Re: [Nano-devel] 2.1.8pre2

[Mike Frysinger]

On Monday 09 February 2009 00:47:14 Chris Allegretta wrote:
> On Sun, Feb 8, 2009 at 1:03 PM, Chris Allegretta <address@hidden> wrote:
> > On Sun, Feb 8, 2009 at 12:47 PM, Chris Allegretta <address@hidden> wrote:
> >> On Sun, Feb 8, 2009 at 1:11 AM, Mike Frysinger <address@hidden> wrote:
> >>> seems to be related to syntax highlighting.  when i have C highlighting
> >>> turned
> >>> on, it's much more crashy.
> >>>
> >>> try this (with highlighting and such turned on):
> >>>  - run `nano foo.c`
> >>>  - type: /**/
> >>>  - hit home (to go to start of line)
> >>>  - hit meta+t (to cut everything)
> >>>  - hit ctrl+u (to paste everything)
> >>>  - hit backspace
> >>>
> >>> some garbage shows up.  my guess is memory corruption.  that's why
> >>> sometimes
> >>> it "just works" for me and other times it segfaults and other times i
> >>> get garbage chars.
> >>>
> >>> i would also look at the test case i posted in the last thread where
> >>> you posted a pre release
> >>> -mike
> >>
> >> I'll do even better, run nano <file that doesnt exist>.c and type any
> >> character and it crashes.
> >> The magicline data wasn't getting initilized, which may also contribute
> >> to the issue you see below, since I couldn't get that to crash...
> >
> > Okay, that should hopefully fixed, as should be some too un-ambitious
> > resetting of multidata.  Apologies; there may be more cases which still
> > cause crashes.
>
> I put up a new tarball: nano-2.1.9pre1 with the --quiet flag implemented as
> well as the syntax highlighting fixes that were committed earlier.  IMO
> it's probably worth getting a more stable release out sooner rather than
> later, given that people are going to be bitten by those syntax bugs pretty
> frequently, at least the un-aggresive highlighting one.  Anyway, it's at:
>
> http://www.nano-editor.org/dist/test/nano-2.1.9pre1.tar.gz
>
> Assuming nothing major comes up which requires further string translation
> changes, I'll submit this PO file to the translators tomorrow so we can
> have 2.1.9 in a little over a week, and start working on the ~ syntax in
> .nanorcs which Eitan found in the meanwhile.

i havent seen any crashes with this.  the colorization behavior though seem to 
be much more "greedy" ... C code that uses comment blocks /* ... */ often 
globs across blocks.  so something like:
/*
 */
some C code
/*
 */
all gets highlighted as a comment ... this is with the start=/end= stuff in 
the nanorc files ...

i'm comparing to nano-2.1.7 btw as every nano-2.1.8pre+ has behaved this way
-mike

signature.asc
Description: This is a digitally signed message part.