[slim-vim] bugs with -u NONE -U NONE

[Larry Clapp]

On Wed, Aug 30, 2006 at 03:11:29PM +0600, Dmitry Petukhov wrote:
> 2006/8/29, Brad Beveridge <brad.beveridge at gmail.com>:
> > This is really perplexing then.  I run 0.9.15, and if it works
> > with Slime, it ought to be fine.  The only way I can think of to
> > debug this is
> > 1) Open Emacs and connect to Slime, open the "slime-messages" buffer
> > 2) Reproduce steps on regular Slime, copy the text is slime-messages
> > 3) Reproduce on Slim-vim and copy the data from --sv-debug--
> 
> i've compared slime and slim-vim conversation with swank, turns out
> what slime d o not send package = nil with
> invoke-nth-restart-for-emacs, and also sends the thread number
> 
> slime:
> (:emacs-rex
>  (swank:invoke-nth-restart-for-emacs 1 0)
>  nil 2 17
> 
> slim-vim:
> (:EMACS-REX (SWANK:INVOKE-NTH-RESTART-FOR-EMACS 1 0) "COMMON-LISP-USER" T 4)
> 
> I tried to fix this, and with the patch below, slim-vim<>swank
> conversation is similar to slime<>swank. And i even observed correct
> behaviour once - but for most times, vim crashes with
> 
> 
> E315: ml_get: invalid lnum: 16
> E315: ml_get: invalid lnum: 16Segmentation violation.
> 
> number varies.

Fwiw, when I get "ml_get: invalid lnum: N", usually I've tried to read
line N when I should've tried to read N-1.

-- L