tcl-core Mailing List for Tcl

Am 04.06.2026 um 16:10 schrieb apnmbx-public--- via Tcl-Core:
> I’m making a couple of (minor, I think) changes to the nmake build system
> 
>   * making the build less verbose, so it does not output so much trace
>     information during builds that hides real issues. Add “verbose” to
>     OPTS to get back the old behavior.
>   * more important, static libraries for **packages** are now installed
>     by default into the same location as Tcl’s static libraries. That
>     makes much more sense to me for building applications that link
>     statically and reflects Unix conventions.
> 
> Hope no one objects. Probably, no one other than Harald or Phil will 
> care but just in case... :-)
> 
> With the above and TIP 755, I’m very close to be being able to build a 
> fully statically linked tcl and wish shells with sqlite, thread, twapi 
> etc. No more writing DLL’s to disk and cleaning up. The downside to 
> single file executables – Microsoft Defender flags the build as some 
> Bear!xx CVE virus. Used to happen  with tclkits in 8.6 as well. Not sure 
> what the solution is there.
Thanks, I appreciate!

Yes, the false positives of the virus engines are an issue.

I normally only have those by fsecure for tcl8.6 starkits...

Thanks for all,
Harald

I'm making a couple of (minor, I think) changes to the nmake build system

 

*	making the build less verbose, so it does not output so much trace
information during builds that hides real issues. Add "verbose" to OPTS to
get back the old behavior.
*	more important, static libraries for *packages* are now installed by
default into the same location as Tcl's static libraries. That makes much
more sense to me for building applications that link statically and reflects
Unix conventions.

 

Hope no one objects. Probably, no one other than Harald or Phil will care
but just in case... :-)

 

With the above and TIP 755, I'm very close to be being able to build a fully
statically linked tcl and wish shells with sqlite, thread, twapi etc. No
more writing DLL's to disk and cleaning up. The downside to single file
executables - Microsoft Defender flags the build as some Bear!xx CVE virus.
Used to happen  with tclkits in 8.6 as well. Not sure what the solution is
there.

 

/Ashok

 

Christian has provided a fix for the syntax.
IMHO, the branch may be merged with an incremented version number.

I have no idea where the version number for pckIndex.tcl is. It might be 
in "configure.ac". Before, it was in "configure.in", but this apparently 
changed or I do not remember.

Thanks for review and help,
Harald

Am 03.06.2026 um 16:04 schrieb Harald Oehlmann:
> Its me again, the agent of the Wizards.
> 2 years ago, I got the task from Wizard Christian to implement the dict 
> optimization into tdbc::sqlite3:
> 
> https://core.tcl-lang.org/tdbcsqlite3/info/0374f302f59cb71f
> 
> Last status was to handle both, arrays and dict, in dependency of the 
> sqlite3 version.
> 
> This is now implemented.
> 
> The implementation is BAD.
> I don't know Tcl-OO and already a command
> 
>      variable -set {*}{
>      -columns -db -needcolumns -resultDict
>      -results -sql -Cursor -RowCount -END}
> 
> is highly strange and the "variable" documentation does not give any 
> idea. Ashoks book helps but it is 100 pages to read to understand this...
> 
> So, any help appreciated. Just look to the dump changes and correct them.
> 
> Thanks for all,
> Harald

Hello Harald

The oo::define man page says:

variable ?-slotOperation? ?name ...?

and your command can also be written as:

variable -set -columns -db -needcolumns -resultDict -results -sql -Cursor -RowCount -END

The -set is the slotOperation and all other args are variable names.

HTH
Rene

-----Ursprüngliche Nachricht-----
Von: Harald Oehlmann <har...@el...> 
Gesendet: Mittwoch, 3. Juni 2026 16:04
An: Tcl Core List <tcl...@li...>
Betreff: ***CAUTION_Invalid_Partial_Signature*** [Ext] [TCLCORE] tdbc::sqlite3 with internal dict processing

Its me again, the agent of the Wizards.
2 years ago, I got the task from Wizard Christian to implement the dict optimization into tdbc::sqlite3:

https://core.tcl-lang.org/tdbcsqlite3/info/0374f302f59cb71f

Last status was to handle both, arrays and dict, in dependency of the
sqlite3 version.

This is now implemented.

The implementation is BAD.
I don't know Tcl-OO and already a command

     variable -set {*}{
	-columns -db -needcolumns -resultDict
	-results -sql -Cursor -RowCount -END}

is highly strange and the "variable" documentation does not give any idea. Ashoks book helps but it is 100 pages to read to understand this...

So, any help appreciated. Just look to the dump changes and correct them.

