tcl-core — Tcl Core Team discussions about Tcl development

[TCLCORE] Ticket for subject A1 in Fossil Tk [Was: Proposal for my contribution to Tk, request for comments]

[elns <el...@xs...>]

A ticket was created in the fossil Tk repository to facilitate the development
of the following subject (of the earlier proposal in this thread):

     A.1 Collect utility procs

Link: https://core.tcl-lang.org/tk/info/718cbc3016

Development has started, and I will direct my future communications
regarding this project there.

News:
- an analysis of the occurrence of utility procs in the Tk test suite;
- a more detailed/concrete relocation plan for the utility procs;

As soon as the opportunity is right, I'll ask for explicit review there.
Regardless, Tk maintainers and developers are welcome to have a peek,
or comment on your own accord (of course!).

Regards,
Erik Leunissen
--

Re: [TCLCORE] Report of BiWeekly Tcl/Tk conference

[Harald O. <har...@el...>]

Sorry, Monday 23rd.
Harald

Am 16.12.2024 um 17:14 schrieb Paul Obermeier:
> Is the next meeting on Friday, 20th or Monday, 23rd?
> Paul
> 
> 
> Am 16.12.2024 um 15:36 schrieb Harald Oehlmann:
>> Dear TCL/Tk team,
>> thanks for the great meeting today. Here is a report.
>>
>> 1) celebrate great TCL/Tk 8.6.16 release !
>> 2) upcomming TCL/Tk 9.0.1 release
>> 2.1) changes.md update -> TIP 707 is not mentioned as it is internal 
>> change
>> -> Ashok adds TIP707
>>
>> 2.2) DKF's Enum work: is it ok on main and for 9.0.1. Should rewrites 
>> like this merged to 9.1 (when development branch open)?
>> https://core.tcl-lang.org/tcl/info/5611b610925a7a30
>>
>> -> Cosmetic change, easy to get typing errors in
>> -> Jan has tested and approved it for 9.0.1
>> -> Those cosmetic changes:
>>   - should first go into a branch for a 2nd eye (risk of error)
>>   - should not be done just before a release
>> Result: go to 9.0.1
>>
>> 2.3) namespace children subnamespace does not match exactly
>> https://core.tcl-lang.org/tcl/info/63449c0514
>> It fixes a real problem. But the fix feels critical as not fully 
>> understood. Nevertheless, a fix is a fix -> go to 9.0.1
>>
>> 2.4) zip bundling - ticket aaa84fbb -> eventual user guidance?
>> -> User guidense difficult as many combination possible
>>
>> 2.5) git/fossil checkin in build-info - tclUuid.h -> eventual user 
>> guidance?
>> -> all useful files are committed
>>
>> 2.6) bundled packages state: ITCL, Thread, TDBC
>> -> SQLite patches by Jan get partly upstream
>> -> may add a warning that ITCL may be unbundled-> ask community about use
>> -> ITCL will break with 9.1 with the long command argument TIP.
>> 2.7) Install/compile/package TCL 8.6 and 9.0 together/on same computer 
>> -> eventual user guidance?
>> -> 8.6,8.7 and 9.0 runs from the same directory.
>> -> development packages (tcl.h) are distinct.
>> -> man pages are distinct-> Torsten
>> -> nothing new with TCL8.5/8.6: but did not ship together.
>>
>> 2.8) Include recent Thread package bug fix in 9.0.1
>> https://core.tcl-lang.org/thread/info/f32864af
>>
>> 3) TIP 705 Licence
>> Vote is pending and expected for next year
>>
>> 4) Tk image copy performance  [e7fc29ce]: image copy optimization: 
>> test for alpha != 0/1 takes more time, than it optimizes.
>> Paul just remarks it but has no confidence in his patch. Only a branch 
>> and experimentation may show issues.
>>
>> 5) Renés Tk bindings proposals
>> Thanks for the core list message. This great input should not be lost, 
>> but is not for 9.0.1.
>> -> selection improvements: type beginning of a text to select -> 
>> requires a TIP to explain
>> -> bindings may conflict with user choice -> Tk 9.1.
>> -> or TkLib
>>
>> Thanks for all ! Remark that the next meeting is in one week: 20th of 
>> December 12:00 UTC.
>>
>> Take care,
>> Harald

