[RFA/RFC] printf (stderr,... ->fprintf_unfiltered (gdb_stderr,... in hpread.c

[RFA/RFC] printf (stderr,... ->fprintf_unfiltered (gdb_stderr,... in hpread.c

[Pierre Muller]

This is a second of the files that I listed in
http://sources.redhat.com/ml/gdb/2002-02/msg00212.html
as still having direct uses of stderr.

I did not fix a 80 char overflow, because I didn't find a good way to do it:
If I try to align the string "Psymtab for %s already read in.  Shouldn't happen.\n"
with gdb_stderr, I still get an overflow, how show I indent the args then?
Should I break the string constant?

ChangeLog entry:

2002-04-24  Pierre Muller  <ics.u-strasbg.fr>
	* hpread.c (hpread_psymtab_to_symtab_1,
	hpread_psymtab_to_symtab): Replace fprintf (stderr,...
	with fprintf_unfiltered (gdb_stderr,....

Index: hpread.c
===================================================================
RCS file: /cvs/src/src/gdb/hpread.c,v
retrieving revision 1.15
diff -u -p -r1.15 hpread.c
--- hpread.c    15 Apr 2002 04:52:08 -0000      1.15
+++ hpread.c    24 Apr 2002 12:41:18 -0000
@@ -2693,7 +2693,7 @@ hpread_psymtab_to_symtab_1 (struct parti
    /* Complain if we've already read in this symbol table.  */
    if (pst->readin)
      {
-      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
+      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
                pst->filename);
        return;
      }
@@ -2748,7 +2748,7 @@ hpread_psymtab_to_symtab (struct partial
    /* Sanity check.  */
    if (pst->readin)
      {
-      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
+      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
                pst->filename);
        return;
      }

Re: [RFA/RFC] printf (stderr,... ->fprintf_unfiltered (gdb_stderr,... in hpread.c

[Daniel Jacobowitz]

On Wed, Apr 24, 2002 at 03:16:16PM +0200, Pierre Muller wrote:
> This is a second of the files that I listed in
> http://sources.redhat.com/ml/gdb/2002-02/msg00212.html
> as still having direct uses of stderr.
> 
> I did not fix a 80 char overflow, because I didn't find a good way to do it:
> If I try to align the string "Psymtab for %s already read in.  Shouldn't happen.\n"
> with gdb_stderr, I still get an overflow, how show I indent the args then?
> Should I break the string constant?

Yes, I'd recommend:

-      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
+      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in.  "
				       "Shouldn't happen.\n",


-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: [RFA/RFC] printf (stderr,... ->fprintf_unfiltered (gdb_stderr,... in hpread.c

[Michael Snyder]

Daniel Jacobowitz wrote:
> 
> On Wed, Apr 24, 2002 at 03:16:16PM +0200, Pierre Muller wrote:
> > This is a second of the files that I listed in
> > http://sources.redhat.com/ml/gdb/2002-02/msg00212.html
> > as still having direct uses of stderr.
> >
> > I did not fix a 80 char overflow, because I didn't find a good way to do it:
> > If I try to align the string "Psymtab for %s already read in.  Shouldn't happen.\n"
> > with gdb_stderr, I still get an overflow, how show I indent the args then?
> > Should I break the string constant?
> 
> Yes, I'd recommend:
> 
> -      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
> +      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in.  "
>                                        "Shouldn't happen.\n",


I thought there was some kind of caveat against the use of
string concatenation ?

Re: [RFA/RFC] printf (stderr,... ->fprintf_unfiltered(gdb_stderr,... in hpread.c

[Michael Snyder]

Pierre Muller wrote:
> 
> This is a second of the files that I listed in
> http://sources.redhat.com/ml/gdb/2002-02/msg00212.html
> as still having direct uses of stderr.
> 
> I did not fix a 80 char overflow, because I didn't find a good way to do it:
> If I try to align the string "Psymtab for %s already read in.  Shouldn't happen.\n"
> with gdb_stderr, I still get an overflow, how show I indent the args then?
> Should I break the string constant?

I've spent more time agonizing over such questions than I care to
admit.  ;-)
And I'm fairly strongly opinionated against greater-than-80-char lines.
But I think when it comes to printf and string constants, we have to be
a little loose about it.

