Re: [RFA]Use prev_lexptr in *-exp.y

Re: [RFA]Use prev_lexptr in *-exp.y

[Michael Elizabeth Chastain]

Pierre Muller writes:
> By the way, I don't have the necessary stuff to run the 
> testsuite... Thus I can not warranty that there are no 
> regressions. There mught be some wrong regressions,
> if some testsuite do check the error output in case of 
> wrong expression entered. The error output should be more
> logical now, but could get some tests  to fail.

Pierre,

I have that covered.  I run the test suite every 5-10 days on many
configurations, and I have a good procedure for diff'ing the results
and finding regressions.  So if there are changes in the gdb output and
they cause regressions in the test suite then I will catch them and post
detailed bug reports.

Michael C

Re: [RFA]Use prev_lexptr in *-exp.y

[Michael Snyder]

Michael Elizabeth Chastain wrote:
> 
> Pierre Muller writes:
> > By the way, I don't have the necessary stuff to run the
> > testsuite... Thus I can not warranty that there are no
> > regressions. There mught be some wrong regressions,
> > if some testsuite do check the error output in case of
> > wrong expression entered. The error output should be more
> > logical now, but could get some tests  to fail.
> 
> Pierre,
> 
> I have that covered.  I run the test suite every 5-10 days on many
> configurations, and I have a good procedure for diff'ing the results
> and finding regressions.  So if there are changes in the gdb output and
> they cause regressions in the test suite then I will catch them and post
> detailed bug reports.

Michael, I think Pierre was talking about Modula 2, Java, 
and Fortran.  Does your testing cover those?

Re: [RFA]Use prev_lexptr in *-exp.y

[Michael Elizabeth Chastain]

Hmmm, if they are f90/f95 specific, and the existing tests are f77,
then I would prefer gdb.f90 or gdb.f95 to gdb.fortran (since they
are different languages).

Other opinions?

Michael C

Re: [RFA]Use prev_lexptr in *-exp.y

[Petr Sorfa]

Hi Michael,

Well f77 is a subset of f90/f95. I guess we need someone else to comment
otherwise, I'll got with your suggestion and create a gdb.f95 directory.

Petr
> Hmmm, if they are f90/f95 specific, and the existing tests are f77,
> then I would prefer gdb.f90 or gdb.f95 to gdb.fortran (since they
> are different languages).
> 
> Other opinions?
> 
> Michael C

Re: [RFA]Use prev_lexptr in *-exp.y

[Michael Elizabeth Chastain]

> Michael, I think Pierre was talking about Modula 2, Java, 
> and Fortran.  Does your testing cover those?

There are no Modula 2 tests in the test suite, as far as I can see.
So I don't have any coverage for Modula 2.

I don't cover Java.  That is on my to-do list.

There are two scripts in gdb.fortran with 116 tests.  I cover those.
The tests are very basic; they don't actually compile any FORTRAN code.
They just exercise the expression parser.

Michael C

Re: [RFA]Use prev_lexptr in *-exp.y

[Michael Snyder]

Michael Elizabeth Chastain wrote:
> 
> > Michael, I think Pierre was talking about Modula 2, Java,
> > and Fortran.  Does your testing cover those?
> 
> There are no Modula 2 tests in the test suite, as far as I can see.
> So I don't have any coverage for Modula 2.
> 
> I don't cover Java.  That is on my to-do list.
> 
> There are two scripts in gdb.fortran with 116 tests.  I cover those.
> The tests are very basic; they don't actually compile any FORTRAN code.
> They just exercise the expression parser.

Well, that's exactly what we're changing, so it sounds like
perhaps fortran is covered.

Re: [RFA]Use prev_lexptr in *-exp.y

[Petr Sorfa]

Hi Michael,


Just to let you know I'm adding about 500 test cases to FORTRAN
(hopefully this week.) Most are based on compiled code. Which is going
to be interesting to automate as the tests are F90/F95 specific and
there is no functioning g95. I'm using the Intel compiler for IA32 and
IA64.

