic (something
like "---" would be a great answer!) that can deliver the improved
rendered appearance, but still read well in the original.

-- 
| Don Porter            Applied and Computational Mathematics Division |
| don...@ni...             Information Technology Laboratory |
| http://math.nist.gov/~DPorter/                                  NIST |
|______________________________________________________________________|

[TCLCORE] Verification request for Tcl 9 changes file

[Donald G P. <don...@ni...>]

The changes.md file for Tcl 9.0.0 says

"Windows platform needs Windows 7 or Windows Server 2008 R2 or later"

Can someone with knowledge either confirm or correct this claim?

Thanks!

-- 
| Don Porter            Applied and Computational Mathematics Division |
| don...@ni...             Information Technology Laboratory |
| http://math.nist.gov/~DPorter/                                  NIST |
|______________________________________________________________________|

[TCLCORE] TK9.0 rc1 on Windows/arm64

[Jan M. <0x...@gm...>]

Hi list.

I'm still not able to produce working Tcl/Tk Windows arm64 dlls on Linux.

I'm using the latest llvm cross toolchain
(llvm-mingw-20240917-ucrt-ubuntu-20.04-x86_64.tar.xz
from https://github.com/mstorsjo/llvm-mingw/releases/tag/20240917).
The .configure && make
approach mostly works with some minor tweaks. This is the part of a
Makefile I'm using in my
project and the target I'm trying zo use for producing the Windows/arm64 stuff:

   130  lib_winarm64: download
   131          if [ "$(GOOS)" != "linux" ]; then exit 1 ; fi
   132          if [ "$(GOARCH)" != "amd64" ]; then exit 1 ; fi
   133          rm -rf ~/tmp/tcl9* ~/tmp/tk9* $(WINARM64)
   134          mkdir -p $(WINARM64)
   135          tar xf $(TAR) -C ~/tmp
   136          tar xf $(TAR2) -C ~/tmp
   137          sh -c "cd ~/tmp/tcl9.0.0/win ; ./configure
--build=x86_64-linux-gnu --host=aarch64-w64-mingw32
--enable-64bit=aarch64"
   138          sh -c "cd ~/tmp/tcl9.0.0/win ; sed -i
's/-DHAVE_CPUID=1/-UHAVE_CPUID/g' *"
   139          make -C ~/tmp/tcl9.0.0/win -j12
   140          cp -v
~/tmp/tcl9.0.0/libtommath/win64-arm/libtommath.dll
~/tmp/tcl9.0.0/compat/zlib/win64-arm/zlib1.dll ~/tmp/tcl9.0.0/win/
   141          cp -v ~/tmp/tcl9.0.0/win/*.dll $(WINARM64)
   142          sh -c "cd ~/tmp/tk9.0.0/win ; ./configure
--build=x86_64-linux-gnu --host=aarch64-w64-mingw32
--with-tcl=$$HOME/tmp/tcl9.0.0/win  --enable-64bit=aarch64"
   143          make -C ~/tmp/tk9.0.0/win -j12
   144          cp -v ~/tmp/tk9.0.0/win/tcl9tk90.dll
~/tmp/tk9.0.0/win/libtk9.0.0.zip $(WINARM64)
   145          zip -j $(WINARM64)/lib.zip.tmp $(WINARM64)/*.dll
$(WINARM64)/*.zip
   146          rm -f $(WINARM64)/*.dll $(WINARM64)/*.zip
   147          mv $(WINARM64)/lib.zip.tmp $(WINARM64)/lib.zip
   148

On lines 137 and 142 the '--enable-64bit=aarch64' works around
producing x86_64 dlls othewrwise.
However libtommath.dll and zlib1.dll produced in the win/ directory
are still x86_64 dlls. Hence line
140 takes the prebuilt ones instead. Line 138 patches out x86_64
CPUID. Without the patch the build fails:

    tclWin32Dll.c:472:2: error: unknown register name '%eax' in asm
     472 |         "%eax", "%ebx", "%ecx", "%edx", "%esi", "%edi", "memory");
         |         ^
    1 error generated.
make[1]: *** [Makefile:766: tclWin32Dll.o] Error 1

However, the preceding guard at tclWin32Dll:445 should have not been
taken, I think:

   #elif defined(__GNUC__) && defined(HAVE_CPUID)

Hence I suspect I have the configuration still wrong even with
--host=aarch64-w64-mingw32 as AFAIK,
HAVE_CPUID should not be defined when targeting aarch64 CPU.

When I try the results that correctly show being for Windows/arm64,
Tcl shell fails before prompt with "illegal instruction". The CPU used
for testing is Cortex-A72.

Please let me know if I can help somehow in solving this issue from
the Linux side, thanks.

