Re: editor

[Mark-Jason Dominus <mjd@xxxxxxxxxx>]

> Suggestion:
> 
> Instead of dying on no-editor-found we should note to the user that
> the dataset is read-only. 

I don't think it makes sense to couple these two things together.
They are not related.  It makes sense to use the user's favorite
editor to view cell contents even if they are read-only, and
conversely, the user can modify a writable dataset with commands like
`break_link' whether or not their editor makes sense.

Also, suppose you try to start up ZigZag, and you end up in read-only
mode.  You didn't want that, so you set about trying to correct the
problem.  You check the permissions on the file, no, the permissions
are all right.  Maybe you check some other things.  But you are not
likely to think to check whether `emacs' is installed in /usr/bin or
/usr/local/bin.  

That is not the Principle of Least Surprise.

> This lets people frob the choose_editor() function to set things to
> read-only if they want.
> 
> Comments?

I would suggest that if the user wants to open a dataset read-only
then we could provide a cprogram command-line argument that says so
explicitly.  Then they could still use their favorite editor to view
cell contents without changing them.

	zigzag -r file.zz		Read-only mode