Re: [TCLCORE] Report of BiWeekly Tcl/Tk conference

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

Is the next meeting on Friday, 20th or Monday, 23rd?
Paul


Am 16.12.2024 um 15:36 schrieb Harald Oehlmann:
> Dear TCL/Tk team,
> thanks for the great meeting today. Here is a report.
>
> 1) celebrate great TCL/Tk 8.6.16 release !
> 2) upcomming TCL/Tk 9.0.1 release
> 2.1) changes.md update -> TIP 707 is not mentioned as it is internal change
> -> Ashok adds TIP707
>
> 2.2) DKF's Enum work: is it ok on main and for 9.0.1. Should rewrites like this merged to 9.1 (when development branch open)?
> https://core.tcl-lang.org/tcl/info/5611b610925a7a30
>
> -> Cosmetic change, easy to get typing errors in
> -> Jan has tested and approved it for 9.0.1
> -> Those cosmetic changes:
>   - should first go into a branch for a 2nd eye (risk of error)
>   - should not be done just before a release
> Result: go to 9.0.1
>
> 2.3) namespace children subnamespace does not match exactly
> https://core.tcl-lang.org/tcl/info/63449c0514
> It fixes a real problem. But the fix feels critical as not fully understood. Nevertheless, a fix is a fix -> go to 9.0.1
>
> 2.4) zip bundling - ticket aaa84fbb -> eventual user guidance?
> -> User guidense difficult as many combination possible
>
> 2.5) git/fossil checkin in build-info - tclUuid.h -> eventual user guidance?
> -> all useful files are committed
>
> 2.6) bundled packages state: ITCL, Thread, TDBC
> -> SQLite patches by Jan get partly upstream
> -> may add a warning that ITCL may be unbundled-> ask community about use
> -> ITCL will break with 9.1 with the long command argument TIP.
> 2.7) Install/compile/package TCL 8.6 and 9.0 together/on same computer -> eventual user guidance?
> -> 8.6,8.7 and 9.0 runs from the same directory.
> -> development packages (tcl.h) are distinct.
> -> man pages are distinct-> Torsten
> -> nothing new with TCL8.5/8.6: but did not ship together.
>
> 2.8) Include recent Thread package bug fix in 9.0.1
> https://core.tcl-lang.org/thread/info/f32864af
>
> 3) TIP 705 Licence
> Vote is pending and expected for next year
>
> 4) Tk image copy performance  [e7fc29ce]: image copy optimization: test for alpha != 0/1 takes more time, than it optimizes.
> Paul just remarks it but has no confidence in his patch. Only a branch and experimentation may show issues.
>
> 5) Renés Tk bindings proposals
> Thanks for the core list message. This great input should not be lost, but is not for 9.0.1.
> -> selection improvements: type beginning of a text to select -> requires a TIP to explain
> -> bindings may conflict with user choice -> Tk 9.1.
> -> or TkLib
>
> Thanks for all ! Remark that the next meeting is in one week: 20th of December 12:00 UTC.
>
> Take care,
> Harald
>
>
>
>
>
>
>
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core

Re: [TCLCORE] Review of tcl_InitStubs documentation enhancements

[Harald O. <har...@el...>]

Don,
thanks for the message.

This is completly valuable for "package require TCL 8.6 9".

But Tcl_InitStubs does not support to specify multiple requirements.
That is one aim of the change I tried to clarify.

In consequence, "package require TCL 8.6 9" can only be implemented by 
two Tcl_InitStubs calls like:
if (NULL = Tcl_InitStubs("9.0")
     && NULL = Tcl_InitStubs("8.6")
) {
    return NULL
}

This may be detailed in the documentation to.

Thanks for the hint !
Harald

Am 16.12.2024 um 16:06 schrieb Donald Porter:
> Take care to understand the language of requirements.
> 
> The requirement “8.5-“ demands a minimum version of 8.5, but will accept anything greater.  It is good for permitting 9, so long as you recognize it also permits 10 and beyond.
> 
> In order to permit only 9 in addition, use [package require]’s capability to take multiple requirement arguments:
> 
> package require Tcl 8.5 9
> 
> I think most apps and extensions will prefer the latter, leaving support for Tcl 10 undetermined until it can be determined.
> 
> DGP
> 
> 
>> On Dec 16, 2024, at 9:55 AM, Harald Oehlmann <har...@el...> wrote:
>>
>> I would love to get feedback on the proposed documentation change for package and Tcl_InitStubs:
>> Branch:
>> https://core.tcl-lang.org/tcl/timeline?r=70f3b23cad-docs-version-requirement&c=2024-12-16+14%3A16%3A12
>> Ticket:
>> https://core.tcl-lang.org/tcl/info/70f3b23cad
>>
>> * Document extended requirement syntax for Tcl_InitStubs("8.5-") (remark the "-" for an min-unbound requirement).
>> * Document the different meaning without USE_TCL_STUBS (eventually not Tipped).
>> * Reorganize the package page to have a "requirement" section.
>>
>> Any comment welcome !
>> Harald

Re: [TCLCORE] Review of tcl_InitStubs documentation enhancements

[Donald P. <d.g...@co...>]

> In order to permit only 9 in addition, use [package require]’s capability to take multiple requirement arguments:
> 
> package require Tcl 8.5 9

You can achieve the same result with

package require Tcl 8.5-10

DGP

Re: [TCLCORE] Review of tcl_InitStubs documentation enhancements

[Donald P. <d.g...@co...>]

Take care to understand the language of requirements.