If you wanted to do something really perverted, you could try this:

   fprintf_unfiltered (gdb_stderr, 
                       "Psymtab for %s %s %s\n", 
			pst->filename, 
			"already read in. ",
			"Shouldn't happen.");

> 
> ChangeLog entry:
> 
> 2002-04-24  Pierre Muller  <ics.u-strasbg.fr>
>         * hpread.c (hpread_psymtab_to_symtab_1,
>         hpread_psymtab_to_symtab): Replace fprintf (stderr,...
>         with fprintf_unfiltered (gdb_stderr,....
> 
> Index: hpread.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/hpread.c,v
> retrieving revision 1.15
> diff -u -p -r1.15 hpread.c
> --- hpread.c    15 Apr 2002 04:52:08 -0000      1.15
> +++ hpread.c    24 Apr 2002 12:41:18 -0000
> @@ -2693,7 +2693,7 @@ hpread_psymtab_to_symtab_1 (struct parti
>     /* Complain if we've already read in this symbol table.  */
>     if (pst->readin)
>       {
> -      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
> +      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
>                 pst->filename);
>         return;
>       }
> @@ -2748,7 +2748,7 @@ hpread_psymtab_to_symtab (struct partial
>     /* Sanity check.  */
>     if (pst->readin)
>       {
> -      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
> +      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
>                 pst->filename);
>         return;
>       }

Re: [RFA/RFC] printf (stderr,... ->fprintf_unfiltered (gdb_stderr,... in hpread.c

[Daniel Jacobowitz]

On Wed, Apr 24, 2002 at 11:50:40AM -0700, Michael Snyder wrote:
> Daniel Jacobowitz wrote:
> > 
> > On Wed, Apr 24, 2002 at 03:16:16PM +0200, Pierre Muller wrote:
> > > This is a second of the files that I listed in
> > > http://sources.redhat.com/ml/gdb/2002-02/msg00212.html
> > > as still having direct uses of stderr.
> > >
> > > I did not fix a 80 char overflow, because I didn't find a good way to do it:
> > > If I try to align the string "Psymtab for %s already read in.  Shouldn't happen.\n"
> > > with gdb_stderr, I still get an overflow, how show I indent the args then?
> > > Should I break the string constant?
> > 
> > Yes, I'd recommend:
> > 
> > -      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
> > +      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in.  "
> >                                        "Shouldn't happen.\n",
> 
> 
> I thought there was some kind of caveat against the use of
> string concatenation ?

Perhaps you're thinking of token concatenation, or stringification?  Or
multiline strings?  I believe "a" "b" has worked even in K&R compilers.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: [RFA/RFC] printf (stderr,... ->fprintf_unfiltered(gdb_stderr,... in hpread.c

[Daniel Jacobowitz]

On Wed, Apr 24, 2002 at 11:49:54AM -0700, Michael Snyder wrote:
> Pierre Muller wrote:
> > 
> > This is a second of the files that I listed in
> > http://sources.redhat.com/ml/gdb/2002-02/msg00212.html
> > as still having direct uses of stderr.
> > 
> > I did not fix a 80 char overflow, because I didn't find a good way to do it:
> > If I try to align the string "Psymtab for %s already read in.  Shouldn't happen.\n"
> > with gdb_stderr, I still get an overflow, how show I indent the args then?
> > Should I break the string constant?
> 
> I've spent more time agonizing over such questions than I care to
> admit.  ;-)
> And I'm fairly strongly opinionated against greater-than-80-char lines.
> But I think when it comes to printf and string constants, we have to be
> a little loose about it.
> 
> If you wanted to do something really perverted, you could try this:
> 
>    fprintf_unfiltered (gdb_stderr, 
>                        "Psymtab for %s %s %s\n", 
> 			pst->filename, 
> 			"already read in. ",
> 			"Shouldn't happen.");

Please don't... if we someday actually use i18n, we'll just have to
kill all of these anyway.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: [RFA/RFC] printf (stderr,... ->fprintf_unfiltered (gdb_stderr,... in hpread.c

[Andrew Cagney]

> On Wed, Apr 24, 2002 at 11:50:40AM -0700, Michael Snyder wrote:
> 
>> Daniel Jacobowitz wrote:
> 
>> > 
>> > On Wed, Apr 24, 2002 at 03:16:16PM +0200, Pierre Muller wrote:
> 
>> > > This is a second of the files that I listed in
>> > > http://sources.redhat.com/ml/gdb/2002-02/msg00212.html
>> > > as still having direct uses of stderr.
>> > >
>> > > I did not fix a 80 char overflow, because I didn't find a good way to do it:
>> > > If I try to align the string "Psymtab for %s already read in.  Shouldn't happen.\n"
>> > > with gdb_stderr, I still get an overflow, how show I indent the args then?
>> > > Should I break the string constant?
> 
>> > 
>> > Yes, I'd recommend:
>> > 
>> > -      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
>> > +      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in.  "
>> > "Shouldn't happen.\n",

The other is the more traditional:

	..... "\
Psymtab for ......\n\
.......\n"

(the ``\'' at the end of the line is needed.