-j

Re: [TCLCORE] Tcl/Tk 9.0.0 Release Candidates

[Paul O. <pa...@po...>]

When running the Tcl test suite on Debian 12.6 with AddressSanitizer enabled, the following errors are issued in some tests:

==291391==ERROR: AddressSanitizer: global-buffer-overflow on address 0x7fa5ca2f2c47 at pc 0x564fa892abf0 bp 0x7ffd2dc05c20 sp 0x7ffd2dc053d0
READ of size 17 at 0x7fa5ca2f2c47 thread T0
     #0 0x564fa892abef in MemcmpInterceptorCommon(void*, int (*)(void const*, void const*, unsigned long), void const*, void const*, unsigned long) (/home/obermeier/poSoft/BawtBuild-9.0.0-9.0.0-Sanitizer/Linux/x64/Debug/Build/Tcl/tcltest+0x190bef)
     #1 0x564fa892b288 in memcmp (/home/obermeier/poSoft/BawtBuild-9.0.0-9.0.0-Sanitizer/Linux/x64/Debug/Build/Tcl/tcltest+0x191288)
     #2 0x7fa5c9d01135 in Tcl_IsChannelExisting /home/obermeier/poSoft/BawtBuild-9.0.0-9.0.0-Sanitizer/Linux/x64/Debug/Build/Tcl/generic/tclIO.c:10764
     #3 0x7fa5c1ac4751 in TransferEventProc generic/threadCmd.c:3530

The following change in Tcl_IsChannelExisting should solve the problem:
As memcmp does not stop comparing after finding a null character, the value of chanNameLen should be either the minimum of strlen(chanName) and strlen(name) or strcmp might be used instead of memcmp.


clock-ivm.test
   Validity default mode: off
Test file error: /home/obermeier/poSoft/BawtBuild-9.0.0-9.0.0-Sanitizer/Linux/x64/Debug/Build/Tcl/generic/tclClockFmt.c:192:10: runtime error: signed integer overflow: -9223372036854775800 - 9 cannot be represented in type 'long long int'
/home/obermeier/poSoft/BawtBuild-9.0.0-9.0.0-Sanitizer/Linux/x64/Debug/Build/Tcl/generic/tclClockFmt.c:182:10: runtime error: signed integer overflow: 9223372036854775800 + 8 cannot be represented in type 'long long int'
/home/obermeier/poSoft/BawtBuild-9.0.0-9.0.0-Sanitizer/Linux/x64/Debug/Build/Tcl/generic/tclClockFmt.c:182:16: runtime error: signed integer overflow: 2767011611056432742 * 10 cannot be represented in type 'long long int'
/home/obermeier/poSoft/BawtBuild-9.0.0-9.0.0-Sanitizer/Linux/x64/Debug/Build/Tcl/generic/tclClockFmt.c:192:16: runtime error: signed integer overflow: -2767011611056432742 * 10 cannot be represented in type 'long long int'

Did not look into the details of this error message.

Paul


Am 20.09.2024 um 21:43 schrieb Paul Obermeier:
> Here are the results of the Tcl9 test suite using my BAWT build environments:
>
> Operating system        Arch. Compiler       Errors
> -----------------------------------------------------
> MacOS BigSur (11.7.10)  x64     clang 13.0.0    0
> MacOS Sonoma (14.6.1)   M2      clang 15.0.0    0
> Debian 12.0             x64     gcc 12.2.0      0
> Suse 15.5               x64     gcc 7.5.0       0
> Ubuntu 23.10            x64     gcc 13.2.0      0
> Raspberry Pi OS         arm64   gcc 12.2.0      0
> Debian (StarFive)       RiscV   gcc 12.2.0      1
> Windows 11              x64     gcc 7.2.0       3
> Windows 11              x64     VS2022          3
>
> StarFive errors:
> ================
>
> ==== binary-40.3 ScanNumber: NaN FAILED
> ==== Contents of test case:
>
>     unset -nocomplain arg1
>     list [binary scan \xFF\xFF\xFF\xFF f1 arg1] $arg1
>
> ---- Result was:
> 1 NaN
> ---- Result should have been (glob matching):
> 1 -NaN*
> ==== binary-40.3 FAILED
>
> Windows 11 errors:
> ==================
>
> ==== fCmd-9.3 file rename: comprehensive: file to new name FAILED
> ==== Contents of test case:
>
>     createfile tf1
>     createfile tf2
>     testchmod 0o444 tf2
>     file rename tf1 tf3
>     file rename tf2 tf4
>     list [lsort [glob tf*]] [file writable tf3] [file writable tf4]
>
> ---- Test generated error; Return code was: 1
> ---- Return code should have been one of: 0 2
> ---- errorInfo: error renaming "tf2" to "tf4": permission denied
>     while executing
> "file rename tf2 tf4"
>     ("uplevel" body line 6)
>     invoked from within
> "uplevel 1 $script"
> ---- errorCode: POSIX EACCES {permission denied}
> ==== fCmd-9.3 FAILED
>
>
> ==== fCmd-9.10 file rename: comprehensive: file to new name and dir FAILED
> ==== Contents of test case:
>
>     createfile tf1
>     createfile tf2
>     file mkdir td1
>     testchmod 0o444 tf2
>     file rename tf1 [file join td1 tf3]
>     file rename tf2 [file join td1 tf4]
>     list [glob tf*] [lsort [glob -directory td1 t*]]  [file writable [file join td1 tf3]] [file writable [file join td1 tf4]]
>
> ---- Test generated error; Return code was: 1
> ---- Return code should have been one of: 0 2
> ---- errorInfo: error renaming "tf2" to "td1/tf4": permission denied
>     while executing
> "file rename tf2 [file join td1 tf4]"
>     ("uplevel" body line 7)
>     invoked from within
> "uplevel 1 $script"
> ---- errorCode: POSIX EACCES {permission denied}
> ==== fCmd-9.10 FAILED
>
>
> ==== winFCmd-3.10 TclpDeleteFile: path is readonly FAILED
> ==== Contents of test case:
>
>     createfile tf1
>     testchmod 0 tf1
>     testfile rm tf1
>     file exists tf1
>
> ---- Test generated error; Return code was: 1
> ---- Return code should have been one of: 0 2
> ---- errorInfo: EACCES
>     while executing
> "testfile rm tf1"
>     ("uplevel" body line 4)
>     invoked from within
> "uplevel 1 $script"
> ---- errorCode: NONE
> ==== winFCmd-3.10 FAILED
>
>
>
> Am 18.09.2024 um 21:32 schrieb Donald G Porter via Tcl-Core:
>>
>> Now available at
>>
>> https://sourceforge.net/projects/tcl/files/Tcl/9.0.0/
>>
>> are RC1 candidate source code distribution pre-releases of Tcl and Tk 9.0.0
>>
>> This is the second of a sequence of candidate releases leading to the release of
>> Tcl/Tk 9.0.0. Testing of builds and operations on multiple platforms is invited.  Open
>> tickets on any problems discovered, or raise the issue in a reply to this message.
>>
>> The Tcl pre-release includes pre-releases of the packages Thread 3.0.0 and Itcl 4.3.1.
>> The same level of vetting on them is also appreciated.  The released packages
>> sqlite 3.45.3 and TDBC* 1.1.9 are also included.
>>
>> Several fixes and improvements brought to light under examination of the RC0
>> release candidates are in these new offerings.
>>
>> Thank you for your contributions and assistance.
>>
>
>
>
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core