The requirement “8.5-“ demands a minimum version of 8.5, but will accept anything greater.  It is good for permitting 9, so long as you recognize it also permits 10 and beyond.

In order to permit only 9 in addition, use [package require]’s capability to take multiple requirement arguments:

package require Tcl 8.5 9

I think most apps and extensions will prefer the latter, leaving support for Tcl 10 undetermined until it can be determined.

DGP


> On Dec 16, 2024, at 9:55 AM, Harald Oehlmann <har...@el...> wrote:
> 
> I would love to get feedback on the proposed documentation change for package and Tcl_InitStubs:
> Branch:
> https://core.tcl-lang.org/tcl/timeline?r=70f3b23cad-docs-version-requirement&c=2024-12-16+14%3A16%3A12
> Ticket:
> https://core.tcl-lang.org/tcl/info/70f3b23cad
> 
> * Document extended requirement syntax for Tcl_InitStubs("8.5-") (remark the "-" for an min-unbound requirement).
> * Document the different meaning without USE_TCL_STUBS (eventually not Tipped).
> * Reorganize the package page to have a "requirement" section.
> 
> Any comment welcome !
> Harald
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core

[TCLCORE] Review of tcl_InitStubs documentation enhancements

[Harald O. <har...@el...>]

I would love to get feedback on the proposed documentation change for 
package and Tcl_InitStubs:
Branch:
https://core.tcl-lang.org/tcl/timeline?r=70f3b23cad-docs-version-requirement&c=2024-12-16+14%3A16%3A12
Ticket:
https://core.tcl-lang.org/tcl/info/70f3b23cad

* Document extended requirement syntax for Tcl_InitStubs("8.5-") (remark 
the "-" for an min-unbound requirement).
* Document the different meaning without USE_TCL_STUBS (eventually not 
Tipped).
* Reorganize the package page to have a "requirement" section.

Any comment welcome !
Harald

[TCLCORE] Report of BiWeekly Tcl/Tk conference

[Harald O. <har...@el...>]

Dear TCL/Tk team,
thanks for the great meeting today. Here is a report.

1) celebrate great TCL/Tk 8.6.16 release !
2) upcomming TCL/Tk 9.0.1 release
2.1) changes.md update -> TIP 707 is not mentioned as it is internal change
-> Ashok adds TIP707

2.2) DKF's Enum work: is it ok on main and for 9.0.1. Should rewrites 
like this merged to 9.1 (when development branch open)?
https://core.tcl-lang.org/tcl/info/5611b610925a7a30

-> Cosmetic change, easy to get typing errors in
-> Jan has tested and approved it for 9.0.1
-> Those cosmetic changes:
   - should first go into a branch for a 2nd eye (risk of error)
   - should not be done just before a release
Result: go to 9.0.1

2.3) namespace children subnamespace does not match exactly
https://core.tcl-lang.org/tcl/info/63449c0514
It fixes a real problem. But the fix feels critical as not fully 
understood. Nevertheless, a fix is a fix -> go to 9.0.1

2.4) zip bundling - ticket aaa84fbb -> eventual user guidance?
-> User guidense difficult as many combination possible

2.5) git/fossil checkin in build-info - tclUuid.h -> eventual user guidance?
-> all useful files are committed

2.6) bundled packages state: ITCL, Thread, TDBC
-> SQLite patches by Jan get partly upstream
-> may add a warning that ITCL may be unbundled-> ask community about use
-> ITCL will break with 9.1 with the long command argument TIP.
2.7) Install/compile/package TCL 8.6 and 9.0 together/on same computer 
-> eventual user guidance?
-> 8.6,8.7 and 9.0 runs from the same directory.
-> development packages (tcl.h) are distinct.
-> man pages are distinct-> Torsten
-> nothing new with TCL8.5/8.6: but did not ship together.

2.8) Include recent Thread package bug fix in 9.0.1
https://core.tcl-lang.org/thread/info/f32864af

3) TIP 705 Licence
Vote is pending and expected for next year

4) Tk image copy performance  [e7fc29ce]: image copy optimization: test 
for alpha != 0/1 takes more time, than it optimizes.
Paul just remarks it but has no confidence in his patch. Only a branch 
and experimentation may show issues.