Thanks for all,
Harald

Its me again, the agent of the Wizards.
2 years ago, I got the task from Wizard Christian to implement the dict 
optimization into tdbc::sqlite3:

https://core.tcl-lang.org/tdbcsqlite3/info/0374f302f59cb71f

Last status was to handle both, arrays and dict, in dependency of the 
sqlite3 version.

This is now implemented.

The implementation is BAD.
I don't know Tcl-OO and already a command

     variable -set {*}{
	-columns -db -needcolumns -resultDict
	-results -sql -Cursor -RowCount -END}

is highly strange and the "variable" documentation does not give any 
idea. Ashoks book helps but it is 100 pages to read to understand this...

So, any help appreciated. Just look to the dump changes and correct them.

Thanks for all,
Harald

TIP #755 is Final and merged.

 

Vote-Summary:  Accepted 6/0/0

Votes-For:  AK, AN, HO, KW, JN, MC, SL

 

Thanks to all those who voted and/or reviewed.

 

/Ashok

 

 

 

From: apnmbx-public--- via Tcl-Core <tcl...@li...> 
Sent: Tuesday, May 26, 2026 9:28 PM
To: 'Harald Oehlmann' <har...@el...>; tcl...@li...
Subject: Re: [TCLCORE] CFV TIP #755 - Add a callback for post-initialization of an interpreter

 

Indeed I forgot, Harald :-)

 

TIP #755: YES

 

Also, proof of concept - single file fully statically linked tclsh with sqlite and thread - tclBIMain.c <https://core.tcl-lang.org/tcl/file?name=generic/tclBIMain.c&ci=6f52e964ecb81c4c>  (without having to modify Tcl sources and making sqlite3 and thread::* available in all interpreters across all threads)

 

-----Original Message-----
From: Harald Oehlmann <har...@el... <mailto:har...@el...> > 
Sent: Tuesday, May 26, 2026 2:14 PM
To: tcl...@li... <mailto:tcl...@li...> 
Subject: Re: [TCLCORE] CFV TIP #755 - Add a callback for post-initialization of an interpreter

 

Am 26.05.2026 um 03:10 schrieb apnmbx-public--- via Tcl-Core:

> This is a CFV for TIP 755: Add a callback for post-initialization of an 

> interpreter < <https://core.tcl-lang.org/tips/doc/trunk/tip/755.md> https://core.tcl-lang.org/tips/doc/trunk/tip/755.md>.

> 

> Voting closes Tuesday June 2 00:00 UTC.

> 

> I have added some use cases <https://core.tcl-lang.org/tips/ 

> forumpost/7779d09116f01b4d> to the Rationale.

> 

> /Ashok

 

Harald: yes

 

@Ashok: you did not vote in your CFV E-Mail...

On Mon, Jun 1, 2026, 2:05 PM Paul Obermeier <pa...@po...> wrote:

> Well, macOS may return whatever color it means is useful.
> It should just be the same when using Tk 8.6.18, 9.0.3 or 9.0.4.
>

I am not so sure about that. There were changes in Tk 9 to make it more
consistent in how colorspaces are used.  Tk 9 is allowed to make changes
that are not compatible with Tk 8.

- Marc



> Paul
>
> Am 31.05.2026 um 15:02 schrieb Marc Culler:
>
> I can confirm that the script prints 0 0 245 on an m3 macbook air running
> macOS 26.
>
> However, that result is not exactly wrong, or not in the way you might
> guess.  The Apple ColorPicker.app also reports the color as (0, 0, 245).
> So Tk is reporting the actual color showing on the screen.  Perhaps this
> has something to do with which colorspace is being used when doing the
> rendering.
>
> - Marc
>
> On Sat, May 30, 2026 at 11:03 AM Paul Obermeier <pa...@po...> wrote:
>
>> Builds fine and Tcl/Tk test suites running successfully (with the usual 3
>> Windows exceptions) on the following platforms:
>> MacOS 15.7.1       M2      clang 17.0.0
>> Suse 15.6          x86_64  gcc 7.5.0
>> Debian ARM         arm64   gcc 12.2.0
>> Windows 11         x86     gcc 14.2.0
>> Windows 11         x86_64  gcc 14.2.0
>> Windows 11         x86_64  VS2022
>>
>> I noticed a strange behaviour using Tk 9.0.4 and the Img extension on
>> macOS:
>> The attached script draws a blue rectangle into a canvas and grabs the
>> canvas content into a photo image using the "window" format of Img.
>> With Tk 8.6.18 and 9.0.3 this returns the correct color value (0,0,255).
>> Using 9.0.4 it returns (0,0,245). Happens only on macOS.
>> Can anybody please confirm this behaviour?
>>
>> Regards,
>> Paul
>>
>> Am 29.05.2026 um 19:59 schrieb Donald G Porter via Tcl-Core:
>> >
>> > Now available at
>> >
>> > https://sourceforge.net/projects/tcl/files/Tcl/9.0.4
>> >
>> > are RC0 candidate source code distribution pre-releases of Tcl 9.0.4.
>> > These are the first candidate releases leading to the release of Tcl
>> 9.0.4
>> >
>> > The Tcl pre-release includes a pre-release of the package Thread 3.0.5.
>> > The released packages Itcl 4.3.7, TDBC* 1.1.13, and sqlite3 3.53.0 are
>> also included.
>> >
>> > Please test these candidates on your platforms and report any blocking
>> issues
>> > you discover.
>> >
>> > Thank you for your contributions and assistance.
>> >
>> _______________________________________________
>> Tcl-Core mailing list
>> Tcl...@li...
>> https://lists.sourceforge.net/lists/listinfo/tcl-core
>>
>
>
> _______________________________________________
> Tcl-Core mailing lis...@li...://lists.sourceforge.net/lists/listinfo/tcl-core
>
>
>

