This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: one-too-many location in breakpoint
> Tom> I fixed all the other problems. I'm running a regression test to make
> Tom> sure I didn't make something else break while doing this. If all goes
> Tom> well I will send out refreshed patches.
As discussed on IRC, there was one problem that I missed in the
proposals I made earlier. Attached is a patch that I wrote
which should solve that problem. In hindsight, perhaps I should
have split it in 2 or 3 patches, for instance one patch for
the new language method, and one patch for the rest. And maybe
one patch on its own for the documentation update of
demangle_for_lookup. Anyways, that's fairly easy to do, but
for now, I think it's more important to get the patch out.
Note that this patch adds new tests.
With that patch, pretty much everything that I suggested, except
for the couple of issues I briefly mentioned (without describing
them), has been addressed. I'm not convinced yet that the remaining
issues have anything to do with the patch series but I will look
into that tomorrow.
For the record, it was testing on x86_64-linux, with the official
testsuite as well as AdaCore's testsuite.
--
Joel
>From 908c8de2fe6bf9be3a2bee1ebc9858ec292ea6a2 Mon Sep 17 00:00:00 2001
From: Joel Brobecker <brobecker@adacore.com>
Date: Tue, 29 Nov 2011 10:35:27 -0800
Subject: [PATCH] new la_symbol_name_compare language method
The goal of this patch is to make sure that GDB is able to handle
Ada linespecs that's just a fully qualified function name with
some of the letters being uppercase characters. For instance:
(gdb) break Pck.Hello
The real issue is that the "natural" name "Pck.Hello" needs to be
encoded for the symbol lookup to be successful. That encoding
was currently only happening inside ada_iterate_over_symbols, which
means that it would only work if we specified a filename with
the function name. For instance:
(gdb) break pck.adb:Pck.Hello
But even there it is a little bit wasteful, because the
la_iterate_over_symbols language method is called for every
symbol table that seems to contain a matching symbol, sometimes
twice. It would be better to perform the name enconding once
at the beginning, and then not have to worry about it during
the symbol lookup.
The other issue is that symbol matching is currently hardcoded to
using strcmp_iw, which does not work for Ada.
So, this patch does the following:
1. It introduces a new language method called la_symbol_name_compare,
which is responsible for comparing two symbol names for a given
language.
This is something we want now because we want to be able to
use a different symbol matching routine for Ada, rather than
the default strcmp_iw.
But this is also something that would be useful to allow
language-specific sorting of partial symbols. Right now,
they are sorted using strcmp_iw_ordered, and the Ada-specific
comparison routine has to be "compatible" with that.
2. It exploits that new language method in the symbol matching
routine. For that, we added a language parameter to the profile
of the name_matcher function (which is passed as an argument to
the expand_symtabs_matching quick_symbol_functions routine).
3. It changes decode_variable so that, for Ada linespecs, function
name get encoded before the lookup starts. This allows us to
avoid doing it everytime ada_iterate_over_symbols is called.
gdb/ChangeLog:
* language.h (struct language_defn): Add new component
la_symbol_name_compare.
* symfile.h (struct quick_symbol_functions): Update the profile
of parameter "name_matcher" for the expand_symtabs_matching
method. Update the documentation accordingly.
* ada-lang.h (ada_name_for_lookup): Add declaration.
* ada-lang.c (ada_name_for_lookup): New function, extracted out
from ada_iterate_over_symbols.
(ada_iterate_over_symbols): Do not encode symbol name anymore.
(ada_expand_partial_symbol_name): Adjust profile.
(ada_language_defn): Add value for la_symbol_name_compare field.
* linespec.c: #include "ada-lang.h".
(iterate_name_matcher): Add language parameter. Replace call
to strcmp_iw by call to language->la_symbol_name_compare.
(decode_variable): Encode COPY if current language is Ada.
* dwarf2read.c (dw2_expand_symtabs_matching): Adjust profile
of name_matcher parameter. Adjust call to name_matcher.
* psymtab.c (expand_symtabs_matching_via_partial): Likewise.
(expand_partial_symbol_names): Update profile of parameter "fun".
* psymtab.h (expand_partial_symbol_names): Update profile of
parameter "fun".
* symtab.c (demangle_for_lookup): Update function documentation.
(search_symbols_name_matches): Add language parameter.
(expand_partial_symbol_name): Likewise.
* c-lang.c (c_language_defn, cplus_language_defn)
(asm_language_defn, minimal_language_defn): Add value for
la_symbol_name_compare field.
* d-lang.c (d_language_defn): Likewise.
* f-lang.c (f_language_defn): Ditto.
* jv-lang.c (java_language_defn): Ditto.
* m2-lang.c (m2_language_defn): Ditto.
* objc-lang.c (objc_language_defn): Ditto.
