[rfc] Print MI fullname even for non-existing files

[rfc] Print MI fullname even for non-existing files

[Jan Kratochvil]

Hi,

symtab_to_fullname now always returns non-NULL.

The goal is that currently with deleted /usr/src/debug one gets:

0x00007ffff637d7d0 in __poll_nocancel () at ../sysdeps/unix/syscall-template.S:81
81      ../sysdeps/unix/syscall-template.S: No such file or directory.

but I would like to get rather:

Program received signal SIGINT, Interrupt.
0x00007ffff637d7d0 in __poll_nocancel () at ../sysdeps/unix/syscall-template.S:81
81      /usr/src/debug/glibc-2.16-75f0d304/sysdeps/unix/syscall-template.S: No such file or directory.

But this patch does not do that part yet, it will be done by the new switch to
fully absolute filenames everywhere, as is being worked on top of this patch
as a rewrite of:
	[patchv2] GDB 7.2: new feature for "backtrace" that cuts path to file (remain filename)
	http://sourceware.org/ml/gdb-patches/2012-11/msg00957.html
and
	set filename-display absolute

Currently this patch only changes from user point of view MI:
(gdb) -file-list-exec-source-file
	^done,line="1",file="deleted.c",macro-info="0"
->
	^done,line="1",file="deleted.c",fullname="/path/to/deleted.c",macro-info="0"

Is there a serious reason why fullname was suppressed for non-existing files?
While it breaks MI compatibility in some way I find it an acceptable change.

No regressions on {x86_64,x86_64-m32,i686}-fedora18-linux-gnu.


Thanks,
Jan


2012-12-17  Jan Kratochvil  <jan.kratochvil@redhat.com>

	* ada-lang.c (is_known_support_routine): New variable fullname.  Use
	access call to verify the symtab_to_fullname result.
	* breakpoint.c (print_breakpoint_location, update_static_tracepoint):
	Remove NULL check of symtab_to_fullname result.
	* cli/cli-cmds.c (edit_command): Likewise.
	* mi/mi-cmd-file.c (mi_cmd_file_list_exec_source_file)
	(mi_cmd_file_list_exec_source_files): Likewise.
	* python/py-symtab.c (stpy_fullname): Likewise.
	* source.c (symtab_to_fullname): Update function comment.  Rename
	variable r to fd, move it to inner block.  Always provide non-NULL
	result.
	(print_source_lines_base): Remove NULL check of symtab_to_fullname
	result.
	* stack.c (print_frame): Likewise.
	* symtab.c (iterate_over_some_symtabs, find_line_symtab, sources_info):
	Likewise.
	* tracepoint.c (print_one_static_tracepoint_marker): Likewise.

gdb/doc/
2012-12-17  Jan Kratochvil  <jan.kratochvil@redhat.com>

	* gdb.texinfo (GDB/MI Data Manipulation) (fullname): Make it always
	present.
	(GDB/MI File Commands) (-file-list-exec-source-files): Make the
	fullname output always present.

gdb/testsuite/
2012-12-17  Jan Kratochvil  <jan.kratochvil@redhat.com>

	* gdb.mi/mi-fullname-deleted.exp: New file.

diff --git a/gdb/ada-lang.c b/gdb/ada-lang.c
index f07e4bc..cb7313b 100644
--- a/gdb/ada-lang.c
+++ b/gdb/ada-lang.c
@@ -11084,6 +11084,7 @@ is_known_support_routine (struct frame_info *frame)
   const char *func_name;
   enum language func_lang;
   int i;
+  const char *fullname;
 
   /* If this code does not have any debugging information (no symtab),
      This cannot be any user code.  */
@@ -11098,7 +11099,8 @@ is_known_support_routine (struct frame_info *frame)
      for the user.  This should also take care of case such as VxWorks
      where the kernel has some debugging info provided for a few units.  */
 
-  if (symtab_to_fullname (sal.symtab) == NULL)
+  fullname = symtab_to_fullname (sal.symtab);
+  if (access (fullname, R_OK) != 0)
     return 1;
 
   /* Check the unit filename againt the Ada runtime file naming.
diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c
index c2ce96c..d587cb4 100644
--- a/gdb/breakpoint.c
+++ b/gdb/breakpoint.c
@@ -5694,8 +5694,7 @@ print_breakpoint_location (struct breakpoint *b,
 	  struct symtab_and_line sal = find_pc_line (loc->address, 0);
 	  const char *fullname = symtab_to_fullname (sal.symtab);
 	  
-	  if (fullname)
-	    ui_out_field_string (uiout, "fullname", fullname);
+	  ui_out_field_string (uiout, "fullname", fullname);
 	}
       
       ui_out_field_int (uiout, "line", loc->line_number);
@@ -13872,8 +13871,7 @@ update_static_tracepoint (struct breakpoint *b, struct symtab_and_line sal)
 	    {
 	      const char *fullname = symtab_to_fullname (sal2.symtab);
 
-	      if (fullname)
-		ui_out_field_string (uiout, "fullname", fullname);
+	      ui_out_field_string (uiout, "fullname", fullname);
 	    }
 
 	  ui_out_field_int (uiout, "line", sal2.line);
diff --git a/gdb/cli/cli-cmds.c b/gdb/cli/cli-cmds.c
index b65262e..fe84256 100644
--- a/gdb/cli/cli-cmds.c
+++ b/gdb/cli/cli-cmds.c
@@ -842,16 +842,7 @@ edit_command (char *arg, int from_tty)
   if ((editor = (char *) getenv ("EDITOR")) == NULL)
       editor = "/bin/ex";
 
-  /* If we don't already know the full absolute file name of the
-     source file, find it now.  */
-  if (!sal.symtab->fullname)
-    {
-      fn = symtab_to_fullname (sal.symtab);
-      if (!fn)
-	fn = "unknown";
-    }
-  else
-    fn = sal.symtab->fullname;
+  fn = symtab_to_fullname (sal.symtab);
 
   /* Quote the file name, in case it has whitespace or other special
      characters.  */
diff --git a/gdb/mi/mi-cmd-file.c b/gdb/mi/mi-cmd-file.c
index 0b2b725..018f505 100644
--- a/gdb/mi/mi-cmd-file.c
+++ b/gdb/mi/mi-cmd-file.c
@@ -48,16 +48,11 @@ mi_cmd_file_list_exec_source_file (char *command, char **argv, int argc)
   if (!st.symtab)
     error (_("-file-list-exec-source-file: No symtab"));
 