Well, macOS may return whatever color it means is useful.
It should just be the same when using Tk 8.6.18, 9.0.3 or 9.0.4.

Paul

Am 31.05.2026 um 15:02 schrieb Marc Culler:
> I can confirm that the script prints 0 0 245 on an m3 macbook air running macOS 26.
>
> However, that result is not exactly wrong, or not in the way you might guess.  The Apple ColorPicker.app also reports the color as (0, 0, 245).  So Tk is reporting the actual color showing on the screen.  Perhaps this has something to do with which colorspace is being used when doing the rendering.
>
> - Marc
>
> On Sat, May 30, 2026 at 11:03 AM Paul Obermeier <pa...@po...> wrote:
>
>     Builds fine and Tcl/Tk test suites running successfully (with the usual 3 Windows exceptions) on the following platforms:
>     MacOS 15.7.1       M2      clang 17.0.0
>     Suse 15.6          x86_64  gcc 7.5.0
>     Debian ARM         arm64   gcc 12.2.0
>     Windows 11         x86     gcc 14.2.0
>     Windows 11         x86_64  gcc 14.2.0
>     Windows 11         x86_64  VS2022
>
>     I noticed a strange behaviour using Tk 9.0.4 and the Img extension on macOS:
>     The attached script draws a blue rectangle into a canvas and grabs the
>     canvas content into a photo image using the "window" format of Img.
>     With Tk 8.6.18 and 9.0.3 this returns the correct color value (0,0,255).
>     Using 9.0.4 it returns (0,0,245). Happens only on macOS.
>     Can anybody please confirm this behaviour?
>
>     Regards,
>     Paul
>
>     Am 29.05.2026 um 19:59 schrieb Donald G Porter via Tcl-Core:
>     >
>     > Now available at
>     >
>     > https://sourceforge.net/projects/tcl/files/Tcl/9.0.4
>     >
>     > are RC0 candidate source code distribution pre-releases of Tcl 9.0.4.
>     > These are the first candidate releases leading to the release of Tcl 9.0.4
>     >
>     > The Tcl pre-release includes a pre-release of the package Thread 3.0.5.
>     > The released packages Itcl 4.3.7, TDBC* 1.1.13, and sqlite3 3.53.0 are also included.
>     >
>     > Please test these candidates on your platforms and report any blocking issues
>     > you discover.
>     >
>     > Thank you for your contributions and assistance.
>     >
>     _______________________________________________
>     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

TIP #687: YES

Thank you,
Harald

Am 01.06.2026 um 17:36 schrieb apnmbx-public--- via Tcl-Core:
> TIP #687: YES

TIP #687: YES

 

Thank you. I am always using latest = greatest ...
________________________________
From: apn...@ya... <apn...@ya...> on behalf of apn...@ya... <apn...@ya...>
Sent: Monday, June 1, 2026 2:49 PM
To: so...@te... <so...@te...>; 'Paul Obermeier' <pa...@po...>; 'Donald G Porter' <don...@ni...>; 'Tcl List Core' <tcl...@li...>
Subject: RE: [TCLCORE] Tcl 9.0.4 Release Candidates


Fixed in the thread repository now.



However, the 9.0.4 release candidate would not be having that check as it includes thread 3.0.5 and the check was added for 3.0.6. Were you using the release candidate or a thread repository checkout?



/Ashok



From: so...@te... <so...@te...>
Sent: Monday, June 1, 2026 1:34 AM
To: Paul Obermeier <pa...@po...>; Donald G Porter <don...@ni...>; Tcl List Core <tcl...@li...>
Subject: Re: [TCLCORE] Tcl 9.0.4 Release Candidates



The thread package now requires autoconf 2.73

Now the scripts fail because these distros have lower versions installed:



RHEL9:  2.69

Debian 12 Bookworm: 2.72

Ubuntu 22 Jammy: 2.71



I did not see any specific features used from this autoconf 2.73 version.



