Peter P. wrote:

Dear Dan and everyone contributing to this issue,

I kinda felt this is a sensitive topic but was not aware of the amount of frustrating work that has already gone into it. Every bit of it is much appreciated, thank you and everyone a lot!

So if I would want to write a patch that uses the [key] and [keyup] objects to work consistently on all three platforms, I would have to tell the user to turn key repeat off in her operating system AND provide a means to filter out repeating key ons (Win and OS X), correct?

In this case I happily deal with the current status.

Often key repetitions are desired, for example if i want to make a beautiful dividing line like this:

=======================================================================

Maybe a safer approach would be to provide an abstraction that does the filtering depending on the OS.

Of course you will need at least [operating_system] from ZEXY for that, so this wouldn't be vanilla only. and of course it has other drawbacks, as it would need to be centralised (only one instance of [keyname] and then distribute the results with a [send])

that's at least the way i do it

attached is an abstraction i made a few years ago, haven't revised it lately, so please just take it as an example (also as it's only for windows ATM)

best

oliver

I might be mistaking what you're saying but, on OSX if key repeats are off in the system preferences (at least when pd starts) then they will not repeat from the [key] & [keyname] objects -seb

-----Original Message----- From: Peter P. peterparker@fastmail.com To: pd-list@lists.iem.at Sent: Tue, May 3, 2022 4:17 am Subject: Re: [PD] [key], [keyup] and operating system key reapat

Dear Dan and everyone contributing to this issue,

I kinda felt this is a sensitive topic but was not aware of the amount of frustrating work that has already gone into it. Every bit of it is much appreciated, thank you and everyone a lot!

So if I would want to write a patch that uses the [key] and [keyup] objects to work consistently on all three platforms, I would have to tell the user to turn key repeat off in her operating system AND provide a means to filter out repeating key ons (Win and OS X), correct?

In this case I happily deal with the current status.

thanks again, Peter

• Dan Wilcox danomatika@gmail.com [2022-05-02 12:38]:

For me, every time this topic comes up, I get recurring nightmares of having to write, set, and apply patches to the Tk source code. From my point of view, we should enjoy what works as it is and remember that for some people, the key handling should respect the general system as opposed to being Pd-specific. Of course, some may have a different opinion, in which case I ask them to contribute working solutions for test/discussion.

The only other solution is to open more issues and *hope* IOhannes fix everything again, but I *think* his feeling is similar to mine above... ;)

...

On May 2, 2022, at 12:00 PM, pd-list-request@lists.iem.at wrote:

Message: 2 Date: Sun, 1 May 2022 18:50:34 +0200 From: "Peter P." <peterparker@fastmail.com mailto:peterparker@fastmail.com> To: pd-list@lists.iem.at mailto:pd-list@lists.iem.at Subject: Re: [PD] [key], [keyup] and operating system key reapat Message-ID: <20220501165034.sq22mbyufuza5icn@fastmail.com mailto:20220501165034.sq22mbyufuza5icn@fastmail.com> Content-Type: text/plain; charset=us-ascii

• Peter P. <peterparker@fastmail.com mailto:peterparker@fastmail.com> [2022-05-01 18:44]:

[...]

...

Shouldn't/couldn't the filtering of repeated note-ons be done within the objects on Win and OS X at least?

Oh, just saw this issue https://github.com/pure-data/pure-data/issues/945 https://github.com/pure-data/pure-data/issues/945 and several related others. Seems complicated and a lot of work has gone into the current behavior already... I am wondering what the current status is, and if there is any future work planned?

Thanks to everyone working on it!!

P

---

Dan Wilcox @danomatika http://twitter.com/danomatika danomatika.com http://danomatika.com/ robotcowboy.com http://robotcowboy.com/

---

Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list

Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list

On 5/4/22 08:04, Peter P. wrote:

It would be great to have more unified behavior since so much work has already gone into this issue.

i guess this would be great for many users of Tcl/Tk. you might want to file a (wishlist) bug there.

gdmasdr IOhannes

PS: i'm happy to fix bugs in the Pd sources; but unlike Dan, I am *ont* willing to patch the Tcl/Tk engine for Pd's sake. i guess this saves me from a few nightmares

On 5/4/22 09:49, Peter P. wrote:

...

i guess this would be great for many users of Tcl/Tk. you might want to file a (wishlist) bug there.

If this doesn't provoke nightmares I'd be happy to do, perhaps for documentation purposes only.

it won't provoke nightmares for me (as i wouldn't even know what you are doing in the Tcl/Tk bugtracker). YMMV.

On 5/4/22 09:49, Peter P. wrote:

...

PS: i'm happy to fix bugs in the Pd sources; but unlike Dan, I am *not* willing to patch the Tcl/Tk engine for Pd's sake. i guess this saves me from a few nightmares

Thanks, just to be clear as I am not involved in this coding: The problem lies within Tcl/Tk right (and is known within that community of coders)?

i have no idea.

the Tcl/Tk community might be aware of the fact, but not see it as a problem.

like so many things this is probably just a result of different decisions that were made on the OS level. (and I assume that Tcl/Tk is just passing through the system characteristics).

e.g. on macOS there is this UX guideline, that whenever you change something in a properties window, the change is applied immediately. otoh, Linux and Windows systems typically use a distinct "Apply" button, that you have to actively press in order to apply your changes. now we *could* create a dialog that behaves the same on all platforms, which is great as there is now a consistent behaviour of Pd across OSs. but if we pick the macOS UX guideline, users of Linux or Windows will probably be unhappy. or vice-versa.

similar with key repeat: the designers of your desktop environment probably (hopefully!) spent some thought on why key repeat behaves as it does, but (unfortunately) they came to different solutions as what should be the "correct" behaviour. but if Tcl/Tk chose to abstract that behaviour away to a (new) behaviour, that is consistent across all platforms, wouldn't this make Tcl/Tk applications behave "weird" in the experience of users who work on a system that usually has a behaviour that *differs* from the one chosen by Tcl/Tk.

the simplistic answer is of course: if you want your dialogs to behave like on Windows, you shouldn't use macOS. if you want your keys to repeat like on Linux, you shouldn't use Windows.

in practice this is of course a straw man, as people don't typically pick their OSs because they like a given keyrepeat or dialog behaviour. but it hopefully explains why it's often not so simple to "fix" differences between OSs.

having said all that: the Tcl/Tk community might be able to give more informed feedback on such a feature request.

gfmadsr IOhannes