Currently I've got all the tests in gdb.fortran. I've also replicated
the existing tests to invoke the new FORTRAN95 parser (and to make sure
that the current FORTRAN interface doesn't get broken.)

Petr
> > Michael, I think Pierre was talking about Modula 2, Java,
> > and Fortran.  Does your testing cover those?
> 
> There are no Modula 2 tests in the test suite, as far as I can see.
> So I don't have any coverage for Modula 2.
> 
> I don't cover Java.  That is on my to-do list.
> 
> There are two scripts in gdb.fortran with 116 tests.  I cover those.
> The tests are very basic; they don't actually compile any FORTRAN code.
> They just exercise the expression parser.
> 
> Michael C

[PATCH]Use prev_lexptr in p-exp.y

[Pierre Muller]

This basically just adds the same functionality
than in the c language parser.
See
http://sources.redhat.com/ml/gdb-patches/2002-04/msg00467.html


ChangeLog entry:

2002-04-25  Pierre Muller  <muller@ics.u-strasbg.fr>

	* p-exp.y: Also use new prev_lexptr variable
	to improve error reporting. Based on Michael Snyder
	2002-04-24 dated patch to c-exp.y.

Index: p-exp.y
===================================================================
RCS file: /cvs/src/src/gdb/p-exp.y,v
retrieving revision 1.9
diff -u -p -r1.9 p-exp.y
--- p-exp.y     18 Apr 2002 15:22:18 -0000      1.9
+++ p-exp.y     25 Apr 2002 08:38:29 -0000
@@ -947,6 +947,8 @@ yylex ()

   retry:

+  prev_lexptr = lexptr;
+
    tokstart = lexptr;
    explen = strlen (lexptr);
    /* See if it is a special token of length 3.  */
@@ -1481,5 +1483,8 @@ void
  yyerror (msg)
       char *msg;
  {
+  if (prev_lexptr)
+    lexptr = prev_lexptr;
+
    error ("A %s in expression, near `%s'.", (msg ? msg : "error"), lexptr);
  }



Pierre Muller
Institut Charles Sadron
6,rue Boussingault
F 67083 STRASBOURG CEDEX (France)
mailto:muller@ics.u-strasbg.fr
Phone : (33)-3-88-41-40-07  Fax : (33)-3-88-41-40-99

Re: [PATCH]Use prev_lexptr in p-exp.y

[Andrew Cagney]

> 2002-04-25  Pierre Muller  <muller@ics.u-strasbg.fr>
> 
> * p-exp.y: Also use new prev_lexptr variable
> 	to improve error reporting. Based on Michael Snyder
> 	2002-04-24 dated patch to c-exp.y.
> 

Nice.  Can I suggest creating bug report so we know that the others 
should be done.

Andrew

[RFA]Use prev_lexptr in *-exp.y

[Pierre Muller]

At 17:30 25/04/2002 , Andrew Cagney a écrit:
>>2002-04-25  Pierre Muller  <muller@ics.u-strasbg.fr>
>>* p-exp.y: Also use new prev_lexptr variable
>>         to improve error reporting. Based on Michael Snyder
>>         2002-04-24 dated patch to c-exp.y.
>
>Nice.  Can I suggest creating bug report so we know that the others should be done.



This is more work than fixing them directly!

So what about that:

ChangeLog entry:

2002-04-25  Pierre Muller  <muller@ics.u-strasbg.fr>
         * f-exp.y: Also use new prev_lexptr variable
         to improve error reporting. Based on Michael Snyder
         2002-04-24 dated patch to c-exp.y.
         * jv-exp.y: Likewise.
         * m2-exp.y: Likewise.