5) Renés Tk bindings proposals
Thanks for the core list message. This great input should not be lost, 
but is not for 9.0.1.
-> selection improvements: type beginning of a text to select -> 
requires a TIP to explain
-> bindings may conflict with user choice -> Tk 9.1.
-> or TkLib

Thanks for all ! Remark that the next meeting is in one week: 20th of 
December 12:00 UTC.

Take care,
Harald

Re: [TCLCORE] [Ext] Tk enhancements

[Zaumseil R. <RZa...@kk...>]

Oh, please remove the "puts $w=$key" before 😉

Von: Zaumseil René via Tcl-Core <tcl...@li...>
Gesendet: Freitag, 13. Dezember 2024 14:42
An: Tcl Core List <tcl...@li...>
Betreff: [Ext] [TCLCORE] Tk enhancements

Hello

I use the following additional tk settings.

# ttk::entry entry scrolling content with mousewheel
bind Entry <MouseWheel> {tk::MouseWheel %W x %D -40.0 units}
bind TEntry <MouseWheel> {tk::MouseWheel %W x %D -40.0 units}

# Combobox enhancement: select item with typing
bind ComboboxListbox <KeyPress>     { ::ttk::combobox::Autoselect %W %A}
bind ComboboxListbox <Key-Delete>   { ::ttk::combobox::Autoselect %W {}}
bind ComboboxListbox <Key-BackSpace>    { ::ttk::combobox::Autoselect %W {}}
set ::ttk::combobox::Autoselect {}
proc ::ttk::combobox::Autoselect {w key} {
    puts $w=$key
    variable Autoselect
    if {$key eq {}} {
        set Autoselect [string range $Autoselect 0 end-1]
    } else {
        append Autoselect $key
    }
    $w selection clear 0 end
    if {$Autoselect eq {}} return
    set myIndex [lsearch -nocase -glob [[::ttk::combobox::LBMain $w] cget -values] $Autoselect*]
    if {$myIndex < 0} return
    $w activate $myIndex
    $w selection set $myIndex
    $w see $myIndex
}
proc ::ttk::combobox::LBCancel {lb} {
    Unpost [::ttk::combobox::LBMain $lb]
    set ::ttk::combobox::Autoselect {}
}


Would it be possible to add these to core tk?

Regards
Rene

Re: [TCLCORE] Fwd: Tcl/Tk 8.6.16 Release Candidates

[Stuart C. <exo...@ya...>]

On Friday, December 13, 2024 at 10:04:36 a.m. EST, Jan Nijtmans <jan...@gm...> wrote: 
That's why we should reverse the flow: start with "main" and
then cherry-pick to core-8-branch and core-8-6-branch from
there (eventually, when needed).

Agreed. Not sure why that's not being done now.


Stu

Re: [TCLCORE] Fwd: Tcl/Tk 8.6.16 Release Candidates

[Stuart C. <exo...@ya...>]

Great, thanks!


 On Friday, December 13, 2024 at 10:04:36 a.m. EST, Jan Nijtmans <jan...@gm...> wrote: 
Op vr 13 dec 2024 om 15:31 schreef Stuart Cassoff:
> I don't find problems - they find me. So if you want to be happy, don't let me look at your software. ;)
Please, continue looking! I'm happy!
 
>  That being said, in this commit:
> https://core.tcl-lang.org/tk/info/9404265d4f47cb01
> I think it would've been better to have the changes to generic/tkBind.c in a separate commit.
Well, it was a separate commit in 8.7 (and 8.6):
    <https://core.tcl-lang.org/tk/info/4dc57e41fe8b64d4>
but it was forward-merged together with another (harmless) fix.
That's why we should reverse the flow: start with "main" and
then cherry-pick to core-8-branch and core-8-6-branch from
there (eventually, when needed).

>  Also, now that I'm here, any chance of looking at https://core.tcl-lang.org/tcl/tktview/63449c0514 ?
> It's not a significant problem, but fixing things is always nice. :)

Now committed to a branch:
    <https://core.tcl-lang.org/tcl/info/ee821ec00fb0f615>

Hope this helps,
      Jan Nijtmans

Re: [TCLCORE] TCL/Tk bi-weekly telco on 2024-12-16 at 12:00 UTC

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

Op vr 13 dec 2024 om 12:43 schreef Harald Oehlmann:

> Proposed topics:
>
> * celebrate great TCL/Tk 8.6.16 release !
> * upcomming TCL/Tk 9.0.1 release
> * changes.md update -> TIP 707 is not mentioned as it is internal change
> * zip bundling - ticket aaa84fbb -> eventual user guidance?
> * git/fossil checkin in build-info - tclUuid.h -> eventual user guidance?
> * bundled packages state: ITCL, Thread, TDBC
> * Install/compile/package TCL 8.6 and 9.0 together/on same computer ->
> eventual user guidance?
> * AOB
>

Can we discuss this ticket?:
    <https://core.tcl-lang.org/tcl/info/63449c0514>
Should it go into 9.0.1 or not?

