[RFC] etags ada parser: doesn't stop when giving broken input

[RFC] etags ada parser: doesn't stop when giving broken input

[Masatake YAMATO]

Look at following bug reproducer:

    % cat /tmp/a.adb
    cat /tmp/a.adb
    ~

    # Make TAGS file with the original version
    % /usr/bin/etags /tmp/a.adb 
    /usr/bin/etags /tmp/a.adb 
    #  Doen't stop C-c C-c^C

    # Patched version
    % ./etags /tmp/a.adb 
    ./etags /tmp/a.adb 
    % LC_ALL=C ls -l TAGS
    -rw-rw-r--. 1 yamato yamato 15 Apr  8 17:09 TAGS

I expect etags stops even if broken input is given.
I wrote a small patch. However, I'm not sure this is correct fix.
Someone who knows both ada and etags, could you review this patch?

    
2013-04-08  Masatake YAMATO  <yamato@redhat.com>

	* etags.c (Ada_funcs): Advance dbp before checking
	input with endtoken.

=== modified file 'lib-src/etags.c'
*** lib-src/etags.c	2013-03-13 07:27:34 +0000
--- lib-src/etags.c	2013-04-08 08:06:37 +0000
***************
*** 4231,4238 ****
  	    }
  
  	  /* Look for the end of the token. */
! 	  while (!endtoken (*dbp))
  	    dbp++;
  
  	} /* advance char */
      } /* advance line */
--- 4231,4239 ----
  	    }
  
  	  /* Look for the end of the token. */
! 	  do
  	    dbp++;
+ 	  while (!endtoken (*dbp));
  
  	} /* advance char */
      } /* advance line */

Re: [RFC] etags ada parser: doesn't stop when giving broken input

[Stephen Leake]

Masatake YAMATO <yamato@redhat.com> writes:

> I expect etags stops even if broken input is given.

That's certainly reasonable

> I wrote a small patch. However, I'm not sure this is correct fix.
> Someone who knows both ada and etags, could you review this patch?

I know Ada, but not etags. 

Your fix is not particular to Ada; it makes sense for any "token". It
looks reasonable to me, but I did not read the entire function (only the
patch). 

Is there a test suite for etags.c? I could contribute some Ada code, if
needed. 

-- 
-- Stephe

Re: [RFC] etags ada parser: doesn't stop when giving broken input

[Francesco Potortì]

>Is there a test suite for etags.c? I could contribute some Ada code, if
>needed. 

I have one.  I am out on a business trip.  Please remind me next week
and I will try to pack it ans send it to whover want sis it with instructions.

Re: [RFC] etags ada parser: doesn't stop when giving broken input

[Stefan Monnier]

> I have one.  I am out on a business trip.  Please remind me next week
> and I will try to pack it ans send it to whover want sis it
> with instructions.

Please add it to emacs/test somewhere.

It's better to put there tests in "raw form" (without scripts
surrounding them to run them conveniently) than to not have them at all.
Of course, it's even better if they come with nice scripts to run them.


        Stefan