Re: [TCLCORE] Tcl/Tk 9.0.0 Release Candidates

[Francois V. <fvo...@fr...>]

Le 20/09/2024 à 21:58, Francois Vogel a écrit :
> Le 20/09/2024 à 21:50, Paul Obermeier a écrit :
>> All Linux based systems have several errors.
>> Is there a Linux reference system having zero errors?
>
> Not really.
>
> The CI runners for Linux (Ubuntu) here have no errors:
>
>     https://github.com/tcltk/tk/actions
>
> but this is simply due to failing tests being constrained by 
> 'failsOnUbuntu', i.e. failures are not visible because failing tests 
> just do not run.

A note about the so-called 'failsOnUbuntu' constraint: most of the time 
a test constrained with 'failsOnUbuntu' actually hasn't anything to do 
with Ubuntu but rather is set on a test that fails at the CI runner, 
which happens to run Ubuntu, AND with xvfb. In many cases the failure is 
due to using xvfb instead of a real display server. Certain tests MUST 
run on top of a window manager to pass, they need more than a virtual 
framebuffer. For these tests a window manager must be installed and Xvfb 
is not enough to let the Tk tests succeed at Github Actions. For 
instance things as simple as the following do need a window manager to 
succeed:
     toplevel .t1 ; update
     wm iconify .t1
     update
     winfo ismapped .t1  ; # expected: 0 ; returns 1 with Xvfb without a WM

Also, note that the constraint is really dedicated to the CI runner 
environment, as it is defined by:

     testConstraint failsOnUbuntu [expr {![info exists ::env(CI)] || 
![string match Linux $::tcl_platform(os)]}]

Regards,

Francois

Re: [TCLCORE] Tcl/Tk 9.0.0 Release Candidates

[<apn...@ya...>]

The three Windows errors reported have been there forever and are related to Unix-style chmod being impossible to implement with Windows file permissions (ACL's, inherited permissions etc.) Failure modes will depend on file system (NTFS/FAT etc.), which directory tests run in and security permissions inherited from ancestors.

-----Original Message-----
From: Paul Obermeier <pa...@po...> 
Sent: Saturday, September 21, 2024 1:14 AM
To: Donald G Porter <don...@ni...>; Tcl List Core <tcl...@li...>
Subject: Re: [TCLCORE] Tcl/Tk 9.0.0 Release Candidates

Here are the results of the Tcl9 test suite using my BAWT build environments:

Operating system        Arch. Compiler       Errors
-----------------------------------------------------
MacOS BigSur (11.7.10)  x64     clang 13.0.0    0
MacOS Sonoma (14.6.1)   M2      clang 15.0.0    0
Debian 12.0             x64     gcc 12.2.0      0
Suse 15.5               x64     gcc 7.5.0       0
Ubuntu 23.10            x64     gcc 13.2.0      0
Raspberry Pi OS         arm64   gcc 12.2.0      0
Debian (StarFive)       RiscV   gcc 12.2.0      1
Windows 11              x64     gcc 7.2.0       3
Windows 11              x64     VS2022          3

StarFive errors:
================

==== binary-40.3 ScanNumber: NaN FAILED
==== Contents of test case:

     unset -nocomplain arg1
     list [binary scan \xFF\xFF\xFF\xFF f1 arg1] $arg1

---- Result was:
1 NaN
---- Result should have been (glob matching):
1 -NaN*
==== binary-40.3 FAILED