Regards,
    Jan Nijtmans

[TCLCORE] Tk enhancements

[Zaumseil R. <RZa...@kk...>]

Hello

I use the following additional tk settings.

# ttk::entry entry scrolling content with mousewheel
bind Entry <MouseWheel> {tk::MouseWheel %W x %D -40.0 units}
bind TEntry <MouseWheel> {tk::MouseWheel %W x %D -40.0 units}

# Combobox enhancement: select item with typing
bind ComboboxListbox <KeyPress>     { ::ttk::combobox::Autoselect %W %A}
bind ComboboxListbox <Key-Delete>   { ::ttk::combobox::Autoselect %W {}}
bind ComboboxListbox <Key-BackSpace>    { ::ttk::combobox::Autoselect %W {}}
set ::ttk::combobox::Autoselect {}
proc ::ttk::combobox::Autoselect {w key} {
    puts $w=$key
    variable Autoselect
    if {$key eq {}} {
        set Autoselect [string range $Autoselect 0 end-1]
    } else {
        append Autoselect $key
    }
    $w selection clear 0 end
    if {$Autoselect eq {}} return
    set myIndex [lsearch -nocase -glob [[::ttk::combobox::LBMain $w] cget -values] $Autoselect*]
    if {$myIndex < 0} return
    $w activate $myIndex
    $w selection set $myIndex
    $w see $myIndex
}
proc ::ttk::combobox::LBCancel {lb} {
    Unpost [::ttk::combobox::LBMain $lb]
    set ::ttk::combobox::Autoselect {}
}


Would it be possible to add these to core tk?

Regards
Rene

Re: [TCLCORE] Fwd: Tcl/Tk 8.6.16 Release Candidates

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

Op vr 13 dec 2024 om 15:31 schreef Stuart Cassoff:

> I don't find problems - they find me. So if you want to be happy, don't
> let me look at your software. ;)
>
Please, continue looking! I'm happy!


> That being said, in this commit:
> https://core.tcl-lang.org/tk/info/9404265d4f47cb01
> I think it would've been better to have the changes to generic/tkBind.c in
> a separate commit.

Well, it was a separate commit in 8.7 (and 8.6):
    <https://core.tcl-lang.org/tk/info/4dc57e41fe8b64d4>
but it was forward-merged together with another (harmless) fix.
That's why we should reverse the flow: start with "main" and
then cherry-pick to core-8-branch and core-8-6-branch from
there (eventually, when needed).

Also, now that I'm here, any chance of looking at
> https://core.tcl-lang.org/tcl/tktview/63449c0514 ?
> It's not a significant problem, but fixing things is always nice. :)
>

Now committed to a branch:
    <https://core.tcl-lang.org/tcl/info/ee821ec00fb0f615>

Hope this helps,
      Jan Nijtmans

Re: [TCLCORE] Fwd: Tcl/Tk 8.6.16 Release Candidates

[Stuart C. <exo...@ya...>]

I don't find problems - they find me. So if you want to be happy, don't let me look at your software. ;)

That being said, in this commit:
https://core.tcl-lang.org/tk/info/9404265d4f47cb01
I think it would've been better to have the changes to generic/tkBind.c in a separate commit.

Also, now that I'm here, any chance of looking at https://core.tcl-lang.org/tcl/tktview/63449c0514 ?
It's not a significant problem, but fixing things is always nice. :)


Stu

On Thursday, December 12, 2024 at 11:32:16 a.m. EST, Jan Nijtmans <jan...@gm...> wrote: 
Op di 10 dec 2024 om 21:28 schreef Jan Nijtmans:> Op di 10 dec 2024 21:10 schreef Stuart Cassoff:
>> Tk 8.6.16 rc0 ships with tkUuid.h and rc2 doesn't. Which is correct?
> 
> rc2 is correct. tKUuid.h is supposed to be built from manifest.uuid

This (however minor) problem won't happen again:
    <https://core.tcl-lang.org/tk/info/9404265d4f47cb01>

Thanks, Stuart, for bringing this to our attention!

