Re: [PATCH 1/2] progmodes/vhdl-mode.el: Backported changes from 3.33.

[Stefan Monnier]

> Oh sorry, I guess I forgot to deactivate wordwrap in KMail. I attached both 
> patches to this email.

Thanks, I've installed the first.

> I'm simply helping to sync the two versions.  The vhdl-mode maintainer
> is Reto Zimmermann.

Good, your help is very appreciated.

>> Regarding the code I haven't reviewed it, but I noticed that it
>> advertises C-c C-c for commenting code, whereas it should advertise M-;
>> instead.
> We could change that downstream.

That's not important enough to warrant a diversion.

> The C-c C-c part is from the current vhdl-mode release.

I know, and I think Reto should change it.  M-; has been able to
comment/uncomment since Emacs-21.  Maybe the change can be made
conditional on the existence of comment-dwim.

>> Also, while I'm here, do you also sync changes from Emacs's repository
>> back to the upstream version?  E.g. the use of syntax-propertize-function.
> I tried to send the patch to the vhdl-mode maintainer but he wants to stay 
> compatible to all kind of old and forked emacs and that's why the patch is 
> probably not acceptable for him.

Which patch?  Regarding syntax-propertize-function, I made an effort to
write the vhdl code in such a way that it gracefully falls back on the
old code if syntax-propertize-function doesn't exist, so it should be
fully backward compatible (although I haven't tested it).

> Maybe we could send at least the documentation fixes upstream.

Sure.  Do it in relatively small chunks, which makes it easier for Reto
review, and maximizes the proportion he can accept.

> Reducing the differences would be useful.

Very much so.

> It took a lot of time with diff-mode and git-blame to figure out what
> was changed in Emacs and what is new in the vhdl-mode release.

I know.  That's why I did not install your second patch, which would
increase the differences, without any significant improvement (while
both are obsolete, they won't disappear soon, especially not
`interactive-p').

You can try and submit a patch upstream to fix those warnings by
changing to code to dynamically use either of the old or new code
depending on what's available.

> @@ -197,6 +197,21 @@ Examples:
>    "Customizations for modes."
>    :group 'vhdl)
 
> +(defcustom vhdl-electric-mode t
> +  "Non-nil enables electrification (automatic template generation).
> +If nil, template generators can still be invoked through key bindings and
> +menu.  Is indicated in the modeline by \"/e\" after the mode name and can be
> +toggled by `\\[vhdl-electric-mode]'."
> +  :type 'boolean
> +  :group 'vhdl-mode)
> +
> +(defcustom vhdl-stutter-mode t
> +  "Non-nil enables stuttering.
> +Is indicated in the modeline by \"/s\" after the mode name and can be toggled
> +by `\\[vhdl-stutter-mode]'."
> +  :type 'boolean
> +  :group 'vhdl-mode)
> +
>  (defcustom vhdl-indent-tabs-mode nil
>    "Non-nil means indentation can insert tabs.
>  Overrides local variable `indent-tabs-mode'."

This is actually undoing a local change where those defcustoms are not
needed any more, because we use define-minor-mode (further down) to
define those vars.


        Stefan