Windows 11 errors:
==================

==== fCmd-9.3 file rename: comprehensive: file to new name FAILED
==== Contents of test case:

     createfile tf1
     createfile tf2
     testchmod 0o444 tf2
     file rename tf1 tf3
     file rename tf2 tf4
     list [lsort [glob tf*]] [file writable tf3] [file writable tf4]

---- Test generated error; Return code was: 1
---- Return code should have been one of: 0 2
---- errorInfo: error renaming "tf2" to "tf4": permission denied
     while executing
"file rename tf2 tf4"
     ("uplevel" body line 6)
     invoked from within
"uplevel 1 $script"
---- errorCode: POSIX EACCES {permission denied}
==== fCmd-9.3 FAILED


==== fCmd-9.10 file rename: comprehensive: file to new name and dir FAILED
==== Contents of test case:

     createfile tf1
     createfile tf2
     file mkdir td1
     testchmod 0o444 tf2
     file rename tf1 [file join td1 tf3]
     file rename tf2 [file join td1 tf4]
     list [glob tf*] [lsort [glob -directory td1 t*]]  [file writable [file join td1 tf3]] [file writable [file join td1 tf4]]

---- Test generated error; Return code was: 1
---- Return code should have been one of: 0 2
---- errorInfo: error renaming "tf2" to "td1/tf4": permission denied
     while executing
"file rename tf2 [file join td1 tf4]"
     ("uplevel" body line 7)
     invoked from within
"uplevel 1 $script"
---- errorCode: POSIX EACCES {permission denied}
==== fCmd-9.10 FAILED


==== winFCmd-3.10 TclpDeleteFile: path is readonly FAILED
==== Contents of test case:

     createfile tf1
     testchmod 0 tf1
     testfile rm tf1
     file exists tf1

---- Test generated error; Return code was: 1
---- Return code should have been one of: 0 2
---- errorInfo: EACCES
     while executing
"testfile rm tf1"
     ("uplevel" body line 4)
     invoked from within
"uplevel 1 $script"
---- errorCode: NONE
==== winFCmd-3.10 FAILED



Am 18.09.2024 um 21:32 schrieb Donald G Porter via Tcl-Core:
>
> Now available at
>
> https://sourceforge.net/projects/tcl/files/Tcl/9.0.0/
>
> are RC1 candidate source code distribution pre-releases of Tcl and Tk 9.0.0
>
> This is the second of a sequence of candidate releases leading to the release of
> Tcl/Tk 9.0.0. Testing of builds and operations on multiple platforms is invited.  Open
> tickets on any problems discovered, or raise the issue in a reply to this message.
>
> The Tcl pre-release includes pre-releases of the packages Thread 3.0.0 and Itcl 4.3.1.
> The same level of vetting on them is also appreciated.  The released packages
> sqlite 3.45.3 and TDBC* 1.1.9 are also included.
>
> Several fixes and improvements brought to light under examination of the RC0
> release candidates are in these new offerings.
>
> Thank you for your contributions and assistance.
>



_______________________________________________
Tcl-Core mailing list
Tcl...@li...
https://lists.sourceforge.net/lists/listinfo/tcl-core

Re: [TCLCORE] TIP 430 experiments.

[Christian W. <Chr...@t-...>]

On 09/20/2024 10:30 PM, Schelte Bron wrote:

Howdy Schelte,

> ...
> It may also be advisable to build tclsh on an old(ish) OS. This will still run on a newer OS. The reverse is probably not true. At least I got an error: /lib64/libm.so.6: version `GLIBC_2.38' not found

from my experience with undroidwish and LUCK, which uses a very similar approach with a zipfs predating TIP#430
I suggest to pick the oldest Linux you can find. Most Linux x86* binaries of the LUCK stuff are made on a CentOS5
box which might support most of the x86* Linuxen of the last 15 years.

BR,
Christian

Re: [TCLCORE] TIP 430 experiments.

[Schelte B. <tc...@tc...>]

Thanks Brian,

I stupidly went by the instructions given in the TIP. But clearly they 
are missing some crucial steps.

First of all, Tcl has to be configured with --disable-shared and 
--without-system-libtommath (because some (most?) target systems will 
not have libtommath installed). Then, after entering the myvfs.vfs 
directory, you have to install a copy of tcl_library in there. This can 
quite easily be done by unzipping ../tclsh. Finally, it's best to use 
the zipfs mkimg command from Tcl to create the binary. If you just 
append the zipped myvfs.vfs to the end of a copy of tclsh, you end up 
with 2 copies of tcl_library in the binary, because there will then 
essentially be 2 zip files appended. The zipfs mkimg command should not 
produce any output. If it does, there is probably something wrong with 
the specified arguments.

It may also be advisable to build tclsh on an old(ish) OS. This will 
still run on a newer OS. The reverse is probably not true. At least I 
got an error: /lib64/libm.so.6: version `GLIBC_2.38' not found

With those changes, I finally managed to create a single file 
executable. Next step: Include Tk.


Schelte.