-  /* Extract the fullname if it is not known yet.  */
-  symtab_to_fullname (st.symtab);
-
   /* Print to the user the line, filename and fullname.  */
   ui_out_field_int (uiout, "line", st.line);
   ui_out_field_string (uiout, "file", st.symtab->filename);
 
-  /* We may not be able to open the file (not available). */
-  if (st.symtab->fullname)
-  ui_out_field_string (uiout, "fullname", st.symtab->fullname);
+  ui_out_field_string (uiout, "fullname", symtab_to_fullname (st.symtab));
 
   ui_out_field_int (uiout, "macro-info", st.symtab->macro_table ? 1 : 0);
 }
@@ -99,12 +94,7 @@ mi_cmd_file_list_exec_source_files (char *command, char **argv, int argc)
     ui_out_begin (uiout, ui_out_type_tuple, NULL);
 
     ui_out_field_string (uiout, "file", s->filename);
-
-    /* Extract the fullname if it is not known yet.  */
-    symtab_to_fullname (s);
-
-    if (s->fullname)
-      ui_out_field_string (uiout, "fullname", s->fullname);
+    ui_out_field_string (uiout, "fullname", symtab_to_fullname (s));
 
     ui_out_end (uiout, ui_out_type_tuple);
   }
diff --git a/gdb/python/py-symtab.c b/gdb/python/py-symtab.c
index d3e95db..deb7330 100644
--- a/gdb/python/py-symtab.c
+++ b/gdb/python/py-symtab.c
@@ -132,11 +132,8 @@ stpy_fullname (PyObject *self, PyObject *args)
   STPY_REQUIRE_VALID (self, symtab);
 
   fullname = symtab_to_fullname (symtab);
-  if (fullname)
-    return PyString_Decode (fullname, strlen (fullname),
-			    host_charset (), NULL);
 
-  Py_RETURN_NONE;
+  return PyString_Decode (fullname, strlen (fullname), host_charset (), NULL);
 }
 
 /* Implementation of gdb.Symtab.is_valid (self) -> Boolean.
diff --git a/gdb/source.c b/gdb/source.c
index 7e14fc6..6e65d1a 100644
--- a/gdb/source.c
+++ b/gdb/source.c
@@ -1075,35 +1075,32 @@ open_source_file (struct symtab *s)
 
 /* Finds the fullname that a symtab represents.
 
-   If this functions finds the fullname, it will save it in s->fullname
-   and it will also return the value.
+   This functions finds the fullname and saves it in s->fullname.
+   It will also return the value.
 
    If this function fails to find the file that this symtab represents,
-   NULL will be returned and s->fullname will be set to NULL.  */
+   the expected fullname is used.  Therefore the files does not have to
+   exist.  */
 
 const char *
 symtab_to_fullname (struct symtab *s)
 {
-  int r;
-
-  if (!s)
-    return NULL;
-
   /* Use cached copy if we have it.
      We rely on forget_cached_source_info being called appropriately
      to handle cases like the file being moved.  */
-  if (s->fullname)
-    return s->fullname;
-
-  r = find_and_open_source (s->filename, s->dirname, &s->fullname);
-
-  if (r >= 0)
+  if (s->fullname == NULL)
     {
-      close (r);
-      return s->fullname;
-    }
+      int fd = find_and_open_source (s->filename, s->dirname, &s->fullname);
 
-  return NULL;
+      if (fd >= 0)
+	close (fd);
+      else if (s->dirname == NULL)
+	s->fullname = xstrdup (s->filename);
+      else
+	s->fullname = concat (s->dirname, SLASH_STRING, s->filename, NULL);
+    } 
+
+  return s->fullname;
 }
 \f
 /* Create and initialize the table S->line_charpos that records
@@ -1306,8 +1303,7 @@ print_source_lines_base (struct symtab *s, int line, int stopline,
 	    {
 	      const char *fullname = symtab_to_fullname (s);
 
-	      if (fullname != NULL)
-		ui_out_field_string (uiout, "fullname", fullname);
+	      ui_out_field_string (uiout, "fullname", fullname);
 	    }
 	  ui_out_text (uiout, "\n");
 	}
diff --git a/gdb/stack.c b/gdb/stack.c
index e246a68..6cf6052 100644
--- a/gdb/stack.c
+++ b/gdb/stack.c
@@ -1189,8 +1189,7 @@ print_frame (struct frame_info *frame, int print_level,
 	{
 	  const char *fullname = symtab_to_fullname (sal.symtab);
 
-	  if (fullname != NULL)
-	    ui_out_field_string (uiout, "fullname", fullname);
+	  ui_out_field_string (uiout, "fullname", fullname);
 	}
       annotate_frame_source_file_end ();
       ui_out_text (uiout, ":");
diff --git a/gdb/symtab.c b/gdb/symtab.c
index 0fd75e5..e3c0700 100644
--- a/gdb/symtab.c
+++ b/gdb/symtab.c
@@ -225,13 +225,13 @@ iterate_over_some_symtabs (const char *name,
       {
         const char *fp = symtab_to_fullname (s);
 
-        if (fp != NULL && FILENAME_CMP (full_path, fp) == 0)
+        if (FILENAME_CMP (full_path, fp) == 0)
           {
 	    if (callback (s, data))
 	      return 1;
           }
 
-	if (fp != NULL && !is_abs && compare_filenames_for_search (fp, name))
+	if (!is_abs && compare_filenames_for_search (fp, name))
 	  {
 	    if (callback (s, data))
 	      return 1;
@@ -241,24 +241,20 @@ iterate_over_some_symtabs (const char *name,
     if (real_path != NULL)
       {
         const char *fullname = symtab_to_fullname (s);
+	char *rp = gdb_realpath (fullname);
 
-        if (fullname != NULL)
-          {
-            char *rp = gdb_realpath (fullname);
-
-            make_cleanup (xfree, rp);
-            if (FILENAME_CMP (real_path, rp) == 0)
-	      {
-		if (callback (s, data))
-		  return 1;
-	      }
+	make_cleanup (xfree, rp);
+	if (FILENAME_CMP (real_path, rp) == 0)
+	  {
+	    if (callback (s, data))
+	      return 1;
+	  }
 
-	    if (!is_abs && compare_filenames_for_search (rp, name))
-	      {
-		if (callback (s, data))
-		  return 1;
-	      }
-          }
+	if (!is_abs && compare_filenames_for_search (rp, name))
+	  {
+	    if (callback (s, data))
+	      return 1;
+	  }
       }
     }
 
@@ -2552,9 +2548,6 @@ find_line_symtab (struct symtab *symtab, int line,
 							 symtab->filename);
       }
 