Regards,
      Jan Nijtmans

[TCLCORE] TCL/Tk bi-weekly telco on 2024-12-16 at 12:00 UTC

[Harald O. <har...@el...>]

Dear TCL/Tk team !

please allow me to invite you to the next Tcl/Tk biweekly telco:
2024-12-16 12:00 UTC on jitzi: https://meet.jit.si/TclMonthlyMeetup

The main goal is to support the release manager:

Proposed topics:

* celebrate great TCL/Tk 8.6.16 release !
* upcomming TCL/Tk 9.0.1 release
* changes.md update -> TIP 707 is not mentioned as it is internal change
* zip bundling - ticket aaa84fbb -> eventual user guidance?
* git/fossil checkin in build-info - tclUuid.h -> eventual user guidance?
* bundled packages state: ITCL, Thread, TDBC
* Install/compile/package TCL 8.6 and 9.0 together/on same computer -> 
eventual user guidance?
* AOB

Please also note the following upcoming telcos:

Biweekly Tcl/Tk:
- 2024-12-23 12:00 UTC
- 2025-01-20 12:00 UTC
   (This is late due to my non-availability. If others organize, it may 
be one week earlier)

Monthly Tk: (1st Tuesday of each month)
- 2025-01-07 18:00 UTC
   Note: the German speaking meeting will directly follow. 
German/Austrian/Swiss/Belgium beer required for participation ;-).

Monthly TCL meetup (2nd Tuesday of each month)
- 2024-01-14 3:00 UTC (to be confirmed)

The meetings are for anybody, not only TCT members.

Thank you for all,
Harald

Re: [TCLCORE] Fwd: Tcl/Tk 8.6.16 Release Candidates

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

Op di 10 dec 2024 om 21:28 schreef Jan Nijtmans:

> Op di 10 dec 2024 21:10 schreef Stuart Cassoff:
>
>> Tk 8.6.16 rc0 ships with tkUuid.h and rc2 doesn't. Which is correct?
>>
>
> rc2 is correct. tKUuid.h is supposed to be built from manifest.uuid
>

This (however minor) problem won't happen again:
    <https://core.tcl-lang.org/tk/info/9404265d4f47cb01>

Thanks, Stuart, for bringing this to our attention!

Regards,
      Jan Nijtmans

Re: [TCLCORE] Sample Extension

[Harald O. <har...@el...>]

Ashok,

thank you, I appreciate.

I have already stopped my actions with the sample extension.

The build-info was not introduced by me. But if this is optional and so 
dependent on whatever, I would strongly urge the original author to 
remove it from the sample extension. It looks like a general solution 
which should always be present.

I also thought to put the command in a namespace. But this also adds 
complexity, so it should not be added.

I will close all my tickets and branches.
Sorry for all the noise!

Thank you all!,
Harald

Am 12.12.2024 um 09:11 schrieb apn...@ya...:
> Harald,
> 
>  From my personal perspective, neither of these is necessary.
> 
> There is no mandate for an extension to have a build-info command. If it wants to have one, it can use whatever fits its build environment (e.g. hgid) or completely leave it out. Note even Tcl's build-info command is undocumented.
> 
> The unload command is fraught with too many perils imho to get right in the general case and also of limited value.
> 
> I would not clutter the sample extension with either of these but of course your opinion may differ.
> 
> /Ashok
> 
> -----Original Message-----
> From: Harald Oehlmann <har...@el...>
> Sent: Tuesday, December 10, 2024 11:34 PM
> To: tcl...@li...
> Subject: [TCLCORE] Sample Extension
> 
> Dear TCL team,
> the sample extension as guidence needs some love.
> 
> https://core.tcl-lang.org/sampleextension/timeline
> 
> It is a complicated subject and it is getting more challenging.
> 
> Here are some thoughts:
> 
> 1) build-info
> 
> To support build-info, many lines were added with a lot of preprocessor
> relative code. It is hard to understand and fills the source code with a
> lot of interesting information.
> I don't think, anybody will change any thing here.
> And it looks unmaintainable in the sense, that if something is changed,
> the source has to be changed.
> It would be great to have this distributed with TCL in tcl.h or
> somewhere else and no package author has to maintain it.
> 
> 2) multiple interpreters and threads
> 
> The sample application uses static variables and thus does not support
> multiple interpreters. IMHO, this is not ok.
> 
> A solution using command client data is contained in the
> "19630c0c-unload" branch.
> It might be added as comments, that command client data may not be used,
> if the data is required outside of a command like in a event handler.
> Then, interpreter associative data has to be used.
> 
> 3) unload support
> 
> IMHO, TCL is very dynamic and unloading of a DLL is crucial.
> This is currently missing.
> A proposal is in the "19630c0c-unload" branch.
> Thanks for the proposal by Emiliano on c.l.t.
> It is so complicated. Interpreter associative data is required to hold
> the tolkens of the created commands.
> 
> -END-
> 
> I think, all this should be shown within the sample application.
> It is better to know bad news and there are a lot !
> 
> Any comments welcome !
> 
> Thank you for all !
> Harald

Re: [TCLCORE] Sample Extension

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

Harald,

>From my personal perspective, neither of these is necessary.

There is no mandate for an extension to have a build-info command. If it wants to have one, it can use whatever fits its build environment (e.g. hgid) or completely leave it out. Note even Tcl's build-info command is undocumented.