On 20/09/2024 14:34, Brian Griffin wrote:
> The mkimg command is not a complete operation. It does only two things: 1) remove any existing zip archive from the original image when copying it to the destination image, and 2) append the new zip archive to the new destination image.
> So, it’s only one small improvement from using cat.
> 
> It is still up to the programmer to create a full library from the original before using mkimg.
> 
>  From my perspective, it only makes sense to use static tclsh or wish images to create custom applications. Using dynamic shared libraries is possible, but complicates construction and installation significantly.
> 
> -Brian
> (from mobile device)
> 
>> On Sep 20, 2024, at 06:29, Schelte Bron <tc...@tc...> wrote:
>>
>> All,
>>
>> As I was interested in the new "Task Executables" functionality provided by TIP 430 (basically an alternative for starpacks), I did some experiments and got unexpected results.
>>
>> I used this main.tcl:
>> puts "Script: [info script]"
>> puts "Library: $tcl_library"
>> puts "Mounted: [dict keys [zipfs mount]]"
>>
>> After compiling tclsh, I performed the following commands:
>> cp tclsh demo
>> zip demo.zip main.tcl
>> cat demo.zip >> demo
>> ./demo
>>
>> First of all, I found a difference between building with and without --disable-shared.
>> Shared:
>> Script: //zipfs:/app/main.tcl
>> Library: //zipfs:/lib/tcl/tcl_library
>> Mounts: //zipfs:/lib/tcl //zipfs:/app
>>
>> Static:
>> Script: //zipfs:/app/main.tcl
>> Library: /home/sbron/tcl/tcl/library
>> Mounts: //zipfs:/app
>>
>> I was surprised to find that with the static build, tcl_library points to the library in the source tree. So then I tried moving the demo binary to another location. That resulted in:
>> application-specific initialization failed: Cannot find a usable init.tcl in the following directories:
>>     {} /home/sbron/usr/lib/tcl9.0 /lib/tcl9.0 /lib/tcl9.0 /library /library /tcl9.0/library /tcl9.0b4/library /tcl9.0b4/library
>>
>>
>>
>> This probably means that Tcl wasn't installed properly.
>>
>> Script: //zipfs:/app/main.tcl
>> Library: /tcl9.0b4/library
>> Mounts: //zipfs:/app
>>
>> Running the static tclsh interactively, I get a different result:
>> % source main.tcl
>> Script: main.tcl
>> Library: //zipfs:/app/tcl_library
>> Mounts: //zipfs:/app
>>
>> I can move the tclsh binary to another location without any issues.
>>
>> As a final attempt, I moved main.tcl into a directory demo.vfs and then ran the following command from the tclsh:
>>
>> % zipfs mkimg demo ./demo.vfs ./tclsh
>> ./demo.vfs/main.tcl
>>
>> This produced a binary that was smaller than the tclsh binary. Running it just gave me a command prompt. Manually sourcing the script then produced the following:
>> % source demo.vfs/main.tcl
>> Script: demo.vfs/main.tcl
>> Library: /home/sbron/tcl/tcl/library
>> Mounts:
>>
>> Is there something I'm doing wrong? Or is the TIP 430 implementation incomplete?
>>
>>
>> Schelte.
>>
>>
>> _______________________________________________
>> Tcl-Core mailing list
>> Tcl...@li...
>> https://lists.sourceforge.net/lists/listinfo/tcl-core
>>

Re: [TCLCORE] Tk 8.6.15 failures on macos Sonoma 14.7 (23H124)

[Paul O. <pa...@po...>]

On my MacBook Air (M2,2022 Sonoma 14.6.1) the Tcl and Tk test suites for 8.6.15 run without errors.

For the Tk test suite I had to set the display resolution to the native resolution of the Retina display, i.e. 2560 x 1600.

Regards,
Paul

Am 20.09.2024 um 09:16 schrieb Alexander Schöpe via Tcl-Core:
> grep FAILED log-tk8.6.15.txt*
> log-tk8.6.15.txt:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-2:==== font-47.2 Bug 3049518 - Canvas FAILED
> log-tk8.6.15.txt-2:==== font-47.2 FAILED
> log-tk8.6.15.txt-2:==== font-47.3 Bug 3049518 - Label FAILED
> log-tk8.6.15.txt-2:==== font-47.3 FAILED
> log-tk8.6.15.txt-2:==== frame-12.3 FrameWorldChanged procedure FAILED
> log-tk8.6.15.txt-2:==== frame-12.3 FAILED
> log-tk8.6.15.txt-2:==== textImage-4.2 alignment checking - baseline FAILED
> log-tk8.6.15.txt-2:==== textImage-4.2 FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-4:==== font-47.2 Bug 3049518 - Canvas FAILED
> log-tk8.6.15.txt-4:==== font-47.2 FAILED
> log-tk8.6.15.txt-4:==== font-47.3 Bug 3049518 - Label FAILED
> log-tk8.6.15.txt-4:==== font-47.3 FAILED
> log-tk8.6.15.txt-4:==== frame-12.3 FrameWorldChanged procedure FAILED
> log-tk8.6.15.txt-4:==== frame-12.3 FAILED
> log-tk8.6.15.txt-4:==== textImage-4.2 alignment checking - baseline FAILED
> log-tk8.6.15.txt-4:==== textImage-4.2 FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 FAILED
>
>
>
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core

Re: [TCLCORE] Tcl/Tk 9.0.0 Release Candidates

[Francois V. <fvo...@fr...>]

Le 20/09/2024 à 21:50, Paul Obermeier a écrit :
> All Linux based systems have several errors.
> Is there a Linux reference system having zero errors?

