[ESS-bugs] ess-mode 18.10.3snapshot [<unknown>]; Indenting in stata mode seems to call ess-indent-line in an infinite loop

[Martin Maechler]

>>>>> Alex Branham 
>>>>>     on Thu, 14 Mar 2019 10:53:15 -0500 writes:

    > On Thu 14 Mar 2019 at 10:48, Brendan Halpin <brendan.halpin using ul.ie> wrote:

    >> On Thu, Mar 14 2019, Alex Branham wrote:
    >> 
    >>> Now the question becomes: Do we want to "fix" this? It seems a little
    >>> crazy to me that stata-mode relies on R's indentation engine. We could
    >>> just let stata-mode use Emacs's default indent-line-function (which is
    >>> indent-relative) and add a note about this in the NEWS.
    >> 
    >> I would prefer something that indents appropriately within loops, and
    >> after line continuation markers:
    >> 
    >> forvalues i = 1/5 {
    >> di "`i'"
    >> }
    >> 
    >> line y x, lc(red) ///
    >> yscale(log)
    >> 
    >> I may be getting it wrong, but I don't think indent-relative will do this.

    > You're right. I should've been clearer. A proper indentation engine for
    > stata-mode would obviously be better. However, no one is actively
    > working on stata-mode, so it's mostly in maintenance-only mode. Assuming
    > no one steps forward and contributes a proper indentation engine for
    > stata-mode, our options are either continue to rely on R's indentation
    > (which, again, seems crazy to me) or rely on Emacs's defaults.

    > Alex

If R's indentation is closer to what stata users expect,
*and* given the fact that this seems to have been the default
now for many years in ESS stata-mode, I strongly believe we
should continue to do so. 

We could possibly "wrap" this by defining  "ess-stata-indent"
(paraphrasing the true names)
and then define -- for the time being "ess-stata-indent" to call "ess-r-indent"

Martin