Index: f-exp.y
===================================================================
RCS file: /cvs/src/src/gdb/f-exp.y,v
retrieving revision 1.5
diff -u -p -r1.5 f-exp.y
--- f-exp.y     6 Mar 2001 08:21:07 -0000       1.5
+++ f-exp.y     25 Apr 2002 16:45:29 -0000
@@ -924,7 +924,9 @@ yylex ()
    char *tokstart;
    
   retry:
-  
+ 
+  prev_lexptr = lexptr;
+ 
    tokstart = lexptr;
    
    /* First of all, let us make sure we are not dealing with the 
@@ -1171,5 +1173,8 @@ void
  yyerror (msg)
       char *msg;
  {
+  if (prev_lexptr)
+    lexptr = prev_lexptr;
+
    error ("A %s in expression, near `%s'.", (msg ? msg : "error"), lexptr);
  }
Index: jv-exp.y
===================================================================
RCS file: /cvs/src/src/gdb/jv-exp.y,v
retrieving revision 1.5
diff -u -p -r1.5 jv-exp.y
--- jv-exp.y    21 Feb 2002 02:54:46 -0000      1.5
+++ jv-exp.y    25 Apr 2002 16:45:29 -0000
@@ -862,6 +862,8 @@ yylex ()
    
   retry:
  
+  prev_lexptr = lexptr;
+
    tokstart = lexptr;
    /* See if it is a special token of length 3.  */
    for (i = 0; i < sizeof tokentab3 / sizeof tokentab3[0]; i++)
@@ -1207,6 +1209,9 @@ void
  yyerror (msg)
       char *msg;
  {
+  if (prev_lexptr)
+    lexptr = prev_lexptr;
+
    error ("A %s in expression, near `%s'.", (msg ? msg : "error"), lexptr);
  }
  
Index: m2-exp.y
===================================================================
RCS file: /cvs/src/src/gdb/m2-exp.y,v
retrieving revision 1.3
diff -u -p -r1.3 m2-exp.y
--- m2-exp.y    6 Mar 2001 08:21:09 -0000       1.3
+++ m2-exp.y    25 Apr 2002 16:45:29 -0000
@@ -821,6 +821,8 @@ yylex ()
  
   retry:
  
+  prev_lexptr = lexptr;
+
    tokstart = lexptr;
  
  
@@ -1090,5 +1092,8 @@ void
  yyerror (msg)
       char *msg;
  {
+  if (prev_lexptr)
+    lexptr = prev_lexptr;
+
    error ("A %s in expression, near `%s'.", (msg ? msg : "error"), lexptr);
  }



Pierre Muller
Institut Charles Sadron
6,rue Boussingault
F 67083 STRASBOURG CEDEX (France)
mailto:muller@ics.u-strasbg.fr
Phone : (33)-3-88-41-40-07  Fax : (33)-3-88-41-40-99

Re: [RFA]Use prev_lexptr in *-exp.y

[Michael Snyder]

Pierre Muller wrote:
> 
> At 17:30 25/04/2002 , Andrew Cagney a écrit:
> >>2002-04-25  Pierre Muller  <muller@ics.u-strasbg.fr>
> >>* p-exp.y: Also use new prev_lexptr variable
> >>         to improve error reporting. Based on Michael Snyder
> >>         2002-04-24 dated patch to c-exp.y.
> >
> >Nice.  Can I suggest creating bug report so we know that the others should be done.
> 
> This is more work than fixing them directly!