The unload command is fraught with too many perils imho to get right in the general case and also of limited value.

I would not clutter the sample extension with either of these but of course your opinion may differ.

/Ashok

-----Original Message-----
From: Harald Oehlmann <har...@el...> 
Sent: Tuesday, December 10, 2024 11:34 PM
To: tcl...@li...
Subject: [TCLCORE] Sample Extension

Dear TCL team,
the sample extension as guidence needs some love.

https://core.tcl-lang.org/sampleextension/timeline

It is a complicated subject and it is getting more challenging.

Here are some thoughts:

1) build-info

To support build-info, many lines were added with a lot of preprocessor 
relative code. It is hard to understand and fills the source code with a 
lot of interesting information.
I don't think, anybody will change any thing here.
And it looks unmaintainable in the sense, that if something is changed, 
the source has to be changed.
It would be great to have this distributed with TCL in tcl.h or 
somewhere else and no package author has to maintain it.

2) multiple interpreters and threads

The sample application uses static variables and thus does not support 
multiple interpreters. IMHO, this is not ok.

A solution using command client data is contained in the 
"19630c0c-unload" branch.
It might be added as comments, that command client data may not be used, 
if the data is required outside of a command like in a event handler.
Then, interpreter associative data has to be used.

3) unload support

IMHO, TCL is very dynamic and unloading of a DLL is crucial.
This is currently missing.
A proposal is in the "19630c0c-unload" branch.
Thanks for the proposal by Emiliano on c.l.t.
It is so complicated. Interpreter associative data is required to hold 
the tolkens of the created commands.

-END-

I think, all this should be shown within the sample application.
It is better to know bad news and there are a lot !

Any comments welcome !

Thank you for all !
Harald

Re: [TCLCORE] tcl.h of Tcl 8.6.16 make nmakehlp generating wrong version

[Eric B. <eri...@gm...>]

OK thanks.... this extension (tcl-opencv) forces the use of its own rules.vc
in rules-ext.vc by forcing _RULES_DIR=.
Removing this line, I don't have the error anymore, but it just lose the
rules to build .cpp. I know how to fix that, thanks.

Eric



Le mer. 11 déc. 2024 à 23:35, Jan Nijtmans <jan...@gm...> a
écrit :

> Op wo 11 dec 2024 om 23:04 schreef Eric Boudaillier:
>
>> Hi,
>>
>> Sorry to see that only now, but after struggling building an extension
>> against 8.6.16...
>> I just saw that it is due to the tcl.h format of TCL_MAJOR_VERSION:
>>
>> In 8.6.13:
>>
>> #define TCL_MAJOR_VERSION   8
>>
>> In 8.6.16:
>>
>> #if !defined(TCL_MAJOR_VERSION)
>> #   define TCL_MAJOR_VERSION   8
>> #endif
>> #if TCL_MAJOR_VERSION != 8
>> #   error "This header-file is for Tcl 8 only"
>> #endif
>>
>> nmakehlp failed to generate TCL_MAJOR_VERSION correctly in version.vc
>> due to the version check in header.
>>
>
>
> That means that this extension didn't pick up the latest
> version of  rules.vc.  This change was already done 2 years ago:
>      <https://core.tcl-lang.org/tcl/info/86b4decffb2c92a5>
>
> Hope this helps,
>      Jan Nijtmans
>
>

Re: [TCLCORE] tcl.h of Tcl 8.6.16 make nmakehlp generating wrong version

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

Op wo 11 dec 2024 om 23:04 schreef Eric Boudaillier:

> Hi,
>
> Sorry to see that only now, but after struggling building an extension
> against 8.6.16...
> I just saw that it is due to the tcl.h format of TCL_MAJOR_VERSION:
>
> In 8.6.13:
>
> #define TCL_MAJOR_VERSION   8
>
> In 8.6.16:
>
> #if !defined(TCL_MAJOR_VERSION)
> #   define TCL_MAJOR_VERSION   8
> #endif
> #if TCL_MAJOR_VERSION != 8
> #   error "This header-file is for Tcl 8 only"
> #endif
>
> nmakehlp failed to generate TCL_MAJOR_VERSION correctly in version.vc due
> to the version check in header.
>


That means that this extension didn't pick up the latest
version of  rules.vc.  This change was already done 2 years ago:
     <https://core.tcl-lang.org/tcl/info/86b4decffb2c92a5>

Hope this helps,
     Jan Nijtmans

[TCLCORE] tcl.h of Tcl 8.6.16 make nmakehlp generating wrong version

[Eric B. <eri...@gm...>]

Hi,

Sorry to see that only now, but after struggling building an extension
against 8.6.16...
I just saw that it is due to the tcl.h format of TCL_MAJOR_VERSION:

In 8.6.13:

#define TCL_MAJOR_VERSION   8

In 8.6.16:

#if !defined(TCL_MAJOR_VERSION)
#   define TCL_MAJOR_VERSION   8
#endif
#if TCL_MAJOR_VERSION != 8
#   error "This header-file is for Tcl 8 only"
#endif

nmakehlp failed to generate TCL_MAJOR_VERSION correctly in version.vc due
to the version check in header.