-      /* Get symbol full file name if possible.  */
-      symtab_to_fullname (symtab);
-
       ALL_SYMTABS (objfile, s)
       {
 	struct linetable *l;
@@ -2562,9 +2555,7 @@ find_line_symtab (struct symtab *symtab, int line,
 
 	if (FILENAME_CMP (symtab->filename, s->filename) != 0)
 	  continue;
-	if (symtab->fullname != NULL
-	    && symtab_to_fullname (s) != NULL
-	    && FILENAME_CMP (symtab->fullname, s->fullname) != 0)
+	if (FILENAME_CMP (symtab->fullname, symtab_to_fullname (s)) != 0)
 	  continue;	
 	l = LINETABLE (s);
 	ind = find_line_common (l, line, &exact, 0);
@@ -3295,7 +3286,7 @@ sources_info (char *ignore, int from_tty)
   {
     const char *fullname = symtab_to_fullname (s);
 
-    output_source_filename (fullname ? fullname : s->filename, &data);
+    output_source_filename (fullname, &data);
   }
   printf_filtered ("\n\n");
 
diff --git a/gdb/tracepoint.c b/gdb/tracepoint.c
index 848179a..2ab8323 100644
--- a/gdb/tracepoint.c
+++ b/gdb/tracepoint.c
@@ -4872,8 +4872,7 @@ print_one_static_tracepoint_marker (int count,
 	{
 	  const char *fullname = symtab_to_fullname (sal.symtab);
 
-	  if (fullname)
-	    ui_out_field_string (uiout, "fullname", fullname);
+	  ui_out_field_string (uiout, "fullname", fullname);
 	}
       else
 	ui_out_field_skip (uiout, "fullname");
diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
index 5abcd93..c40cc82 100644
--- a/gdb/doc/gdb.texinfo
+++ b/gdb/doc/gdb.texinfo
@@ -31086,9 +31086,13 @@ file name or a relative file name depending on the compile command
 used.
 
 @item fullname
-This field is optional.  If it is present it will contain an absolute
-file name of @samp{file}.  If this field is not present then
-@value{GDBN} was unable to determine the absolute file name.
+Absolute file name of @samp{file}.  It is converted to a canonical form
+using the source file search path
+(@pxref{Source Path, ,Specifying Source Directories}).
+and after resolving all the symbolic links.
+
+If the source file is not found this field will contain the path as
+present in the debug information.
 
 @item line_asm_insn
 This is a list of tuples containing the disassembly for @samp{line} in
@@ -32269,8 +32273,8 @@ The @value{GDBN} equivalent is @samp{info source}
 
 List the source files for the current executable.
 
-It will always output the filename, but only when @value{GDBN} can find
-the absolute file name of a source file, will it output the fullname.
+It will always output both the filename and fullname (absolute file
+name) of a source file.
 
 @subsubheading @value{GDBN} Command
 
diff --git a/gdb/testsuite/gdb.mi/mi-fullname-deleted.exp b/gdb/testsuite/gdb.mi/mi-fullname-deleted.exp
new file mode 100644
index 0000000..c40f6b2
--- /dev/null
+++ b/gdb/testsuite/gdb.mi/mi-fullname-deleted.exp
@@ -0,0 +1,40 @@
+# Copyright 2012 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+load_lib mi-support.exp
+set MIFLAGS "-i=mi"
+
+gdb_exit
+if [mi_gdb_start] {
+    continue
+}
+
+standard_testfile
+
+set f [open ${objdir}/${subdir}/${srcfile} "w"]
+puts $f "int main (void) { return 0; }"
+close $f
+
+if { [gdb_compile "${objdir}/${subdir}/${srcfile}" "${binfile}" executable {debug}] != "" } {
+    untested $testname
+    return -1
+}
+
+file delete -- ${objdir}/${subdir}/${srcfile}
+
+mi_gdb_reinitialize_dir $srcdir/$subdir
+mi_gdb_load ${binfile}
+
+mi_gdb_test "-file-list-exec-source-file" ".*\",fullname=\"[string_to_regexp ${objdir}/${subdir}/${srcfile}]\".*" "fullname present"

Re: [rfc] Print MI fullname even for non-existing files

[Andrew Burgess]

On 17/12/2012 3:58 PM, Jan Kratochvil wrote:

> Currently this patch only changes from user point of view MI:
> (gdb) -file-list-exec-source-file
> 	^done,line="1",file="deleted.c",macro-info="0"
> ->
> 	^done,line="1",file="deleted.c",fullname="/path/to/deleted.c",macro-info="0"

Seems like a good change to me.

> Is there a serious reason why fullname was suppressed for non-existing files?

I added the fullname field in source.c, there was no good reason then
for dropping the field if we got NULL back except that (1) it matched
the behaviour in stack.c, and (2) if we got null back then I'd have just
left the field empty which didn't feel very useful.

> While it breaks MI compatibility in some way I find it an acceptable change.

I agree.

Thanks,
Andrew

Re: [rfc] Print MI fullname even for non-existing files

[Eli Zaretskii]

> Date: Mon, 17 Dec 2012 16:58:59 +0100
> From: Jan Kratochvil <jan.kratochvil@redhat.com>
> 
> +Absolute file name of @samp{file}.  It is converted to a canonical form
> +using the source file search path
> +(@pxref{Source Path, ,Specifying Source Directories}).
> +and after resolving all the symbolic links.
   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Is the last part some uncompleted sentence?  If it's part of the
previous sentence, the period there should go away.

The documentation part is OK with this gotcha fixed.

Thanks.

Re: [rfc] Print MI fullname even for non-existing files

[Jan Kratochvil]

On Mon, 17 Dec 2012 17:33:22 +0100, Eli Zaretskii wrote:
> > Date: Mon, 17 Dec 2012 16:58:59 +0100
> > From: Jan Kratochvil <jan.kratochvil@redhat.com>
> > 
> > +Absolute file name of @samp{file}.  It is converted to a canonical form
> > +using the source file search path
> > +(@pxref{Source Path, ,Specifying Source Directories}).
> > +and after resolving all the symbolic links.
>    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> Is the last part some uncompleted sentence?  If it's part of the
> previous sentence, the period there should go away.

That dot is a copy-paste error, thanks.


Jan

Re: [rfc] Print MI fullname even for non-existing files

[Tom Tromey]

>>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:

Jan> symtab_to_fullname now always returns non-NULL.