ROTFL!  You rock, Pierre!   <g>

> 
> So what about that:
> 
> ChangeLog entry:
> 
> 2002-04-25  Pierre Muller  <muller@ics.u-strasbg.fr>
>          * f-exp.y: Also use new prev_lexptr variable
>          to improve error reporting. Based on Michael Snyder
>          2002-04-24 dated patch to c-exp.y.
>          * jv-exp.y: Likewise.
>          * m2-exp.y: Likewise.
> 
> Index: f-exp.y
> ===================================================================
> RCS file: /cvs/src/src/gdb/f-exp.y,v
> retrieving revision 1.5
> diff -u -p -r1.5 f-exp.y
> --- f-exp.y     6 Mar 2001 08:21:07 -0000       1.5
> +++ f-exp.y     25 Apr 2002 16:45:29 -0000
> @@ -924,7 +924,9 @@ yylex ()
>     char *tokstart;
> 
>    retry:
> -
> +
> +  prev_lexptr = lexptr;
> +
>     tokstart = lexptr;
> 
>     /* First of all, let us make sure we are not dealing with the
> @@ -1171,5 +1173,8 @@ void
>   yyerror (msg)
>        char *msg;
>   {
> +  if (prev_lexptr)
> +    lexptr = prev_lexptr;
> +
>     error ("A %s in expression, near `%s'.", (msg ? msg : "error"), lexptr);
>   }
> Index: jv-exp.y
> ===================================================================
> RCS file: /cvs/src/src/gdb/jv-exp.y,v
> retrieving revision 1.5
> diff -u -p -r1.5 jv-exp.y
> --- jv-exp.y    21 Feb 2002 02:54:46 -0000      1.5
> +++ jv-exp.y    25 Apr 2002 16:45:29 -0000
> @@ -862,6 +862,8 @@ yylex ()
> 
>    retry:
> 
> +  prev_lexptr = lexptr;
> +
>     tokstart = lexptr;
>     /* See if it is a special token of length 3.  */
>     for (i = 0; i < sizeof tokentab3 / sizeof tokentab3[0]; i++)
> @@ -1207,6 +1209,9 @@ void
>   yyerror (msg)
>        char *msg;
>   {
> +  if (prev_lexptr)
> +    lexptr = prev_lexptr;
> +
>     error ("A %s in expression, near `%s'.", (msg ? msg : "error"), lexptr);
>   }
> 
> Index: m2-exp.y
> ===================================================================
> RCS file: /cvs/src/src/gdb/m2-exp.y,v
> retrieving revision 1.3
> diff -u -p -r1.3 m2-exp.y
> --- m2-exp.y    6 Mar 2001 08:21:09 -0000       1.3
> +++ m2-exp.y    25 Apr 2002 16:45:29 -0000
> @@ -821,6 +821,8 @@ yylex ()
> 
>    retry:
> 
> +  prev_lexptr = lexptr;
> +
>     tokstart = lexptr;
> 
> 
> @@ -1090,5 +1092,8 @@ void
>   yyerror (msg)
>        char *msg;
>   {
> +  if (prev_lexptr)
> +    lexptr = prev_lexptr;
> +
>     error ("A %s in expression, near `%s'.", (msg ? msg : "error"), lexptr);
>   }
> 
> Pierre Muller
> Institut Charles Sadron
> 6,rue Boussingault
> F 67083 STRASBOURG CEDEX (France)
> mailto:muller@ics.u-strasbg.fr
> Phone : (33)-3-88-41-40-07  Fax : (33)-3-88-41-40-99

Re: [RFA]Use prev_lexptr in *-exp.y

[Pierre Muller]

At 20:38 25/04/2002 , Michael Snyder a écrit:
>Pierre Muller wrote:
> > 
> > At 17:30 25/04/2002 , Andrew Cagney a écrit:
> > >>2002-04-25  Pierre Muller  <muller@ics.u-strasbg.fr>
> > >>* p-exp.y: Also use new prev_lexptr variable
> > >>         to improve error reporting. Based on Michael Snyder
> > >>         2002-04-24 dated patch to c-exp.y.
> > >
> > >Nice.  Can I suggest creating bug report so we know that the others should be done.
> > 
> > This is more work than fixing them directly!
>
>ROTFL!  You rock, Pierre!   <g>

May I consider this as an approval and commit this patch?