Not really.

The CI runners for Linux (Ubuntu) here have no errors:

     https://github.com/tcltk/tk/actions

but this is simply due to failing tests being constrained by 
'failsOnUbuntu', i.e. failures are not visible because failing tests 
just do not run.

For quite a long time now I'm trying to address these failures in the 
less_tests_constraints branch. Some failures are due to font issues, 
some others to distribution specificities, and finally some are real Tk 
bugs. Help in this effort is most welcome!

https://core.tcl-lang.org/tk/timeline?r=less_tests_constraints

Regards,

Francois

Re: [TCLCORE] Tcl/Tk 9.0.0 Release Candidates

[Paul O. <pa...@po...>]

Here are the results of the Tk9 test suite using my BAWT build environments:

Operating system        Arch. Compiler       Errors
-----------------------------------------------------
MacOS BigSur (11.7.10)  x64     clang 13.0.0    0
MacOS Sonoma (14.6.1)   M2      clang 15.0.0    0
Debian 12.0             x64     gcc 12.2.0     59
Suse 15.5               x64     gcc 7.5.0      76
Ubuntu 23.10            x64     gcc 13.2.0     54
Raspberry Pi OS         arm64   gcc 12.2.0    111
Debian (StarFive)       RiscV   gcc 12.2.0     47
Windows 11              x64     gcc 7.2.0       0
Windows 11              x64     VS2022          0

All Linux based systems have several errors.
Is there a Linux reference system having zero errors?

Paul

Am 18.09.2024 um 21:32 schrieb Donald G Porter via Tcl-Core:
>
> Now available at
>
> https://sourceforge.net/projects/tcl/files/Tcl/9.0.0/
>
> are RC1 candidate source code distribution pre-releases of Tcl and Tk 9.0.0
>
> This is the second of a sequence of candidate releases leading to the release of
> Tcl/Tk 9.0.0. Testing of builds and operations on multiple platforms is invited.  Open
> tickets on any problems discovered, or raise the issue in a reply to this message.
>
> The Tcl pre-release includes pre-releases of the packages Thread 3.0.0 and Itcl 4.3.1.
> The same level of vetting on them is also appreciated.  The released packages
> sqlite 3.45.3 and TDBC* 1.1.9 are also included.
>
> Several fixes and improvements brought to light under examination of the RC0
> release candidates are in these new offerings.
>
> Thank you for your contributions and assistance.
>

Re: [TCLCORE] Tcl/Tk 9.0.0 Release Candidates

[Paul O. <pa...@po...>]

Here are the results of the Tcl9 test suite using my BAWT build environments:

Operating system        Arch. Compiler       Errors
-----------------------------------------------------
MacOS BigSur (11.7.10)  x64     clang 13.0.0    0
MacOS Sonoma (14.6.1)   M2      clang 15.0.0    0
Debian 12.0             x64     gcc 12.2.0      0
Suse 15.5               x64     gcc 7.5.0       0
Ubuntu 23.10            x64     gcc 13.2.0      0
Raspberry Pi OS         arm64   gcc 12.2.0      0
Debian (StarFive)       RiscV   gcc 12.2.0      1
Windows 11              x64     gcc 7.2.0       3
Windows 11              x64     VS2022          3

StarFive errors:
================

==== binary-40.3 ScanNumber: NaN FAILED
==== Contents of test case:

     unset -nocomplain arg1
     list [binary scan \xFF\xFF\xFF\xFF f1 arg1] $arg1

---- Result was:
1 NaN
---- Result should have been (glob matching):
1 -NaN*
==== binary-40.3 FAILED

Windows 11 errors:
==================

==== fCmd-9.3 file rename: comprehensive: file to new name FAILED
==== Contents of test case:

     createfile tf1
     createfile tf2
     testchmod 0o444 tf2
     file rename tf1 tf3
     file rename tf2 tf4
     list [lsort [glob tf*]] [file writable tf3] [file writable tf4]

---- Test generated error; Return code was: 1
---- Return code should have been one of: 0 2
---- errorInfo: error renaming "tf2" to "tf4": permission denied
     while executing
"file rename tf2 tf4"
     ("uplevel" body line 6)
     invoked from within
"uplevel 1 $script"
---- errorCode: POSIX EACCES {permission denied}
==== fCmd-9.3 FAILED


==== fCmd-9.10 file rename: comprehensive: file to new name and dir FAILED
==== Contents of test case:

     createfile tf1
     createfile tf2
     file mkdir td1
     testchmod 0o444 tf2
     file rename tf1 [file join td1 tf3]
     file rename tf2 [file join td1 tf4]
     list [glob tf*] [lsort [glob -directory td1 t*]]  [file writable [file join td1 tf3]] [file writable [file join td1 tf4]]

---- Test generated error; Return code was: 1
---- Return code should have been one of: 0 2
---- errorInfo: error renaming "tf2" to "td1/tf4": permission denied
     while executing
"file rename tf2 [file join td1 tf4]"
     ("uplevel" body line 7)
     invoked from within
"uplevel 1 $script"
---- errorCode: POSIX EACCES {permission denied}
==== fCmd-9.10 FAILED