Eric

[TCLCORE] Tcl/Tk 8.6.16 RELEASED

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

Tcl/Tk 8.6.16 Release Announcement
December 11, 2024

The Tcl Core Team is pleased to announce the 8.6.16 releases of the Tcl
dynamic language and the Tk toolkit.  This is the sixteenth patch release
of Tcl/Tk 8.6.  More details can be found below.

We would like to express our gratitude to all those who submit bug
reports and patches.  This information is invaluable in enabling us
to identify and eliminate problems in the core. Such reports can be
submitted here.

         https://core.tcl-lang.org/tcl/ticket
         https://core.tcl-lang.org/tk/ticket

We ask that you log in (anonymous if you wish) to create tickets.
This deters abuse of the ticketing system:

         https://core.tcl-lang.org/tcl/login
         https://core.tcl-lang.org/tcl/login

Where to get the new releases:
------------------------------

Tcl/Tk 8.6.16 sources are freely available as open source from the Tcl
SourceForge project's file distribution area:

         http://sourceforge.net/projects/tcl/files/

This distribution is source code only.  We keep links to some third
parties offering pre-built binaries for various systems here:

         http://www.tcl-lang.org/software/tcltk/bindist.html

For additional information:
---------------------------

Please visit the Tcl Developer Xchange web site:

         http://www.tcl-lang.org/

This site contains a variety of information about Tcl/Tk in general, the
core Tcl and Tk distributions, Tcl development tools, and much more.

Summary of Changes since Tcl/Tk 8.6.15:
--------------------------------------

This is a patch release, so it primarily includes bug fixes and corrections
to erratic behavior.  Highlighted changes are noted below.  The changes file
at the root of the source tree contains a more complete list.  The Timelines
of all changes are online.

     http://core.tcl-lang.org/tcl/timeline
     http://core.tcl-lang.org/tk/timeline

    * Regression in [clock] timezones due to revised tzdata format

    * Improper startup if [pwd] contains a file named init.tcl

    * Fix crashes or hangs in...
       - TclOO + coroutine, oo-1.25
       - lifecycle management of the attributes of a menu entry
       - [grid] and [pack] handling of half-dead argument
       - Tk_DeleteErrorHandler()
       - overwrite of thread data by Tk initialization in second interp

    * Prevent negative zlib stream checksums, zlib-15.1

    * Filesystem path efficiency from skipping unnecessary normalization

    * Revised [clock scan] consistent with leap second timestamp validation

    * Updated bundled packages, libraries, standards, data
       - Itcl 4.3.2
       - sqlite3 3.47.2
       - Thread 2.8.11
       - TDBC* 1.1.10
       - tcltest 2.5.9
       - tzdata 2024b, corrected

--
Tcl Core Team and Maintainers
Don Porter, Tcl Core Release Manager

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

Re: [TCLCORE] Fwd: Tcl/Tk 8.6.16 Release Candidates

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

Op di 10 dec 2024 21:10 schreef Stuart Cassoff:

> Tk 8.6.16 rc0 ships with tkUuid.h and rc2 doesn't. Which is correct?
>

rc2 is correct. tKUuid.h is supposed to be built from manifest.uuid

Hope this helps,
        Jan Nijtmans

>

Re: [TCLCORE] Fwd: Tcl/Tk 8.6.16 Release Candidates

[Stuart C. <exo...@ya...>]

Tk 8.6.16 rc0 ships with tkUuid.h and rc2 doesn't. Which is correct?


Stu





On Tuesday, December 10, 2024 at 12:43:05 p.m. EST, Donald G Porter via Tcl-Core <tcl...@li...> wrote: 






Now available at

https://sourceforge.net/projects/tcl/files/Tcl/8.6.16/

are RC2 candidate source code distribution pre-releases of Tcl 8.6.16 and Tk 8.6.16.

Unless some serious blocking issue is reported, it is intended to promote these
candidates to the releases on December 11.

The Tcl pre-release includes pre-releases of the packages Thread 2.8.11, Itcl 4.3.2,
and TDBC* 1.1.10.  The released package sqlite3 3.47.2 is also included.

Compared to earlier pre-releases, one test in Tcl has been made more portable,
the bundled version of sqlite3 has been updated, and release dates have been
advanced.


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