Could it be downgraded to 2.69 at least?



BR





________________________________

From: Paul Obermeier <pa...@po...<mailto:pa...@po...>>
Sent: Saturday, May 30, 2026 18:02
To: Donald G Porter <don...@ni...<mailto:don...@ni...>>; Tcl List Core <tcl...@li...<mailto:tcl...@li...>>
Subject: Re: [TCLCORE] Tcl 9.0.4 Release Candidates



Builds fine and Tcl/Tk test suites running successfully (with the usual 3 Windows exceptions) on the following platforms:
MacOS 15.7.1       M2      clang 17.0.0
Suse 15.6          x86_64  gcc 7.5.0
Debian ARM         arm64   gcc 12.2.0
Windows 11         x86     gcc 14.2.0
Windows 11         x86_64  gcc 14.2.0
Windows 11         x86_64  VS2022

I noticed a strange behaviour using Tk 9.0.4 and the Img extension on macOS:
The attached script draws a blue rectangle into a canvas and grabs the
canvas content into a photo image using the "window" format of Img.
With Tk 8.6.18 and 9.0.3 this returns the correct color value (0,0,255).
Using 9.0.4 it returns (0,0,245). Happens only on macOS.
Can anybody please confirm this behaviour?

Regards,
Paul

Am 29.05.2026 um 19:59 schrieb Donald G Porter via Tcl-Core:
>
> Now available at
>
> https://sourceforge.net/projects/tcl/files/Tcl/9.0.4
>
> are RC0 candidate source code distribution pre-releases of Tcl 9.0.4.
> These are the first candidate releases leading to the release of Tcl 9.0.4
>
> The Tcl pre-release includes a pre-release of the package Thread 3.0.5.
> The released packages Itcl 4.3.7, TDBC* 1.1.13, and sqlite3 3.53.0 are also included.
>
> Please test these candidates on your platforms and report any blocking issues
> you discover.
>
> Thank you for your contributions and assistance.
>

Fixed in the thread repository now.

 

However, the 9.0.4 release candidate would not be having that check as it
includes thread 3.0.5 and the check was added for 3.0.6. Were you using the
release candidate or a thread repository checkout?

 

/Ashok

 

From: so...@te... <so...@te...> 
Sent: Monday, June 1, 2026 1:34 AM
To: Paul Obermeier <pa...@po...>; Donald G Porter
<don...@ni...>; Tcl List Core <tcl...@li...>
Subject: Re: [TCLCORE] Tcl 9.0.4 Release Candidates

 

The thread package now requires autoconf 2.73

Now the scripts fail because these distros have lower versions installed:

 

RHEL9:  2.69

Debian 12 Bookworm: 2.72

Ubuntu 22 Jammy: 2.71

 

I did not see any specific features used from this autoconf 2.73 version.

 

Could it be downgraded to 2.69 at least?

 

BR

 

 

  _____  

From: Paul Obermeier <pa...@po... <mailto:pa...@po...> >
Sent: Saturday, May 30, 2026 18:02
To: Donald G Porter <don...@ni... <mailto:don...@ni...>
>; Tcl List Core <tcl...@li...
<mailto:tcl...@li...> >
Subject: Re: [TCLCORE] Tcl 9.0.4 Release Candidates 

 

Builds fine and Tcl/Tk test suites running successfully (with the usual 3
Windows exceptions) on the following platforms:
MacOS 15.7.1       M2      clang 17.0.0
Suse 15.6          x86_64  gcc 7.5.0
Debian ARM         arm64   gcc 12.2.0
Windows 11         x86     gcc 14.2.0
Windows 11         x86_64  gcc 14.2.0
Windows 11         x86_64  VS2022

I noticed a strange behaviour using Tk 9.0.4 and the Img extension on macOS:
The attached script draws a blue rectangle into a canvas and grabs the
canvas content into a photo image using the "window" format of Img.
With Tk 8.6.18 and 9.0.3 this returns the correct color value (0,0,255).
Using 9.0.4 it returns (0,0,245). Happens only on macOS.
Can anybody please confirm this behaviour?

Regards,
Paul

Am 29.05.2026 um 19:59 schrieb Donald G Porter via Tcl-Core:
>
> Now available at
>
> https://sourceforge.net/projects/tcl/files/Tcl/9.0.4
>
> are RC0 candidate source code distribution pre-releases of Tcl 9.0.4.
> These are the first candidate releases leading to the release of Tcl 9.0.4
>
> The Tcl pre-release includes a pre-release of the package Thread 3.0.5.
> The released packages Itcl 4.3.7, TDBC* 1.1.13, and sqlite3 3.53.0 are
also included.
>
> Please test these candidates on your platforms and report any blocking
issues
> you discover.
>
> Thank you for your contributions and assistance.
>

Dear Tcl/Tk team,

here is the report of the Tcl/Tk biweekly telco:
1t of June at 12:00 UTC
(one hour earlier than last meeting)
Duration: 1 hour
At: https://meet.jit.si/TclMonthlyMeetup

Top 1) Release calender (TIP 713)
Current release calender:
May	9.0.4

- 2 Issues
- Thread outdated: may be updated to 3.0.6 - fixes a memory leak - Don 
may decide, small change
- thread package autoconf now 2.73. RHEL9 uses 2.69 -> could it be 
downgraded?

