Re: Broken "C-b remapped to ..." advice in tutorial.

[Lennart Borgman]

Chong Yidong wrote:
> "Kim F. Storm" <address@hidden> writes:
>
>   
> > emacs -Q
> > (global-set-key "\C-b" 'ignore)
> > C-h C-t
> >
> >
> > It now contains this misplaced advice (with C-b in read):
> >
> >         C-x C-b   List buffers
> > ** The key C-b has been rebound, but you can use <left> instead [More 
> > information] **
> >
> >
> > (notice also that the info line is >80 chars long).
> >     
>
> I checked in a fix.
>   

Thanks. I fixed it once and wondered what happened to my old fix.


> By the way, there are some pretty dubious features in the new
> tutorial.  I think we should revisit them before the release.
>
> 1. There's no call for a special tutorial--tab-map to tab between
>    warning messages.  Why bother?
>   

No big deal, but that is a way it often works in other applications.

> 2. Displaying one line of warning message for each rebound key
>    sequence is really annoying.  We should just highlight them in a
>    different face and display the message in a tooltip, and/or when
>    the text is clicked on.  After all, there is already a big warning
>    message at the top of the tutorial page saying keys have been
>    rebound; no need to hammer it in.
>   

I prefer the suggestion I think Kim made to show the warning only for 
the first occurrence of a rebound key.

> 3. The tutorial-position saving mechanism is very opaque.  When the
>    buffer is killed, the current position in the tutorial, without
>    prompting the user (it used to save into a directory ~/.emacstut,
>    but I moved this into .emacs.d to avoid homedir pollution).
>    However, this saving does not occur if the user does C-x C-c or C-x
>    C-s.  This feature is confusing enough that I'd propose removing it
>    entirely.
>   

That the tutorial state is not saved for C-x C-c is probably a bug in 
Emacs. I think the hooks for killing buffers is not run at exit. Should 
they not be run then?

That the tutorial state is not saved for C-x C-s is ... well I never 
thought of it. Why should the tutorial state be saved when trying to 
save the buffer? Is not that quite a different thing than the state of 
the tutorial?

The feature to save the tutorial state was there before also, but it 
never worked. Now it actually works I believe (it is the buffer+position 
that is saved) . But yes, it is a bit obscure. However I think it is 
helpful since the tutorial is long. Currently the user is offered to 
restart the tutorial fresh or use the saved state when starting the 
tutorial. I think that is good.