Pierre Muller
Institut Charles Sadron
6,rue Boussingault
F 67083 STRASBOURG CEDEX (France)
mailto:muller@ics.u-strasbg.fr
Phone : (33)-3-88-41-40-07  Fax : (33)-3-88-41-40-99

Re: [RFA]Use prev_lexptr in *-exp.y

[Michael Snyder]

Pierre Muller wrote:
> 
> At 20:38 25/04/2002 , Michael Snyder a écrit:
> >Pierre Muller wrote:
> > >
> > > At 17:30 25/04/2002 , Andrew Cagney a écrit:
> > > >>2002-04-25  Pierre Muller  <muller@ics.u-strasbg.fr>
> > > >>* p-exp.y: Also use new prev_lexptr variable
> > > >>         to improve error reporting. Based on Michael Snyder
> > > >>         2002-04-24 dated patch to c-exp.y.
> > > >
> > > >Nice.  Can I suggest creating bug report so we know that the others should be done.
> > >
> > > This is more work than fixing them directly!
> >
> >ROTFL!  You rock, Pierre!   <g>
> 
> May I consider this as an approval and commit this patch?

Aha, the files have no maintainer!  OK, yes, donning my omnipotence hat, 
I approve your change.  ;-)

Re: [RFA]Use prev_lexptr in *-exp.y

[Pierre Muller]

At 19:49 02/05/2002 , Michael Snyder a écrit:
>Pierre Muller wrote:
> > 
> > At 20:38 25/04/2002 , Michael Snyder a écrit:
> > >Pierre Muller wrote:
> > > >
> > > > At 17:30 25/04/2002 , Andrew Cagney a écrit:
> > > > >>2002-04-25  Pierre Muller  <muller@ics.u-strasbg.fr>
> > > > >>* p-exp.y: Also use new prev_lexptr variable
> > > > >>         to improve error reporting. Based on Michael Snyder
> > > > >>         2002-04-24 dated patch to c-exp.y.
> > > > >
> > > > >Nice.  Can I suggest creating bug report so we know that the others should be done.
> > > >
> > > > This is more work than fixing them directly!
> > >
> > >ROTFL!  You rock, Pierre!   <g>
> > 
> > May I consider this as an approval and commit this patch?
>
>Aha, the files have no maintainer!  OK, yes, donning my omnipotence hat, 
>I approve your change.  ;-)

Thanks, committed.

By the way, I don't have the necessary stuff to run the 
testsuite... Thus I can not warranty that there are no 
regressions. There mught be some wrong regressions,
if some testsuite do check the error output in case of 
wrong expression entered. The error output should be more
logical now, but could get some tests  to fail.



Pierre Muller
Institut Charles Sadron
6,rue Boussingault
F 67083 STRASBOURG CEDEX (France)
mailto:muller@ics.u-strasbg.fr
Phone : (33)-3-88-41-40-07  Fax : (33)-3-88-41-40-99

Re: [RFA]Use prev_lexptr in *-exp.y

[Michael Snyder]

Pierre Muller wrote:
> 
> At 19:49 02/05/2002 , Michael Snyder a écrit:
> >Pierre Muller wrote:
> > >
> > > At 20:38 25/04/2002 , Michael Snyder a écrit:
> > > >Pierre Muller wrote:
> > > > >
> > > > > At 17:30 25/04/2002 , Andrew Cagney a écrit:
> > > > > >>2002-04-25  Pierre Muller  <muller@ics.u-strasbg.fr>
> > > > > >>* p-exp.y: Also use new prev_lexptr variable
> > > > > >>         to improve error reporting. Based on Michael Snyder
> > > > > >>         2002-04-24 dated patch to c-exp.y.
> > > > > >
> > > > > >Nice.  Can I suggest creating bug report so we know that the others should be done.
> > > > >
> > > > > This is more work than fixing them directly!
> > > >
> > > >ROTFL!  You rock, Pierre!   <g>
> > >
> > > May I consider this as an approval and commit this patch?
> >
> >Aha, the files have no maintainer!  OK, yes, donning my omnipotence hat,
> >I approve your change.  ;-)
> 
> Thanks, committed.
> 
> By the way, I don't have the necessary stuff to run the
> testsuite... Thus I can not warranty that there are no
> regressions. There mught be some wrong regressions,
> if some testsuite do check the error output in case of
> wrong expression entered. The error output should be more
> logical now, but could get some tests  to fail.

Yes, I'm waiting for someone to complain.  None of the C or C++
tests have failed because of this (and I presume you've tested
pascal).