>> 
>> 
>> I thought there was some kind of caveat against the use of
>> string concatenation ?
> 
> 
> Perhaps you're thinking of token concatenation, or stringification?  Or
> multiline strings?  I believe "a" "b" has worked even in K&R compilers.

Some, not all.  But yes string concatenation like the above is fine for GDB.

(Pierre, yes, ok with a tweak).

Andrew

Re: [RFA/RFC] printf (stderr,... ->fprintf_unfiltered (gdb_stderr,... in hpread.c

[Pierre Muller]

At 17:38 25/04/2002 , Andrew Cagney a écrit:
>>On Wed, Apr 24, 2002 at 11:50:40AM -0700, Michael Snyder wrote:
>>
>>>Daniel Jacobowitz wrote:
>>
>>> > > On Wed, Apr 24, 2002 at 03:16:16PM +0200, Pierre Muller wrote:
>>
>>> > > This is a second of the files that I listed in
>>> > > http://sources.redhat.com/ml/gdb/2002-02/msg00212.html
>>> > > as still having direct uses of stderr.
>>> > >
>>> > > I did not fix a 80 char overflow, because I didn't find a good way to do it:
>>> > > If I try to align the string "Psymtab for %s already read in.  Shouldn't happen.\n"
>>> > > with gdb_stderr, I still get an overflow, how show I indent the args then?
>>> > > Should I break the string constant?
>>
>>> > > Yes, I'd recommend:
>>> > > -      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
>>> > +      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in.  "
>>> > "Shouldn't happen.\n",
>
>The other is the more traditional:
>
>         ..... "\
>Psymtab for ......\n\
>.......\n"
>
>(the ``\'' at the end of the line is needed.
>
>>>
>>>I thought there was some kind of caveat against the use of
>>>string concatenation ?
>>
>>Perhaps you're thinking of token concatenation, or stringification?  Or
>>multiline strings?  I believe "a" "b" has worked even in K&R compilers.
>
>Some, not all.  But yes string concatenation like the above is fine for GDB.
>
>(Pierre, yes, ok with a tweak).
Sorry, but I din't understand which version of the
modification I should use here...



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/RFC] printf (stderr,... ->fprintf_unfiltered (gdb_stderr,... in hpread.c

[Andrew Cagney]

> (Pierre, yes, ok with a tweak).
> 
> Sorry, but I din't understand which version of the
> modification I should use here...

Either is acceptable and either works.  (Personally, it really doesn't 
worry me :-)  I suspect danielJ slightly prefers the ISO C form, so 
perhaps go with that.

Andrew

Re: [RFA/RFC] printf (stderr,... ->fprintf_unfiltered (gdb_stderr,... in hpread.c

[Daniel Jacobowitz]

On Thu, Apr 25, 2002 at 01:16:15PM -0400, Andrew Cagney wrote:
> >(Pierre, yes, ok with a tweak).
> >
> >Sorry, but I din't understand which version of the
> >modification I should use here...
> 
> Either is acceptable and either works.  (Personally, it really doesn't 
> worry me :-)  I suspect danielJ slightly prefers the ISO C form, so 
> perhaps go with that.

Nod.  I find that the \-newline form is uglier w.r.t. indentation, so
if no one else cares...

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

[RFA 2nd] hpread.c printf (stderr,... ->fprintf_unfiltered (gdb_stderr,..

[Pierre Muller]

At 15:16 24/04/2002 , Pierre Muller a écrit:
>This is a second of the files that I listed in
>http://sources.redhat.com/ml/gdb/2002-02/msg00212.html
>as still having direct uses of stderr.
>
>I did not fix a 80 char overflow, because I didn't find a good way to do it:
>If I try to align the string "Psymtab for %s already read in.  Shouldn't happen.\n"
>with gdb_stderr, I still get an overflow, how show I indent the args then?
>Should I break the string constant?

This is a second version which replaces
[RFA/RFC] printf (stderr,... ->fprintf_unfiltered (gdb_stderr,... in hpread.c
http://sources.redhat.com/ml/gdb-patches/2002-04/msg00930.html

I finally cut the string in two part,
but like for the other patch, I can't compile it to test
the patch :(

In that case I even find the compilation error
rather strange:
$ (cd ../../build/gdb;make hpread.o)
gcc -c -g -O2    -I. -I../../src/gdb -I../../src/gdb/config -DHAVE_CONFIG_H -I..
/../src/gdb/../include/opcode -I../../src/gdb/../readline/.. -I../bfd -I../../sr
c/gdb/../bfd  -I../../src/gdb/../include -I../intl -I../../src/gdb/../intl  -DMI
_OUT=1 -DUI_OUT=1 -Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wpare
ntheses -Wpointer-arith -Wuninitialized  ../../src/gdb/hpread.c
../../src/gdb/hpread.c:29: syms.h: No such file or directory
make: *** [hpread.o] Error 1

in hpread.c
there is a line with
#include "syms.h"
but there is no syms.h 
in the gdb sources...