   Ok the check for 2.73 is to harsh.
   Don't do any check and regenerate.
   https://core.tcl-lang.org/thread/info/1215387dc5183637

No other issues. 9.0.4 is in great state!

June 	9.1b0 	Tcl 9.1 Feature Freeze
Deadline 3rd of June, features are fixed.

Top 2) TIP 755: interpreter post-initialization callback
-> in vote
Top 3) TIP 754: Cache Composed Pixmaps for Ttk Image Elements
-> withdrawn and moved to tk ticket 7caf9e9e
Work in progress. Great work.

Top 4) TIP 752: Timer_API with long long
-> in vote
Great stuff - will be in.

Top 5) TIP 751: rotated text for label widgets
-> ok
Finished

Top 6) TIP 740: treeview enhancements
-> menu test failures - false positive - fixed

Top 7) TIP 687: locale support for word handling in text and entry
-> in vote
Great stuff - will be in.

TOP 8) Tk bidi support
Great work - refinement
Configure warning on Linux X11: font config warning, using without "fcinit"
-> no idea why it appears.

Top 9) wayland
Far from finished - great work
It already builds - great milestone achieved
So multiple people can work on it

Does Wayland feature a remote display?
Apparently this was recently added.

Top 10) AOB

Top 11) next meetings
- German telco: 2nd of June at 20:00 EST
- monthly meetup: 9th of June 10:00 UTC
- biweekly telco: 15th of June 2026 12:00 UTC

Asking for presentations and help:
- present new treewidget

Top 12) OpenACS/Tcl/Tk Vienna Conference:
- 19th of June: tentative Program
- 30th of June: registration end
- 15th of June meet in Luftburg (location normally the conf dinner)

Thank you for all,
Harald

Op di 26 mei 2026 om 03:11 schreef Ashok:
>
> This is a CFV for TIP 755: Add a callback for post-initialization of an interpreter.

TIP #755:   Yes

Regards,
    Jan Nijtmans

Just FYI I will not be able to attend.

> On May 31, 2026, at 11:29 AM, Harald Oehlmann <har...@el...> wrote:
> 
> Dear Tcl/Tk team,
> 
> Please feel invited to the next Tcl/Tk biweekly telco:
> 1t of June at 12:00 UTC
> (one hour earlier than last meeting)
> Duration: 1 hour
> At: https://meet.jit.si/TclMonthlyMeetup
> 
> Agenda proposal:
> 
> Top 1) Release calender (TIP 713)
> Current release calender:
> May	9.0.4
> June 	9.1b0 	Tcl 9.1 Feature Freeze
> July Vacation month - no action
> August 	9.1b1
> September 	9.1.0, 9.0.5 	New stable release sequence
> October 	
> November 	9.1.1
> December 	9.0.6 	Final Tcl 9.0 release
> 
> Top 2) TIP 755: interpreter post-initialization callback
> -> in vote
> Top 3) TIP 754: Cache Composed Pixmaps for Ttk Image Elements
> -> withdrawn and moved to tk ticket 7caf9e9e
> Top 4) TIP 752: Timer_API with long long
> -> in vote
> Top 5) TIP 751: rotated text for label widgets
> -> ok
> Top 6) TIP 740: treeview enhancements
> -> menu test failures
> Top 7) TIP 687: locale support for word handling in text and entry
> -> in vote
> TOP 8) Tk bidi support
> Top 9) wayland
> Top 10) AOB
> Top 11) next meetings
> - German telco: 2nd of June at 20:00 EST
> - monthly meetup: 9th of June 10:00 UTC
> - biweekly telco: 15th of June 2026 12:00 UTC
> 
> Top 12) OpenACS/Tcl/Tk Vienna Conference:
> - 19th of June: tnetative Program
> - 30th of June: registration end
> - 15th of June meet in Luftburg (location normally the conf dinner)
> 
> Thank you for all,
> Harald
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core

The thread package now requires autoconf 2.73
Now the scripts fail because these distros have lower versions installed:

RHEL9:  2.69
Debian 12 Bookworm: 2.72
Ubuntu 22 Jammy: 2.71

I did not see any specific features used from this autoconf 2.73 version.

Could it be downgraded to 2.69 at least?