==== winFCmd-3.10 TclpDeleteFile: path is readonly FAILED
==== Contents of test case:

     createfile tf1
     testchmod 0 tf1
     testfile rm tf1
     file exists tf1

---- Test generated error; Return code was: 1
---- Return code should have been one of: 0 2
---- errorInfo: EACCES
     while executing
"testfile rm tf1"
     ("uplevel" body line 4)
     invoked from within
"uplevel 1 $script"
---- errorCode: NONE
==== winFCmd-3.10 FAILED



Am 18.09.2024 um 21:32 schrieb Donald G Porter via Tcl-Core:
>
> Now available at
>
> https://sourceforge.net/projects/tcl/files/Tcl/9.0.0/
>
> are RC1 candidate source code distribution pre-releases of Tcl and Tk 9.0.0
>
> This is the second of a sequence of candidate releases leading to the release of
> Tcl/Tk 9.0.0. Testing of builds and operations on multiple platforms is invited.  Open
> tickets on any problems discovered, or raise the issue in a reply to this message.
>
> The Tcl pre-release includes pre-releases of the packages Thread 3.0.0 and Itcl 4.3.1.
> The same level of vetting on them is also appreciated.  The released packages
> sqlite 3.45.3 and TDBC* 1.1.9 are also included.
>
> Several fixes and improvements brought to light under examination of the RC0
> release candidates are in these new offerings.
>
> Thank you for your contributions and assistance.
>

Re: [TCLCORE] Tk 8.6.15 failures on macos Sonoma 14.7 (23H124)

[Alexander S. <a.s...@gm...>]

Hi Marc,

that’s a good tip.
If I remove the second monitor, I’m left with just the error I’ve been having the last few times.

==== focus-1.1 Tk_FocusCmd procedure FAILED

Best
Alex

> Am 20.09.2024 um 16:30 schrieb Marc Culler <cul...@gm...>:
> 
> In my experience, those failures are related to having a second monitor, which affects the maxsize value.  Try turning off your other monitor before running the tests.
> 
> - Marc
> 
> On Fri, Sep 20, 2024, 1:17 AM Alexander Schöpe via Tcl-Core <tcl...@li...> wrote:
> grep FAILED log-tk8.6.15.txt*
> log-tk8.6.15.txt:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-2:==== font-47.2 Bug 3049518 - Canvas FAILED
> log-tk8.6.15.txt-2:==== font-47.2 FAILED
> log-tk8.6.15.txt-2:==== font-47.3 Bug 3049518 - Label FAILED
> log-tk8.6.15.txt-2:==== font-47.3 FAILED
> log-tk8.6.15.txt-2:==== frame-12.3 FrameWorldChanged procedure FAILED
> log-tk8.6.15.txt-2:==== frame-12.3 FAILED
> log-tk8.6.15.txt-2:==== textImage-4.2 alignment checking - baseline FAILED
> log-tk8.6.15.txt-2:==== textImage-4.2 FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-4:==== font-47.2 Bug 3049518 - Canvas FAILED
> log-tk8.6.15.txt-4:==== font-47.2 FAILED
> log-tk8.6.15.txt-4:==== font-47.3 Bug 3049518 - Label FAILED
> log-tk8.6.15.txt-4:==== font-47.3 FAILED
> log-tk8.6.15.txt-4:==== frame-12.3 FrameWorldChanged procedure FAILED
> log-tk8.6.15.txt-4:==== frame-12.3 FAILED
> log-tk8.6.15.txt-4:==== textImage-4.2 alignment checking - baseline FAILED
> log-tk8.6.15.txt-4:==== textImage-4.2 FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 FAILED
> 
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core

Re: [TCLCORE] Tcl/Tk 9.0.0 Release Candidates

[Paul O. <pa...@po...>]

On my MacBook Air (M2,2022 Sonoma 14.6.1) the Tcl and Tk test suites run without errors.

For the Tk test suite I had to set the display resolution to the native resolution of the Retina display, i.e. 2560 x 1600.