Jan> Currently this patch only changes from user point of view MI:
Jan> (gdb) -file-list-exec-source-file
Jan> 	^done,line="1",file="deleted.c",macro-info="0"
-> 
Jan> 	^done,line="1",file="deleted.c",fullname="/path/to/deleted.c",macro-info="0"

Jan> Is there a serious reason why fullname was suppressed for
Jan> non-existing files?  While it breaks MI compatibility in some way I
Jan> find it an acceptable change.

It seems pretty reasonable to me as well.

Jan> +set f [open ${objdir}/${subdir}/${srcfile} "w"]

standard_output_file, here and elsewhere.

Tom

Re: [rfc] Print MI fullname even for non-existing files

[Jan Kratochvil]

On Mon, 17 Dec 2012 17:12:23 +0100, Andrew Burgess wrote:
> On 17/12/2012 3:58 PM, Jan Kratochvil wrote:
> > Is there a serious reason why fullname was suppressed for non-existing files?
> 
> I added the fullname field in source.c, there was no good reason then
> for dropping the field if we got NULL back except that (1) it matched
> the behaviour in stack.c, and (2) if we got null back then I'd have just
> left the field empty which didn't feel very useful.

Thanks; although I was more curious why symtab_to_fullname itself returns NULL
for non-existing files, ignoring possibly present symtab->dirname.


> > While it breaks MI compatibility in some way I find it an acceptable change.
> 
> I agree.

Thanks for this confirmation.


Jan

Re: [rfc] Print MI fullname even for non-existing files

[Jan Kratochvil]

On Mon, 17 Dec 2012 19:33:06 +0100, Tom Tromey wrote:
> >>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:
> Jan> Is there a serious reason why fullname was suppressed for
> Jan> non-existing files?  While it breaks MI compatibility in some way I
> Jan> find it an acceptable change.
> 
> It seems pretty reasonable to me as well.

Great.


> Jan> +set f [open ${objdir}/${subdir}/${srcfile} "w"]
> 
> standard_output_file, here and elsewhere.

It cannot be, as it is ${objdir}, not ${srcdir}, as the source file needs to
be modified (=deleted).


Thanks,
Jan

Re: [rfc] Print MI fullname even for non-existing files

[Tom Tromey]

>>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:

Jan> +set f [open ${objdir}/${subdir}/${srcfile} "w"]

>> standard_output_file, here and elsewhere.

Jan> It cannot be, as it is ${objdir}, not ${srcdir}, as the source file
Jan> needs to be modified (=deleted).

standard_output_file makes a file name in objdir.

Tom

Re: [rfc] Print MI fullname even for non-existing files

[Jan Kratochvil]

On Mon, 17 Dec 2012 19:41:38 +0100, Tom Tromey wrote:
> Jan> It cannot be, as it is ${objdir}, not ${srcdir}, as the source file
> Jan> needs to be modified (=deleted).
> 
> standard_output_file makes a file name in objdir.

OK, sorry.


Jan


gdb/
2012-12-17  Jan Kratochvil  <jan.kratochvil@redhat.com>

	* ada-lang.c (is_known_support_routine): New variable fullname.  Use
	access call to verify the symtab_to_fullname result.
	* breakpoint.c (print_breakpoint_location, update_static_tracepoint):
	Remove NULL check of symtab_to_fullname result.
	* cli/cli-cmds.c (edit_command): Likewise.
	* mi/mi-cmd-file.c (mi_cmd_file_list_exec_source_file)
	(mi_cmd_file_list_exec_source_files): Likewise.
	* python/py-symtab.c (stpy_fullname): Likewise.
	* source.c (symtab_to_fullname): Update function comment.  Rename
	variable r to fd, move it to inner block.  Always provide non-NULL
	result.
	(print_source_lines_base): Remove NULL check of symtab_to_fullname
	result.
	* stack.c (print_frame): Likewise.
	* symtab.c (iterate_over_some_symtabs, find_line_symtab, sources_info):
	Likewise.
	* tracepoint.c (print_one_static_tracepoint_marker): Likewise.

gdb/doc/
2012-12-17  Jan Kratochvil  <jan.kratochvil@redhat.com>

	* gdb.texinfo (GDB/MI Data Manipulation) (fullname): Make it always
	present.
	(GDB/MI File Commands) (-file-list-exec-source-files): Make the
	fullname output always present.

gdb/testsuite/
2012-12-17  Jan Kratochvil  <jan.kratochvil@redhat.com>

	* gdb.mi/mi-fullname-deleted.exp: New file.

diff --git a/gdb/ada-lang.c b/gdb/ada-lang.c
index f07e4bc..cb7313b 100644
--- a/gdb/ada-lang.c
+++ b/gdb/ada-lang.c
@@ -11084,6 +11084,7 @@ is_known_support_routine (struct frame_info *frame)
   const char *func_name;
   enum language func_lang;
   int i;
+  const char *fullname;
 
   /* If this code does not have any debugging information (no symtab),
      This cannot be any user code.  */
@@ -11098,7 +11099,8 @@ is_known_support_routine (struct frame_info *frame)
      for the user.  This should also take care of case such as VxWorks
      where the kernel has some debugging info provided for a few units.  */
 
-  if (symtab_to_fullname (sal.symtab) == NULL)
+  fullname = symtab_to_fullname (sal.symtab);
+  if (access (fullname, R_OK) != 0)
     return 1;
 
   /* Check the unit filename againt the Ada runtime file naming.
diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c
index c2ce96c..d587cb4 100644
--- a/gdb/breakpoint.c
+++ b/gdb/breakpoint.c
@@ -5694,8 +5694,7 @@ print_breakpoint_location (struct breakpoint *b,
 	  struct symtab_and_line sal = find_pc_line (loc->address, 0);
 	  const char *fullname = symtab_to_fullname (sal.symtab);
 	  
-	  if (fullname)
-	    ui_out_field_string (uiout, "fullname", fullname);
+	  ui_out_field_string (uiout, "fullname", fullname);
 	}
       
       ui_out_field_int (uiout, "line", loc->line_number);
@@ -13872,8 +13871,7 @@ update_static_tracepoint (struct breakpoint *b, struct symtab_and_line sal)
 	    {
 	      const char *fullname = symtab_to_fullname (sal2.symtab);
 
-	      if (fullname)
-		ui_out_field_string (uiout, "fullname", fullname);
+	      ui_out_field_string (uiout, "fullname", fullname);
 	    }
 
 	  ui_out_field_int (uiout, "line", sal2.line);
diff --git a/gdb/cli/cli-cmds.c b/gdb/cli/cli-cmds.c
index b65262e..fe84256 100644
--- a/gdb/cli/cli-cmds.c
+++ b/gdb/cli/cli-cmds.c
@@ -842,16 +842,7 @@ edit_command (char *arg, int from_tty)
   if ((editor = (char *) getenv ("EDITOR")) == NULL)
       editor = "/bin/ex";
 
-  /* If we don't already know the full absolute file name of the
-     source file, find it now.  */
-  if (!sal.symtab->fullname)
-    {
-      fn = symtab_to_fullname (sal.symtab);
-      if (!fn)
-	fn = "unknown";
-    }
-  else
-    fn = sal.symtab->fullname;
+  fn = symtab_to_fullname (sal.symtab);
 
   /* Quote the file name, in case it has whitespace or other special
      characters.  */
diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
index 5abcd93..c1ba745 100644
--- a/gdb/doc/gdb.texinfo
+++ b/gdb/doc/gdb.texinfo
@@ -31086,9 +31086,13 @@ file name or a relative file name depending on the compile command
 used.
 