BR


________________________________
From: Paul Obermeier <pa...@po...>
Sent: Saturday, May 30, 2026 18:02
To: Donald G Porter <don...@ni...>; Tcl List Core <tcl...@li...>
Subject: Re: [TCLCORE] Tcl 9.0.4 Release Candidates

Builds fine and Tcl/Tk test suites running successfully (with the usual 3 Windows exceptions) on the following platforms:
MacOS 15.7.1       M2      clang 17.0.0
Suse 15.6          x86_64  gcc 7.5.0
Debian ARM         arm64   gcc 12.2.0
Windows 11         x86     gcc 14.2.0
Windows 11         x86_64  gcc 14.2.0
Windows 11         x86_64  VS2022

I noticed a strange behaviour using Tk 9.0.4 and the Img extension on macOS:
The attached script draws a blue rectangle into a canvas and grabs the
canvas content into a photo image using the "window" format of Img.
With Tk 8.6.18 and 9.0.3 this returns the correct color value (0,0,255).
Using 9.0.4 it returns (0,0,245). Happens only on macOS.
Can anybody please confirm this behaviour?

Regards,
Paul

Am 29.05.2026 um 19:59 schrieb Donald G Porter via Tcl-Core:
>
> Now available at
>
> https://sourceforge.net/projects/tcl/files/Tcl/9.0.4
>
> are RC0 candidate source code distribution pre-releases of Tcl 9.0.4.
> These are the first candidate releases leading to the release of Tcl 9.0.4
>
> The Tcl pre-release includes a pre-release of the package Thread 3.0.5.
> The released packages Itcl 4.3.7, TDBC* 1.1.13, and sqlite3 3.53.0 are also included.
>
> Please test these candidates on your platforms and report any blocking issues
> you discover.
>
> Thank you for your contributions and assistance.
>

Dear Tcl/Tk team,

Please feel invited to the next Tcl/Tk biweekly telco:
1t of June at 12:00 UTC
(one hour earlier than last meeting)
Duration: 1 hour
At: https://meet.jit.si/TclMonthlyMeetup

Agenda proposal:

Top 1) Release calender (TIP 713)
Current release calender:
May	9.0.4
June 	9.1b0 	Tcl 9.1 Feature Freeze
July Vacation month - no action
August 	9.1b1
September 	9.1.0, 9.0.5 	New stable release sequence
October 	
November 	9.1.1
December 	9.0.6 	Final Tcl 9.0 release

Top 2) TIP 755: interpreter post-initialization callback
-> in vote
Top 3) TIP 754: Cache Composed Pixmaps for Ttk Image Elements
-> withdrawn and moved to tk ticket 7caf9e9e
Top 4) TIP 752: Timer_API with long long
-> in vote
Top 5) TIP 751: rotated text for label widgets
-> ok
Top 6) TIP 740: treeview enhancements
-> menu test failures
Top 7) TIP 687: locale support for word handling in text and entry
-> in vote
TOP 8) Tk bidi support
Top 9) wayland
Top 10) AOB
Top 11) next meetings
- German telco: 2nd of June at 20:00 EST
- monthly meetup: 9th of June 10:00 UTC
- biweekly telco: 15th of June 2026 12:00 UTC

Top 12) OpenACS/Tcl/Tk Vienna Conference:
- 19th of June: tnetative Program
- 30th of June: registration end
- 15th of June meet in Luftburg (location normally the conf dinner)

Thank you for all,
Harald

I can confirm that the script prints 0 0 245 on an m3 macbook air running
macOS 26.

However, that result is not exactly wrong, or not in the way you might
guess.  The Apple ColorPicker.app also reports the color as (0, 0, 245).
So Tk is reporting the actual color showing on the screen.  Perhaps this
has something to do with which colorspace is being used when doing the
rendering.

- Marc

On Sat, May 30, 2026 at 11:03 AM Paul Obermeier <pa...@po...> wrote:

> Builds fine and Tcl/Tk test suites running successfully (with the usual 3
> Windows exceptions) on the following platforms:
> MacOS 15.7.1       M2      clang 17.0.0
> Suse 15.6          x86_64  gcc 7.5.0
> Debian ARM         arm64   gcc 12.2.0
> Windows 11         x86     gcc 14.2.0
> Windows 11         x86_64  gcc 14.2.0
> Windows 11         x86_64  VS2022
>
> I noticed a strange behaviour using Tk 9.0.4 and the Img extension on
> macOS:
> The attached script draws a blue rectangle into a canvas and grabs the
> canvas content into a photo image using the "window" format of Img.
> With Tk 8.6.18 and 9.0.3 this returns the correct color value (0,0,255).
> Using 9.0.4 it returns (0,0,245). Happens only on macOS.
> Can anybody please confirm this behaviour?
>
> Regards,
> Paul
>
> Am 29.05.2026 um 19:59 schrieb Donald G Porter via Tcl-Core:
> >
> > Now available at
> >
> > https://sourceforge.net/projects/tcl/files/Tcl/9.0.4
> >
> > are RC0 candidate source code distribution pre-releases of Tcl 9.0.4.
> > These are the first candidate releases leading to the release of Tcl
> 9.0.4
> >
> > The Tcl pre-release includes a pre-release of the package Thread 3.0.5.
> > The released packages Itcl 4.3.7, TDBC* 1.1.13, and sqlite3 3.53.0 are
> also included.
> >
> > Please test these candidates on your platforms and report any blocking
> issues
> > you discover.
> >
> > Thank you for your contributions and assistance.
> >
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core
>