Regards,
Paul


Am 19.09.2024 um 10:36 schrieb Torsten Berg:
> Hi,
>
> running the test suite on a Mac mini (M1 processor from 2020) with macOS Ventura 13.6.9:
>
> - all tests pass for Tcl
>
> - three failures for Tk: the first two do not seem to be related to Tk, for the third is weird and I can't see why maxsize reports values bigger than the scree size. However, this does not seem too critical, or?
>
> ==== treetags-2.1 Events delivered to tags FAILED
>
> ==== Contents of test case:
>
>
> focus -force $tv ; update;# needed so [event generate] delivers Key
>
>     $tv focus item1
>
> event generate $tv <a>
>
>     set ::KEY
>
>
> ---- Test generated error; Return code was: 1
>
> ---- Return code should have been one of: 0 2
>
> ---- errorInfo: can't read "::KEY": no such variable
>
> while executing
>
> "set ::KEY"
>
> ("uplevel" body line 5)
>
> invoked from within
>
> "uplevel 1 $script"
>
> ---- errorCode: TCL LOOKUP VARNAME ::KEY
>
> ==== treetags-2.1 FAILED
>
>
>
>
> ==== treetags-2.2 Events delivered to correct tags FAILED
>
> ==== Contents of test case:
>
>
>     $tv tag bind tag2 <Key> [list set ::KEY2 %A]
>
>
>     $tv focus item1
>
> event generate $tv <b>
>
>     $tv focus item2
>
> event generate $tv <c>
>
>
>     list $::KEY $::KEY2
>
>
> ---- Test generated error; Return code was: 1
>
> ---- Return code should have been one of: 0 2
>
> ---- errorInfo: can't read "::KEY": no such variable
>
> while executing
>
> "list $::KEY $::KEY2"
>
> ("uplevel" body line 9)
>
> invoked from within
>
> "uplevel 1 $script"
>
> ---- errorCode: TCL LOOKUP VARNAME ::KEY
>
> ==== treetags-2.2 FAILED
>
>
>
>
> ==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
>
> ==== Contents of test case:
>
>
> toplevel .t
>
> lassign [wm maxsize .t] t_width t_height
>
> set s_width [winfo screenwidth .t]
>
> set s_height [winfo screenheight .t]
>
> expr {($t_width <= $s_width) && ($t_height <= $s_height)}
>
>
> ---- Result was:
>
> 0
>
> ---- Result should have been (exact matching):
>
> 1
>
> ==== wm-maxsize-1.7 FAILED
>
>
>
>
>
>
>
>> Am 18.09.2024 um 21:32 schrieb Donald G Porter via Tcl-Core <tcl...@li...>:
>>
>>
>> Now available at
>>
>> https://sourceforge.net/projects/tcl/files/Tcl/9.0.0/
>>
>> are RC1 candidate source code distribution pre-releases of Tcl and Tk 9.0.0
>>
>> This is the second of a sequence of candidate releases leading to the release of
>> Tcl/Tk 9.0.0. Testing of builds and operations on multiple platforms is invited.  Open
>> tickets on any problems discovered, or raise the issue in a reply to this message.
>>
>> The Tcl pre-release includes pre-releases of the packages Thread 3.0.0 and Itcl 4.3.1.
>> The same level of vetting on them is also appreciated. The released packages
>> sqlite 3.45.3 and TDBC* 1.1.9 are also included.
>>
>> Several fixes and improvements brought to light under examination of the RC0
>> release candidates are in these new offerings.
>>
>> Thank you for your contributions and assistance.
>>
>> -- 
>> | Don Porter            Applied and Computational Mathematics Division |
>> | don...@ni...             Information Technology Laboratory |
>> | http://math.nist.gov/~DPorter/                                  NIST |
>> |______________________________________________________________________|
>>
>>
>>
>> _______________________________________________
>> Tcl-Core mailing list
>> Tcl...@li...
>> https://lists.sourceforge.net/lists/listinfo/tcl-core
>
>
>
>
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core

Re: [TCLCORE] Tk 8.6.15 failures on macos Sonoma 14.7 (23H124)

[Marc C. <cul...@gm...>]

In my experience, those failures are related to having a second monitor,
which affects the maxsize value.  Try turning off your other monitor before
running the tests.

- Marc

On Fri, Sep 20, 2024, 1:17 AM Alexander Schöpe via Tcl-Core <
tcl...@li...> wrote:

> grep FAILED log-tk8.6.15.txt*
> log-tk8.6.15.txt:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt:==== wm-maxsize-1.7 maxsize must be <= screen size FAILED
> log-tk8.6.15.txt:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-2:==== font-47.2 Bug 3049518 - Canvas FAILED
> log-tk8.6.15.txt-2:==== font-47.2 FAILED
> log-tk8.6.15.txt-2:==== font-47.3 Bug 3049518 - Label FAILED
> log-tk8.6.15.txt-2:==== font-47.3 FAILED
> log-tk8.6.15.txt-2:==== frame-12.3 FrameWorldChanged procedure FAILED
> log-tk8.6.15.txt-2:==== frame-12.3 FAILED
> log-tk8.6.15.txt-2:==== textImage-4.2 alignment checking - baseline FAILED
> log-tk8.6.15.txt-2:==== textImage-4.2 FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 maxsize must be <= screen size
> FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 maxsize must be <= screen size
> FAILED
> log-tk8.6.15.txt-2:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 maxsize must be <= screen size
> FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 maxsize must be <= screen size
> FAILED
> log-tk8.6.15.txt-3:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 maxsize must be <= screen size
> FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 FAILED
> log-tk8.6.15.txt-4:==== font-47.2 Bug 3049518 - Canvas FAILED
> log-tk8.6.15.txt-4:==== font-47.2 FAILED
> log-tk8.6.15.txt-4:==== font-47.3 Bug 3049518 - Label FAILED
> log-tk8.6.15.txt-4:==== font-47.3 FAILED
> log-tk8.6.15.txt-4:==== frame-12.3 FrameWorldChanged procedure FAILED
> log-tk8.6.15.txt-4:==== frame-12.3 FAILED
> log-tk8.6.15.txt-4:==== textImage-4.2 alignment checking - baseline FAILED
> log-tk8.6.15.txt-4:==== textImage-4.2 FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 maxsize must be <= screen size
> FAILED
> log-tk8.6.15.txt-4:==== wm-maxsize-1.7 FAILED
>
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core
>