Re: sh-mode highlight issue

[Thomas Dickey]

On Mon, Jan 31, 2022 at 03:20:14PM +0000, Wayne Cuddy via discussions 
concerning the vile text editor wrote:
> On Sun, Jan 30, 2022 at 03:45:15PM -0500, Thomas Dickey wrote:
> > On Sat, Jan 29, 2022 at 04:09:09PM -0500, Thomas Dickey wrote:
> > > On Sat, Jan 29, 2022 at 07:51:53PM +0000, Wayne Cuddy via discussions 
> > > concerning the vile text editor wrote:
> > > > I'm not sure if this has been reported before, please forgive me if
> > > > this is repetitious.
> > > > 
> > > > I'm running 9.8t and have not had a chance to upgrade to the
> > > > latest/greatest yet but I don't see mention of this issue being
> > > > corrected in the changelog. The closest I could find was comments from
> > > > 9.7y:
> > > > 
> > > >     + improve vile-sh-filt for these cases:
> > > >     + "${number}"
> > > >     + "$(commands)"
> > > >     + "$((expression))"
> > > >     + "${name#value}", etc.
> > > > 
> > > > 
> > > > The issue concerns $((expression)) expansions in variable assignments,
> > > > but may not be limited to just this case.
> > > > 
> > > > var="prefix$((5*5))suffix"
> > >        ^ string
> > >                 ^expression
> > >                      ^ string?
> > > > 
> > > > prefix appears correctly highlighted, suffix does not, it seems the
> > > > syntax highlighting terminates after the final ')' character.
> > > 
> > > Just looking at the source, the line
> > > 
> > >   <EXPRS>{BACKTIC2_UP}    { handle_backtic2(yytext, yyleng, 
> > > LEN_BACKTIC2_DOWN, 0); }
> > > 
> > > is probably what should trigger on the "))", which takes it into
> > > 
> > >     if (FLTSTACK_OK && (which <= 0) && (FLTSTACK_THIS.backtic == test)) {
> > > 
> > > and fails there for some reason -- otherwise it would go back to the
> > > string/literal coloring.
> > > 
> > > (I'm in the middle of ncurses, but may dig into it tomorrow to get a 
> > > better
> > > answer).
> > 
> > See attached
> > 
> > -- 
> > Thomas E. Dickey <dickey@invisible-island.net>
> > https://invisible-island.net
> > ftp://ftp.invisible-island.net
> 
> > --- sh-filt.l       2021/12/12 01:26:53     1.183
> > +++ sh-filt.l       2022/01/30 20:27:01
> > @@ -70,6 +70,7 @@
> >  static void handle_backtic2(const char *text, int length, int test, int 
> > which);
> >  static void handle_backtic1(const char *text, int length, int test, int 
> > which);
> >  static void handle_parens(const char *text, int length, int test, int 
> > which, int state);
> > +static void resume_state(void);
> >  static void save_here(const char *text, int length);
> >  static void write_vname(const char *text, int length);
> >  
> > @@ -198,8 +199,10 @@
> >                         FLEX_PRINTF((stderr, "cannot pop '$((' level %d\n", 
> > stk_level));
> >                         if (!pop_backtic1(yytext, 1, LEN_BACKTIC1_DOWN)) {
> >                           FLEX_PRINTF((stderr, "cannot pop '$(' level 
> > %d\n", stk_level));
> > -                         if (FLTSTACK_OK && stk_state[stk_level-1].state 
> > == NORMAL)
> > +                         if (FLTSTACK_OK && stk_state[stk_level-1].state 
> > == NORMAL) {
> >                             pop_state();
> > +                           resume_state();
> > +                         }
> >                           flt_putc(R_PAREN);
> >                         }
> >                         unput(R_PAREN);
> > @@ -213,8 +216,10 @@
> >                     }
> >  <NORMAL>{BACKTIC1_UP}      { if (!pop_backtic1(yytext, yyleng, 
> > LEN_BACKTIC1_DOWN)) {
> >                         FLEX_PRINTF((stderr, "cannot pop '$(' level %d\n", 
> > stk_level));
> > -                       if (FLTSTACK_OK && stk_state[stk_level-1].state == 
> > NORMAL)
> > +                       if (FLTSTACK_OK && stk_state[stk_level-1].state == 
> > NORMAL) {
> >                           pop_state();
> > +                         resume_state();
> > +                       }
> >                         ECHO;
> >                       }
> >                     }
> > @@ -293,6 +298,7 @@
> >  <KSH_MATH>"))"             {
> >                       WriteToken(Action_attr); 
> >                       pop_state();
> > +                     resume_state();
> >                     }
> >  <KSH_MATH>{IDENT1} |
> >  <KSH_MATH>{VNAME}  { WriteToken(Ident_attr); }
> > @@ -312,6 +318,7 @@
> >                       flt_bfr_append("'", 1);
> >                       flt_bfr_finish();
> >                       pop_state();
> > +                     resume_state();
> >                     }
> >  <KSH_CQUOTE>\\[0-7]{1,3}  {
> >                       flt_bfr_embed(yytext, 1, Action_attr);
> > @@ -874,6 +881,7 @@
> >     flt_bfr_append(value + used, length - used);
> >     flt_bfr_finish();
> >     pop_state();
> > +   resume_state();
> >     FreeAndNull(here_tag);
> >     here_exp = 0;
> >     here_next = 0;
> 
> I'll make time to upgrade and test this patch later this week and post
> the results.

sounds good.  Actually I didn't have anything in my collection of scripts
which I could have used to see this bug, and other using the test-case
which I constructed, saw no differences in output.
 
> Many thanks as usual, 

no problem (I started working on a markdown filter, but can see that
xterm's going to use up this week).

-- 
Thomas E. Dickey <dickey@invisible-island.net>
https://invisible-island.net
ftp://ftp.invisible-island.net

signature.asc
Description: PGP signature