Builds fine and Tcl/Tk test suites running successfully (with the usual 3 Windows exceptions) on the following platforms:
MacOS 15.7.1       M2      clang 17.0.0
Suse 15.6          x86_64  gcc 7.5.0
Debian ARM         arm64   gcc 12.2.0
Windows 11         x86     gcc 14.2.0
Windows 11         x86_64  gcc 14.2.0
Windows 11         x86_64  VS2022

I noticed a strange behaviour using Tk 9.0.4 and the Img extension on macOS:
The attached script draws a blue rectangle into a canvas and grabs the
canvas content into a photo image using the "window" format of Img.
With Tk 8.6.18 and 9.0.3 this returns the correct color value (0,0,255).
Using 9.0.4 it returns (0,0,245). Happens only on macOS.
Can anybody please confirm this behaviour?

Regards,
Paul

Am 29.05.2026 um 19:59 schrieb Donald G Porter via Tcl-Core:
>
> Now available at
>
> https://sourceforge.net/projects/tcl/files/Tcl/9.0.4
>
> are RC0 candidate source code distribution pre-releases of Tcl 9.0.4.
> These are the first candidate releases leading to the release of Tcl 9.0.4
>
> The Tcl pre-release includes a pre-release of the package Thread 3.0.5.
> The released packages Itcl 4.3.7, TDBC* 1.1.13, and sqlite3 3.53.0 are also included.
>
> Please test these candidates on your platforms and report any blocking issues
> you discover.
>
> Thank you for your contributions and assistance.
>

Tested on Win 10, VS 2022, x64 and x86. Configs shared, shared-nozipfs,
static.

Tcl + pkgs passes except for the usual broken testchmod tests and
unavailability of DB drivers.

Wish passes except for the usual clipboard failures and with one additional
failure, presumably benign, that showed up on two of the runs:

event-c3f5c85322

---- Result was:
|1(000)|2(001)|3(002)|4(010)|<Enter>
.|5(101)|6(012)|7(110)|8(111)|9(112)|10(120)|<Leave> .|<Enter>
.f|11(211)|<Leave> .f|<Enter> .|12(122)|<Leave> .|<Enter>
.f|13(220)|14(221)|15(222)|16(200)|<Leave> .f|<Leave> .|17(021)|<Enter>
.|<Enter> .f|18(202)|<Leave> .f|<Leave> .|
---- Result should have been (exact matching):
|1(000)|2(001)|3(002)|4(010)|<Enter> .|5(101)|<Leave> .|6(012)|<Enter>
.|7(110)|8(111)|9(112)|10(120)|<Leave> .|<Enter> .f|11(211)|<Leave>
.f|<Enter> .|12(122)|<Leave> .|<Enter>
.f|13(220)|14(221)|15(222)|16(200)|<Leave> .f|<Leave> .|17(021)|<Enter>
.|<Enter> .f|18(202)|<Leave> .f|<Leave> .|
==== event-c3f5c85322 FAILED

-----Original Message-----
From: Donald G Porter via Tcl-Core <tcl...@li...> 
Sent: Friday, May 29, 2026 11:30 PM
To: Tcl List Core <tcl...@li...>
Subject: [TCLCORE] Tcl 9.0.4 Release Candidates


Now available at

https://sourceforge.net/projects/tcl/files/Tcl/9.0.4

are RC0 candidate source code distribution pre-releases of Tcl 9.0.4.
These are the first candidate releases leading to the release of Tcl 9.0.4

The Tcl pre-release includes a pre-release of the package Thread 3.0.5.
The released packages Itcl 4.3.7, TDBC* 1.1.13, and sqlite3 3.53.0 are also
included.

Please test these candidates on your platforms and report any blocking
issues
you discover.

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

*	And if there is in fact a reason this is necessary, why only for core packages?

 

And come to think of it, even for core packages why is the check for safe interps only necessary for zipfs builds and not for the –disable-zipfs configuration?

 

Most curious...

