If undo-only why not a redo-only?

If undo-only why not a redo-only?

[Ergus]

[-- Attachment #1: Type: text/plain, Size: 358 bytes --]

Hi:

I've seen that we have an undo-only command and a way to customize a variable to do so for undo-redo.
Is it too complex to implement the same to do a redo-only equivalent in order to bind that to a different key if the user wants so?
There are different external packages for that, but they add some unneeded complexity and are somehow buggy.
Best,Ergus

[-- Attachment #2: Type: text/html, Size: 543 bytes --]

Re: If undo-only why not a redo-only?

[Stefan Monnier]

> I've seen that we have an undo-only command and a way to customize
> a variable to do so for undo-redo.
> Is it too complex to implement the same to do a redo-only equivalent in
> order to bind that to a different key if the user wants so?

I don't understand what you're asking for.
Could you give some examples of what you'd like to see?

The current code is designed so you can have a "plain old undo+redo"
system by using the commands `undo-only` and `undo-redo`.


        Stefan

Re: If undo-only why not a redo-only?

[Ergus]

[-- Attachment #1: Type: text/plain, Size: 1541 bytes --]

Hi Stefan:

It is actually simple. The idea is to have more or less what undo-fu provides but in vanilla. I don't actually know if it is possible to do that for redo with what we already have there. (Probably yes and I don't know how)
The idea is to navigate only in one direction until the initial opposite action. or the other with the *-only commands, without the danger of switching unintentionally from undo to redo or vice-versa. If there is not more undo or redo or the next action is the opposite (redo/undo), just stop and do nothing (print a message maybe)

Basically I don't want to depend of undo-fu anymore for such a basic editing functionality as we already have the half of the work and because it tries to reinvent part of the wheel in my opinion.

Does it makes sense?

-----Original Message-----
From: Stefan Monnier <monnier@iro.umontreal.ca>
To: Ergus <spacibba@aol.com>
Cc: emacs-devel@gnu.org <emacs-devel@gnu.org>
Sent: Mon, Jun 8, 2020 8:10 pm
Subject: Re: If undo-only why not a redo-only?

> I've seen that we have an undo-only command and a way to customize
> a variable to do so for undo-redo.
> Is it too complex to implement the same to do a redo-only equivalent in
> order to bind that to a different key if the user wants so?

I don't understand what you're asking for.
Could you give some examples of what you'd like to see?

The current code is designed so you can have a "plain old undo+redo"
system by using the commands `undo-only` and `undo-redo`.


        Stefan



[-- Attachment #2: Type: text/html, Size: 2187 bytes --]

Re: If undo-only why not a redo-only?

[Ergus]

[-- Attachment #1: Type: text/plain, Size: 799 bytes --]

BTW: https://gitlab.com/ideasman42/emacs-undo-fu


-----Original Message-----
From: Stefan Monnier <monnier@iro.umontreal.ca>
To: Ergus <spacibba@aol.com>
Cc: emacs-devel@gnu.org <emacs-devel@gnu.org>
Sent: Mon, Jun 8, 2020 8:10 pm
Subject: Re: If undo-only why not a redo-only?

> I've seen that we have an undo-only command and a way to customize
> a variable to do so for undo-redo.
> Is it too complex to implement the same to do a redo-only equivalent in
> order to bind that to a different key if the user wants so?

I don't understand what you're asking for.
Could you give some examples of what you'd like to see?

The current code is designed so you can have a "plain old undo+redo"
system by using the commands `undo-only` and `undo-redo`.


        Stefan



[-- Attachment #2: Type: text/html, Size: 1437 bytes --]

Re: If undo-only why not a redo-only?

[Stefan Monnier]

> The idea is to navigate only in one direction until the initial opposite
> action. or the other with the *-only commands, without the danger of
> switching unintentionally from undo to redo or vice-versa. If there is not
> more undo or redo or the next action is the opposite (redo/undo), just stop
> and do nothing (print a message maybe)

Yes, that's what `undo-only` and `undo-redo` attempt to provide, IIUC.
Have you bumped into problems with them?


        Stefan

Re: If undo-only why not a redo-only?

[Ergus]

[-- Attachment #1: Type: text/plain, Size: 1488 bytes --]

Ohh, now I got it. 

Sorry, but the name `undo-redo` was not clear for me at all (it seems more appropriated for the normal emacs undo as it actually does undo and redo) and the function documentation was confusing itself "Undo the last ARG undos" (it seems a prophecy of an oracle ;p ) also I didn't find anything more explicit in the manual either. Actually there are some useful undo related functions in simple.el I wasn't aware of and I just discovered looking into it. Maybe it would be fine to rename it or provide an alias more "symmetric" like `redo-only` or just `redo`?

So the only missing functions available in undo-fu are undo-fu-[redo/undo]-all but I have never used them so I don't actually care them.
Thanks for everything.Ergus




-----Original Message-----
From: Stefan Monnier <monnier@iro.umontreal.ca>
To: Ergus <spacibba@aol.com>
Cc: emacs-devel@gnu.org <emacs-devel@gnu.org>
Sent: Mon, Jun 8, 2020 9:27 pm
Subject: Re: If undo-only why not a redo-only?

> The idea is to navigate only in one direction until the initial opposite
> action. or the other with the *-only commands, without the danger of
> switching unintentionally from undo to redo or vice-versa. If there is not
> more undo or redo or the next action is the opposite (redo/undo), just stop
> and do nothing (print a message maybe)

Yes, that's what `undo-only` and `undo-redo` attempt to provide, IIUC.
Have you bumped into problems with them?


        Stefan



[-- Attachment #2: Type: text/html, Size: 2100 bytes --]

Re: If undo-only why not a redo-only?

[João Távora]

[-- Attachment #1: Type: text/plain, Size: 455 bytes --]

On Mon, Jun 8, 2020, 22:19 Ergus <spacibba@aol.com> wrote:

> Ohh, now I got it.
>
> Sorry, but the name `undo-redo` was not clear for me at all
>

Can't resist pointing out that that is quite  understandable. The
"prefix-with-library" name convention is fraught with these little
misunderstandings. It's arguably the last bad option we have now, but we
would need a proper namespace system to fix them: renaming can only do so
much.

João

[-- Attachment #2: Type: text/html, Size: 894 bytes --]

RE: If undo-only why not a redo-only?

[Drew Adams]

>> Ohh, now I got it. 
>> Sorry, but the name `undo-redo` was not clear for me at all 
>
> Can't resist pointing out that that is quite  understandable.
> The "prefix-with-library" name convention is fraught with
> these little misunderstandings. It's arguably the last bad
> option we have now, but we would need a proper namespace
> system to fix them: renaming can only do so much.

I was thinking the same thing.

Using hyphen to separate both pkg prefix and other parts of
a name is confusing.  Better, perhaps, to use something like
`/' or `:' as pkg separator, while waiting for a real,
namespace solution.

But didn't you mean "least bad", not "last bad".  Surely it's
not the last one that can be found. ;-)

Re: If undo-only why not a redo-only?

[João Távora]

[-- Attachment #1: Type: text/plain, Size: 275 bytes --]

On Tue, Jun 9, 2020 at 4:19 PM Drew Adams <drew.adams@oracle.com> wrote:

>
> But didn't you mean "least bad", not "last bad".  Surely it's
> not the last one that can be found. ;-)
>

Ehe, thanks, I did mean "least".  Indeed, as we know,
it's always possible to do worse...

[-- Attachment #2: Type: text/html, Size: 629 bytes --]