 @item fullname
-This field is optional.  If it is present it will contain an absolute
-file name of @samp{file}.  If this field is not present then
-@value{GDBN} was unable to determine the absolute file name.
+Absolute file name of @samp{file}.  It is converted to a canonical form
+using the source file search path
+(@pxref{Source Path, ,Specifying Source Directories})
+and after resolving all the symbolic links.
+
+If the source file is not found this field will contain the path as
+present in the debug information.
 
 @item line_asm_insn
 This is a list of tuples containing the disassembly for @samp{line} in
@@ -32269,8 +32273,8 @@ The @value{GDBN} equivalent is @samp{info source}
 
 List the source files for the current executable.
 
-It will always output the filename, but only when @value{GDBN} can find
-the absolute file name of a source file, will it output the fullname.
+It will always output both the filename and fullname (absolute file
+name) of a source file.
 
 @subsubheading @value{GDBN} Command
 
diff --git a/gdb/mi/mi-cmd-file.c b/gdb/mi/mi-cmd-file.c
index 0b2b725..018f505 100644
--- a/gdb/mi/mi-cmd-file.c
+++ b/gdb/mi/mi-cmd-file.c
@@ -48,16 +48,11 @@ mi_cmd_file_list_exec_source_file (char *command, char **argv, int argc)
   if (!st.symtab)
     error (_("-file-list-exec-source-file: No symtab"));
 
-  /* Extract the fullname if it is not known yet.  */
-  symtab_to_fullname (st.symtab);
-
   /* Print to the user the line, filename and fullname.  */
   ui_out_field_int (uiout, "line", st.line);
   ui_out_field_string (uiout, "file", st.symtab->filename);
 
-  /* We may not be able to open the file (not available). */
-  if (st.symtab->fullname)
-  ui_out_field_string (uiout, "fullname", st.symtab->fullname);
+  ui_out_field_string (uiout, "fullname", symtab_to_fullname (st.symtab));
 
   ui_out_field_int (uiout, "macro-info", st.symtab->macro_table ? 1 : 0);
 }
@@ -99,12 +94,7 @@ mi_cmd_file_list_exec_source_files (char *command, char **argv, int argc)
     ui_out_begin (uiout, ui_out_type_tuple, NULL);
 
     ui_out_field_string (uiout, "file", s->filename);
-
-    /* Extract the fullname if it is not known yet.  */
-    symtab_to_fullname (s);
-
-    if (s->fullname)
-      ui_out_field_string (uiout, "fullname", s->fullname);
+    ui_out_field_string (uiout, "fullname", symtab_to_fullname (s));
 
     ui_out_end (uiout, ui_out_type_tuple);
   }
diff --git a/gdb/python/py-symtab.c b/gdb/python/py-symtab.c
index d3e95db..deb7330 100644
--- a/gdb/python/py-symtab.c
+++ b/gdb/python/py-symtab.c
@@ -132,11 +132,8 @@ stpy_fullname (PyObject *self, PyObject *args)
   STPY_REQUIRE_VALID (self, symtab);
 
   fullname = symtab_to_fullname (symtab);
-  if (fullname)
-    return PyString_Decode (fullname, strlen (fullname),
-			    host_charset (), NULL);
 
-  Py_RETURN_NONE;
+  return PyString_Decode (fullname, strlen (fullname), host_charset (), NULL);
 }
 
 /* Implementation of gdb.Symtab.is_valid (self) -> Boolean.
diff --git a/gdb/source.c b/gdb/source.c
index 7e14fc6..6e65d1a 100644
--- a/gdb/source.c
+++ b/gdb/source.c
@@ -1075,35 +1075,32 @@ open_source_file (struct symtab *s)
 
 /* Finds the fullname that a symtab represents.
 
-   If this functions finds the fullname, it will save it in s->fullname
-   and it will also return the value.
+   This functions finds the fullname and saves it in s->fullname.
+   It will also return the value.
 
    If this function fails to find the file that this symtab represents,
-   NULL will be returned and s->fullname will be set to NULL.  */
+   the expected fullname is used.  Therefore the files does not have to
+   exist.  */
 
 const char *
 symtab_to_fullname (struct symtab *s)
 {
-  int r;
-
-  if (!s)
-    return NULL;
-
   /* Use cached copy if we have it.
      We rely on forget_cached_source_info being called appropriately
      to handle cases like the file being moved.  */
-  if (s->fullname)
-    return s->fullname;
-
-  r = find_and_open_source (s->filename, s->dirname, &s->fullname);
-
-  if (r >= 0)
+  if (s->fullname == NULL)
     {
-      close (r);
-      return s->fullname;
-    }
+      int fd = find_and_open_source (s->filename, s->dirname, &s->fullname);
 
-  return NULL;
+      if (fd >= 0)
+	close (fd);
+      else if (s->dirname == NULL)
+	s->fullname = xstrdup (s->filename);
+      else
+	s->fullname = concat (s->dirname, SLASH_STRING, s->filename, NULL);
+    } 
+
+  return s->fullname;
 }
 \f
 /* Create and initialize the table S->line_charpos that records
@@ -1306,8 +1303,7 @@ print_source_lines_base (struct symtab *s, int line, int stopline,
 	    {
 	      const char *fullname = symtab_to_fullname (s);
 
-	      if (fullname != NULL)
-		ui_out_field_string (uiout, "fullname", fullname);
+	      ui_out_field_string (uiout, "fullname", fullname);
 	    }
 	  ui_out_text (uiout, "\n");
 	}
diff --git a/gdb/stack.c b/gdb/stack.c
index b01c8f0..867c802 100644
--- a/gdb/stack.c
+++ b/gdb/stack.c
@@ -1189,8 +1189,7 @@ print_frame (struct frame_info *frame, int print_level,
 	{
 	  const char *fullname = symtab_to_fullname (sal.symtab);
 