 

From: apnmbx-public--- via Tcl-Core <tcl...@li...> 
Sent: Saturday, May 30, 2026 8:31 AM
To: 'Jan Nijtmans' <jan...@gm...>
Cc: 'Tcl Core List' <tcl...@li...>
Subject: Re: [TCLCORE] Utility of manifest.txt?

 

Thanks, Jan.

 

I don't really understand much of the rationale of the poster. For example, the statement 

 

... a manifest file that tells the interpreter, for clear and for certain, which packages may be loaded in a normal interpreter and which packages may be loaded in a safe interpreter.

 

This is the first time I’ve heard of a package loading mechanism needing to check whether the interpreter is safe or not. What for? Unsafe commands would be unavailable to the package anyways in safe interpreters. And if there is in fact a reason this is necessary, why only for core packages?

 

(Just thinking out aloud, the question is not directed towards you!)

 

Not an urgent matter but just ran into it in the course of building a static BI distribution. I’ll investigate further at some point and perhaps understand the reason for its existence.

 

/Ashok

 

-----Original Message-----
From: Jan Nijtmans <jan...@gm... <mailto:jan...@gm...> > 
Sent: Friday, May 29, 2026 9:10 PM
To: apn...@ya... <mailto:apn...@ya...> 
Cc: Tcl Core List <tcl...@li... <mailto:tcl...@li...> >
Subject: Re: [TCLCORE] Utility of manifest.txt?

 

Op vr 29 mei 2026 om 17:25 schreef apnmbx-public:

> 

> Any hints as to the reason for this duplication?

 

The complete discussion can be found here:

     < <https://core.tcl-lang.org/tcl/info/4e812aadae> https://core.tcl-lang.org/tcl/info/4e812aadae>

 

Hope this helps,

       Jan Nijtmans

The current head of Thread is 3.0.6. Is the inclusion of 3.0.5 intentional
or an oversight?

-----Original Message-----
From: Donald G Porter via Tcl-Core <tcl...@li...> 
Sent: Friday, May 29, 2026 11:30 PM
To: Tcl List Core <tcl...@li...>
Subject: [TCLCORE] Tcl 9.0.4 Release Candidates


Now available at

https://sourceforge.net/projects/tcl/files/Tcl/9.0.4

are RC0 candidate source code distribution pre-releases of Tcl 9.0.4.
These are the first candidate releases leading to the release of Tcl 9.0.4

The Tcl pre-release includes a pre-release of the package Thread 3.0.5.
The released packages Itcl 4.3.7, TDBC* 1.1.13, and sqlite3 3.53.0 are also
included.

Please test these candidates on your platforms and report any blocking
issues
you discover.

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

Thanks, Jan.

 

I don't really understand much of the rationale of the poster. For example, the statement 

 

... a manifest file that tells the interpreter, for clear and for certain, which packages may be loaded in a normal interpreter and which packages may be loaded in a safe interpreter.

 

This is the first time I’ve heard of a package loading mechanism needing to check whether the interpreter is safe or not. What for? Unsafe commands would be unavailable to the package anyways in safe interpreters. And if there is in fact a reason this is necessary, why only for core packages?

 

(Just thinking out aloud, the question is not directed towards you!)

 

Not an urgent matter but just ran into it in the course of building a static BI distribution. I’ll investigate further at some point and perhaps understand the reason for its existence.

 

/Ashok

 

-----Original Message-----
From: Jan Nijtmans <jan...@gm...> 
Sent: Friday, May 29, 2026 9:10 PM
To: apn...@ya...
Cc: Tcl Core List <tcl...@li...>
Subject: Re: [TCLCORE] Utility of manifest.txt?

 

Op vr 29 mei 2026 om 17:25 schreef apnmbx-public:

> 

> Any hints as to the reason for this duplication?

 

The complete discussion can be found here:

     < <https://core.tcl-lang.org/tcl/info/4e812aadae> https://core.tcl-lang.org/tcl/info/4e812aadae>

 

Hope this helps,

       Jan Nijtmans

Now available at

https://sourceforge.net/projects/tcl/files/Tcl/9.0.4

are RC0 candidate source code distribution pre-releases of Tk 9.0.4.
These are the first candidate releases leading to the release of Tk 9.0.4

Please test these candidates on your platforms and report any blocking issues
you discover.

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 |
|______________________________________________________________________|

Now available at

https://sourceforge.net/projects/tcl/files/Tcl/9.0.4

are RC0 candidate source code distribution pre-releases of Tcl 9.0.4.
These are the first candidate releases leading to the release of Tcl 9.0.4

The Tcl pre-release includes a pre-release of the package Thread 3.0.5.
The released packages Itcl 4.3.7, TDBC* 1.1.13, and sqlite3 3.53.0 are also included.

Please test these candidates on your platforms and report any blocking issues
you discover.

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 |
|______________________________________________________________________|