If I do a 'grep -n syms.h *.c in gdb directory,
I get only this:
$ grep -n syms.h *.c
hpread.c:29:#include "syms.h"
somread.c:25:#include <syms.h>

Is the hpread.c file obsolete?
Or shouldn't it be also
#include <syms.h>
like in somread.c?

ChangeLog entry unchanged.

>ChangeLog entry:
>
>2002-04-24  Pierre Muller  <ics.u-strasbg.fr>
>         * hpread.c (hpread_psymtab_to_symtab_1,
>         hpread_psymtab_to_symtab): Replace fprintf (stderr,...
>         with fprintf_unfiltered (gdb_stderr,....


Index: hpread.c
===================================================================
RCS file: /cvs/src/src/gdb/hpread.c,v
retrieving revision 1.15
diff -u -p -r1.15 hpread.c
--- hpread.c    15 Apr 2002 04:52:08 -0000      1.15
+++ hpread.c    26 Apr 2002 09:01:32 -0000
@@ -2693,8 +2693,9 @@ hpread_psymtab_to_symtab_1 (struct parti
    /* Complain if we've already read in this symbol table.  */
    if (pst->readin)
      {
-      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
-              pst->filename);
+      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in."
+                         "  Shouldn't happen.\n",
+                         pst->filename);
        return;
      }

@@ -2748,8 +2749,9 @@ hpread_psymtab_to_symtab (struct partial
    /* Sanity check.  */
    if (pst->readin)
      {
-      fprintf (stderr, "Psymtab for %s already read in.  Shouldn't happen.\n",
-              pst->filename);
+      fprintf_unfiltered (gdb_stderr, "Psymtab for %s already read in."
+                         "  Shouldn't happen.\n",
+                         pst->filename);
        return;
      }







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 2nd] hpread.c printf (stderr,... ->fprintf_unfiltered (gdb_stderr,..

[Andrew Cagney]

> in hpread.c
> there is a line with
> #include "syms.h"
> but there is no syms.h 
> in the gdb sources...
> 
> If I do a 'grep -n syms.h *.c in gdb directory,
> I get only this:
> $ grep -n syms.h *.c
> hpread.c:29:#include "syms.h"
> somread.c:25:#include <syms.h>
> 
> Is the hpread.c file obsolete?
> Or shouldn't it be also
> #include <syms.h>
> like in somread.c?

I need to tread carefuly here, skating on thin ice :-)  There is the 
strict ISO C defined behavour of "" vs <> and then there are accepted 
conventions (note plural).

Within GDB, the accepted convention is to use <> as as "syms.h" is a 
system header.

In theory, it should be possible for you to compile hpread.c on any 
system, because of includes like the above, it isn't (hence the 
MAINTAINERS file marks it as broken).  Consequently, yes, ok.

Andrew

Re: [RFA 2nd] hpread.c printf (stderr,... ->fprintf_unfiltered (gdb_stderr,..

[Pierre Muller]

At 15:32 26/04/2002 , Andrew Cagney a écrit:
>>in hpread.c
>>there is a line with
>>#include "syms.h"
>>but there is no syms.h in the gdb sources...
>>If I do a 'grep -n syms.h *.c in gdb directory,
>>I get only this:
>>$ grep -n syms.h *.c
>>hpread.c:29:#include "syms.h"
>>somread.c:25:#include <syms.h>
>>Is the hpread.c file obsolete?
>>Or shouldn't it be also
>>#include <syms.h>
>>like in somread.c?
>
>I need to tread carefuly here, skating on thin ice :-)  There is the strict ISO C defined behavour of "" vs <> and then there are accepted conventions (note plural).
>
>Within GDB, the accepted convention is to use <> as as "syms.h" is a system header.
>
>In theory, it should be possible for you to compile hpread.c on any system, because of includes like the above, it isn't (hence the MAINTAINERS file marks it as broken).  Consequently, yes, ok.
>
>Andrew

I hope I did understand you correctly.
I can commit this patch without testing that it complies correctly
because it is known to belong to a broken target.

Thus, I committed it.



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