-	  if (fullname != NULL)
-	    ui_out_field_string (uiout, "fullname", fullname);
+	  ui_out_field_string (uiout, "fullname", fullname);
 	}
       annotate_frame_source_file_end ();
       ui_out_text (uiout, ":");
diff --git a/gdb/symtab.c b/gdb/symtab.c
index 0fd75e5..e3c0700 100644
--- a/gdb/symtab.c
+++ b/gdb/symtab.c
@@ -225,13 +225,13 @@ iterate_over_some_symtabs (const char *name,
       {
         const char *fp = symtab_to_fullname (s);
 
-        if (fp != NULL && FILENAME_CMP (full_path, fp) == 0)
+        if (FILENAME_CMP (full_path, fp) == 0)
           {
 	    if (callback (s, data))
 	      return 1;
           }
 
-	if (fp != NULL && !is_abs && compare_filenames_for_search (fp, name))
+	if (!is_abs && compare_filenames_for_search (fp, name))
 	  {
 	    if (callback (s, data))
 	      return 1;
@@ -241,24 +241,20 @@ iterate_over_some_symtabs (const char *name,
     if (real_path != NULL)
       {
         const char *fullname = symtab_to_fullname (s);
+	char *rp = gdb_realpath (fullname);
 
-        if (fullname != NULL)
-          {
-            char *rp = gdb_realpath (fullname);
-
-            make_cleanup (xfree, rp);
-            if (FILENAME_CMP (real_path, rp) == 0)
-	      {
-		if (callback (s, data))
-		  return 1;
-	      }
+	make_cleanup (xfree, rp);
+	if (FILENAME_CMP (real_path, rp) == 0)
+	  {
+	    if (callback (s, data))
+	      return 1;
+	  }
 
-	    if (!is_abs && compare_filenames_for_search (rp, name))
-	      {
-		if (callback (s, data))
-		  return 1;
-	      }
-          }
+	if (!is_abs && compare_filenames_for_search (rp, name))
+	  {
+	    if (callback (s, data))
+	      return 1;
+	  }
       }
     }
 
@@ -2552,9 +2548,6 @@ find_line_symtab (struct symtab *symtab, int line,
 							 symtab->filename);
       }
 
-      /* Get symbol full file name if possible.  */
-      symtab_to_fullname (symtab);
-
       ALL_SYMTABS (objfile, s)
       {
 	struct linetable *l;
@@ -2562,9 +2555,7 @@ find_line_symtab (struct symtab *symtab, int line,
 
 	if (FILENAME_CMP (symtab->filename, s->filename) != 0)
 	  continue;
-	if (symtab->fullname != NULL
-	    && symtab_to_fullname (s) != NULL
-	    && FILENAME_CMP (symtab->fullname, s->fullname) != 0)
+	if (FILENAME_CMP (symtab->fullname, symtab_to_fullname (s)) != 0)
 	  continue;	
 	l = LINETABLE (s);
 	ind = find_line_common (l, line, &exact, 0);
@@ -3295,7 +3286,7 @@ sources_info (char *ignore, int from_tty)
   {
     const char *fullname = symtab_to_fullname (s);
 
-    output_source_filename (fullname ? fullname : s->filename, &data);
+    output_source_filename (fullname, &data);
   }
   printf_filtered ("\n\n");
 
diff --git a/gdb/testsuite/gdb.mi/mi-fullname-deleted.exp b/gdb/testsuite/gdb.mi/mi-fullname-deleted.exp
new file mode 100644
index 0000000..d76225e
--- /dev/null
+++ b/gdb/testsuite/gdb.mi/mi-fullname-deleted.exp
@@ -0,0 +1,41 @@
+# Copyright 2012 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+load_lib mi-support.exp
+set MIFLAGS "-i=mi"
+
+gdb_exit
+if [mi_gdb_start] {
+    continue
+}
+
+standard_testfile
+set srcfileabs [standard_output_file $srcfile]
+
+set f [open $srcfileabs "w"]
+puts $f "int main (void) { return 0; }"
+close $f
+
+if { [gdb_compile "$srcfileabs" "${binfile}" executable {debug}] != "" } {
+    untested $testname
+    return -1
+}
+
+file delete -- $srcfileabs
+
+mi_gdb_reinitialize_dir $srcdir/$subdir
+mi_gdb_load ${binfile}
+
+mi_gdb_test "-file-list-exec-source-file" ".*\",fullname=\"[string_to_regexp $srcfileabs]\".*" "fullname present"
diff --git a/gdb/tracepoint.c b/gdb/tracepoint.c
index f61ede7..647c31e 100644
--- a/gdb/tracepoint.c
+++ b/gdb/tracepoint.c
@@ -4873,8 +4873,7 @@ print_one_static_tracepoint_marker (int count,
 	{
 	  const char *fullname = symtab_to_fullname (sal.symtab);
 
-	  if (fullname)
-	    ui_out_field_string (uiout, "fullname", fullname);
+	  ui_out_field_string (uiout, "fullname", fullname);
 	}
       else
 	ui_out_field_skip (uiout, "fullname");

RE: [rfc] Print MI fullname even for non-existing files

[Marc Khouzam]

> -----Original Message-----
> From: gdb-patches-owner@sourceware.org 
> [mailto:gdb-patches-owner@sourceware.org] On Behalf Of Tom Tromey
> Sent: Monday, December 17, 2012 1:33 PM
> To: Jan Kratochvil
> Cc: gdb-patches@sourceware.org
> Subject: Re: [rfc] Print MI fullname even for non-existing files
> 
> >>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:
> 
> Jan> symtab_to_fullname now always returns non-NULL.
> 
> Jan> Currently this patch only changes from user point of view MI:
> Jan> (gdb) -file-list-exec-source-file
> Jan> 	^done,line="1",file="deleted.c",macro-info="0"
> -> 
> Jan> 	
> ^done,line="1",file="deleted.c",fullname="/path/to/deleted.c",
> macro-info="0"
> 
> Jan> Is there a serious reason why fullname was suppressed for
> Jan> non-existing files?  While it breaks MI compatibility in 
> some way I
> Jan> find it an acceptable change.
> 
> It seems pretty reasonable to me as well.

For what it's worth, that is an ok change for Eclipse (we don't
use that command).

Marc 

[commit] [rfc] Print MI fullname even for non-existing files

[Jan Kratochvil]

On Thu, 20 Dec 2012 15:32:55 +0100, Marc Khouzam wrote:
> > Jan> symtab_to_fullname now always returns non-NULL.
> > 
> > Jan> Currently this patch only changes from user point of view MI:
> > Jan> (gdb) -file-list-exec-source-file
> > Jan> 	^done,line="1",file="deleted.c",macro-info="0"
> > -> 
> > Jan> 	
> > ^done,line="1",file="deleted.c",fullname="/path/to/deleted.c",
> > macro-info="0"
> > 
> > Jan> Is there a serious reason why fullname was suppressed for
> > Jan> non-existing files?  While it breaks MI compatibility in 
> > some way I
> > Jan> find it an acceptable change.
> > 
> > It seems pretty reasonable to me as well.
> 
> For what it's worth, that is an ok change for Eclipse (we don't
> use that command).

This specific command was just an example.  The change affects any MI output
where fullname="/abs/path" was printed.  Now it will be printed even if that
source file is not present on the disk.  While that may be rather rate case it
would be bad if Eclipse breaks.

I have tested now eclipse-cdt-8.1.1-1.fc18.x86_64, if I remove /usr/src/debug
I get:

unpatched/patched present file:
Subwindow tab title shows the short filename "ioputs.c".

unpatched missing file:
Can't find a source file at "ioputs.c" 
Locate the file or edit the source lookup path to include its location.
Subwindow tab title shows the short filename "ioputs.c":
	http://people.redhat.com/~jkratoch/eclipse-gdb-unpatched.png

patched missing file:
Can't find a source file at "/usr/src/debug/glibc-2.16-75f0d304/libio/ioputs.c"
Locate the file or edit the source lookup path to include its location.
Subwindow tab title shows the fullname "/usr/src/debug/glibc-2.16-75f0d304/libio/ioputs.c":
	http://people.redhat.com/~jkratoch/eclipse-gdb-fullname.png

So I find the patch in fact an improvement even for Eclipse.
Therefore I have checked it in:
	http://sourceware.org/ml/gdb-cvs/2012-12/msg00184.html

Besides that the compatibility of Eclipse <-> GDB across versions is not too
great, FSF GDB HEAD fails with eclipse-cdt-8.0.1-4.fc16.x86_64:
	Error in final launch sequence
	Failed to execute MI command:
	maintenance set python print-stack off
	Error message from debugger back end:
	Undefined maintenance set command: "python print-stack off".  Try "help maintenance set".


Thanks,
Jan

Re: [rfc] Print MI fullname even for non-existing files

[Edjunior Barbosa Machado]

On 12/17/2012 04:59 PM, Jan Kratochvil wrote:
> gdb/
> 2012-12-17  Jan Kratochvil  <jan.kratochvil@redhat.com>
> 
> 	* ada-lang.c (is_known_support_routine): New variable fullname.  Use
> 	access call to verify the symtab_to_fullname result.
> 	* breakpoint.c (print_breakpoint_location, update_static_tracepoint):
> 	Remove NULL check of symtab_to_fullname result.
> 	* cli/cli-cmds.c (edit_command): Likewise.
> 	* mi/mi-cmd-file.c (mi_cmd_file_list_exec_source_file)
> 	(mi_cmd_file_list_exec_source_files): Likewise.
> 	* python/py-symtab.c (stpy_fullname): Likewise.
> 	* source.c (symtab_to_fullname): Update function comment.  Rename
> 	variable r to fd, move it to inner block.  Always provide non-NULL
> 	result.
> 	(print_source_lines_base): Remove NULL check of symtab_to_fullname
> 	result.
> 	* stack.c (print_frame): Likewise.
> 	* symtab.c (iterate_over_some_symtabs, find_line_symtab, sources_info):
> 	Likewise.
> 	* tracepoint.c (print_one_static_tracepoint_marker): Likewise.
> 
> gdb/doc/
> 2012-12-17  Jan Kratochvil  <jan.kratochvil@redhat.com>
> 
> 	* gdb.texinfo (GDB/MI Data Manipulation) (fullname): Make it always
> 	present.
> 	(GDB/MI File Commands) (-file-list-exec-source-files): Make the
> 	fullname output always present.
> 
> gdb/testsuite/
> 2012-12-17  Jan Kratochvil  <jan.kratochvil@redhat.com>
> 
> 	* gdb.mi/mi-fullname-deleted.exp: New file.
> 

I've seen a high number of failures on gdb.base/list.exp testcase on ppc64 after
this change and noticed that gdb was segfaulting when issuing a list with
specific arguments:

$ ./gdb ./testsuite/gdb.base/list -q -ex 'list list0.h:1'
Reading symbols from /home/emachado/gdb/build/gdb/testsuite/gdb.base/list...done.
Segmentation fault (core dumped)

This seem to be due to a strcmp with a null string 'symtab->fullname' in:

        if (FILENAME_CMP (symtab->fullname, symtab_to_fullname (s)) != 0)

With the patch below, this problem is fixed and gdb.base/list.exp no longer
crashes gdb. Ok?


2012-12-27  Edjunior Machado  <emachado@linux.vnet.ibm.com>

	* symtab.c (find_line_symtab): Update symtab->fullname before compare
	full file name strings.

diff --git a/gdb/symtab.c b/gdb/symtab.c
index a39e5bf..bd6c3b4 100644
--- a/gdb/symtab.c
+++ b/gdb/symtab.c
@@ -2547,6 +2547,9 @@ find_line_symtab (struct symtab *symtab, int line,
 							 symtab->filename);
       }
 
+      /* Get symbol full file name if possible.  */
+      symtab_to_fullname (symtab);
+
       ALL_SYMTABS (objfile, s)
       {
 	struct linetable *l;

[commit] [rfc] Print MI fullname even for non-existing files

[Jan Kratochvil]

On Fri, 28 Dec 2012 01:17:06 +0100, Edjunior Barbosa Machado wrote:
> This seem to be due to a strcmp with a null string 'symtab->fullname' in:
> 
>         if (FILENAME_CMP (symtab->fullname, symtab_to_fullname (s)) != 0)
> 
> With the patch below, this problem is fixed and gdb.base/list.exp no longer
> crashes gdb. Ok?

Thanks for the bugreport.  I have checked in an equivalent fix below.


Jan


http://sourceware.org/ml/gdb-cvs/2012-12/msg00192.html

--- src/gdb/ChangeLog	2012/12/25 17:27:47	1.14982
+++ src/gdb/ChangeLog	2012/12/28 09:05:39	1.14983
@@ -1,3 +1,8 @@
+2012-12-28  Jan Kratochvil  <jan.kratochvil@redhat.com>
+
+	* symtab.c (find_line_symtab): Call symtab_to_fullname instead of
+	accessing FULLNAME directly.
+
 2012-12-25  Sergio Durigan Junior  <sergiodj@redhat.com>
 
 	* frv-linux-tdep.c (frv_linux_pc_in_sigtramp): Fix typo.
--- src/gdb/symtab.c	2012/12/25 08:03:31	1.333
+++ src/gdb/symtab.c	2012/12/28 09:05:40	1.334
@@ -2554,7 +2554,8 @@
 
 	if (FILENAME_CMP (symtab->filename, s->filename) != 0)
 	  continue;
-	if (FILENAME_CMP (symtab->fullname, symtab_to_fullname (s)) != 0)
+	if (FILENAME_CMP (symtab_to_fullname (symtab),
+			  symtab_to_fullname (s)) != 0)
 	  continue;	
 	l = LINETABLE (s);
 	ind = find_line_common (l, line, &exact, 0);

RE: [commit] [rfc] Print MI fullname even for non-existing files

[Marc Khouzam]

> -----Original Message-----
> From: Jan Kratochvil [mailto:jan.kratochvil@redhat.com] 
> Sent: Tuesday, December 25, 2012 3:04 AM
> To: Marc Khouzam
> Cc: 'Tom Tromey'; 'gdb-patches@sourceware.org'
> Subject: [commit] [rfc] Print MI fullname even for non-existing files
> 
> On Thu, 20 Dec 2012 15:32:55 +0100, Marc Khouzam wrote:
> > > Jan> symtab_to_fullname now always returns non-NULL.
> > > 
> > > Jan> Currently this patch only changes from user point of view MI:
> > > Jan> (gdb) -file-list-exec-source-file
> > > Jan> 	^done,line="1",file="deleted.c",macro-info="0"
> > > -> 
> > > Jan> 	
> > > ^done,line="1",file="deleted.c",fullname="/path/to/deleted.c",
> > > macro-info="0"
> > > 
> > > Jan> Is there a serious reason why fullname was suppressed for
> > > Jan> non-existing files?  While it breaks MI compatibility in 
> > > some way I
> > > Jan> find it an acceptable change.
> > > 
> > > It seems pretty reasonable to me as well.
> > 
> > For what it's worth, that is an ok change for Eclipse (we don't
> > use that command).
> 
> This specific command was just an example.  The change 
> affects any MI output
> where fullname="/abs/path" was printed.  Now it will be 
> printed even if that
> source file is not present on the disk.  While that may be 
> rather rate case it
> would be bad if Eclipse breaks.
> 
> I have tested now eclipse-cdt-8.1.1-1.fc18.x86_64, if I 

Thanks for doing this!

> remove /usr/src/debug
> I get:
> 
> unpatched/patched present file:
> Subwindow tab title shows the short filename "ioputs.c".
> 
> unpatched missing file:
> Can't find a source file at "ioputs.c" 
> Locate the file or edit the source lookup path to include its 
> location.
> Subwindow tab title shows the short filename "ioputs.c":
> 	http://people.redhat.com/~jkratoch/eclipse-gdb-unpatched.png
> 
> patched missing file:
> Can't find a source file at 
> "/usr/src/debug/glibc-2.16-75f0d304/libio/ioputs.c"
> Locate the file or edit the source lookup path to include its 
> location.
> Subwindow tab title shows the fullname 
> "/usr/src/debug/glibc-2.16-75f0d304/libio/ioputs.c":
> 	http://people.redhat.com/~jkratoch/eclipse-gdb-fullname.png
> 
> So I find the patch in fact an improvement even for Eclipse.

Agreed.

> Therefore I have checked it in:
> 	http://sourceware.org/ml/gdb-cvs/2012-12/msg00184.html
> 
> Besides that the compatibility of Eclipse <-> GDB across 
> versions is not too
> great, FSF GDB HEAD fails with eclipse-cdt-8.0.1-4.fc16.x86_64:
> 	Error in final launch sequence
> 	Failed to execute MI command:
> 	maintenance set python print-stack off
> 	Error message from debugger back end:
> 	Undefined maintenance set command: "python print-stack 
> off".  Try "help maintenance set".

This has been fixed in cdt-8.0.2:
https://bugs.eclipse.org/bugs/show_bug.cgi?id=367788

Thanks again

Marc

Re: [commit] [rfc] Print MI fullname even for non-existing files

[Jan Kratochvil]

On Thu, 03 Jan 2013 16:38:04 +0100, Marc Khouzam wrote:
> > From: Jan Kratochvil [mailto:jan.kratochvil@redhat.com] 
> > Besides that the compatibility of Eclipse <-> GDB across 
> > versions is not too
> > great, FSF GDB HEAD fails with eclipse-cdt-8.0.1-4.fc16.x86_64:
> > 	Error in final launch sequence
> > 	Failed to execute MI command:
> > 	maintenance set python print-stack off
> > 	Error message from debugger back end:
> > 	Undefined maintenance set command: "python print-stack 
> > off".  Try "help maintenance set".
> 
> This has been fixed in cdt-8.0.2:
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=367788

This does not matter, the point was that GDB does not fully keep Front Ends
backward compatibility.

I remember some discussion around the deprecation so it was intentional.


Jan

Re: [commit] [rfc] Print MI fullname even for non-existing files

[Tom Tromey]

Jan> This does not matter, the point was that GDB does not fully keep Front Ends
Jan> backward compatibility.

I think we should never promise compatibility for "maint" commands.

Tom

Re: [commit] [rfc] Print MI fullname even for non-existing files

[Joel Brobecker]

> I think we should never promise compatibility for "maint" commands.

Agreed!

-- 
Joel

Re: [rfc] Print MI fullname even for non-existing files

[Pedro Alves]

Hello,

On 12/17/2012 03:58 PM, Jan Kratochvil wrote:

> gdb/doc/
> 2012-12-17  Jan Kratochvil  <jan.kratochvil@redhat.com>
> 
> 	* gdb.texinfo (GDB/MI Data Manipulation) (fullname): Make it always
> 	present.
> 	(GDB/MI File Commands) (-file-list-exec-source-files): Make the
> 	fullname output always present.

Going through NEWS in 7.6, I noticed this doesn't have an entry under
"MI changes" (there maybe be other fullname related changes missing too;
haven't checked those).

-- 
Pedro Alves