* opencl-lang.c (opencl_language_defn): Ditto.
* p-lang.c (pascal_language_defn): Ditto.
* language.c (unknown_language_defn, auto_language_defn)
(local_language_defn): Ditto.
gdb/testsuite/ChangeLog:
* gdb.ada/fullname_bp.exp: Add tests for other valid linespecs
involving a fully qualified function name.
---
gdb/ada-lang.c | 40 +++++++++++++++++++++++----------
gdb/ada-lang.h | 2 +
gdb/c-lang.c | 4 +++
gdb/d-lang.c | 1 +
gdb/dwarf2read.c | 13 +++++-----
gdb/f-lang.c | 1 +
gdb/jv-lang.c | 1 +
gdb/language.c | 3 ++
gdb/language.h | 14 +++++++++++
gdb/linespec.c | 13 +++++++++-
gdb/m2-lang.c | 1 +
gdb/objc-lang.c | 1 +
gdb/opencl-lang.c | 1 +
gdb/p-lang.c | 1 +
gdb/psymtab.c | 20 +++++++++-------
gdb/psymtab.h | 3 +-
gdb/symfile.h | 20 ++++++++++------
gdb/symtab.c | 22 +++++++++++++++--
gdb/testsuite/gdb.ada/fullname_bp.exp | 23 +++++++++++++++++++
19 files changed, 143 insertions(+), 41 deletions(-)
diff --git a/gdb/ada-lang.c b/gdb/ada-lang.c
index 6d69199..db9de01 100644
--- a/gdb/ada-lang.c
+++ b/gdb/ada-lang.c
@@ -5052,29 +5052,43 @@ done:
return ndefns;
}
-/* Implementation of the la_iterate_over_symbols method. */
+/* If NAME is the name of an entity, return a string that should
+ be used to look that entity up in Ada units. This string should
+ be deallocated after use using xfree.
-static void
-ada_iterate_over_symbols (const struct block *block,
- const char *name, domain_enum domain,
- int (*callback) (struct symbol *, void *),
- void *data)
+ NAME can have any form that the "break" or "print" commands might
+ recognize. In other words, it does not have to be the "natural"
+ name, or the "encoded" name. */
+
+char *
+ada_name_for_lookup (const char *name)
{
- int ndefs, i;
- struct ada_symbol_info *results;
char *canon;
int nlen = strlen (name);
if (name[0] == '<' && name[nlen - 1] == '>')
{
- canon = alloca (nlen - 1);
+ canon = xmalloc (nlen - 1);
memcpy (canon, name + 1, nlen - 2);
canon[nlen - 2] = '\0';
}
else
- canon = ada_encode (ada_fold_name (name));
+ canon = xstrdup (ada_encode (ada_fold_name (name)));
+ return canon;
+}
+
+/* Implementation of the la_iterate_over_symbols method. */
+
+static void
+ada_iterate_over_symbols (const struct block *block,
+ const char *name, domain_enum domain,
+ int (*callback) (struct symbol *, void *),
+ void *data)
+{
+ int ndefs, i;
+ struct ada_symbol_info *results;
- ndefs = ada_lookup_symbol_list (canon, block, domain, &results);
+ ndefs = ada_lookup_symbol_list (name, block, domain, &results);
for (i = 0; i < ndefs; ++i)
{
if (! (*callback) (results[i].sym, data))
@@ -5663,7 +5677,8 @@ struct add_partial_datum
/* A callback for expand_partial_symbol_names. */
static int
-ada_expand_partial_symbol_name (const char *name, void *user_data)
+ada_expand_partial_symbol_name (const struct language_defn *language,
+ const char *name, void *user_data)
{
struct add_partial_datum *data = user_data;
@@ -12368,6 +12383,7 @@ const struct language_defn ada_language_defn = {
ada_print_array_index,
default_pass_by_reference,
c_get_string,
+ compare_names,
ada_iterate_over_symbols,
LANG_MAGIC
};
diff --git a/gdb/ada-lang.h b/gdb/ada-lang.h
index d6404ca..fa4bb51 100644
--- a/gdb/ada-lang.h
+++ b/gdb/ada-lang.h
@@ -365,6 +365,8 @@ extern char *ada_breakpoint_rewrite (char *, int *);
extern char *ada_main_name (void);
+extern char *ada_name_for_lookup (const char *name);
+
/* Tasking-related: ada-tasks.c */
extern int valid_task_id (int);
diff --git a/gdb/c-lang.c b/gdb/c-lang.c
index 81e7849..7be916c 100644
--- a/gdb/c-lang.c
+++ b/gdb/c-lang.c
@@ -863,6 +863,7 @@ const struct language_defn c_language_defn =
default_print_array_index,
default_pass_by_reference,
c_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
@@ -985,6 +986,7 @@ const struct language_defn cplus_language_defn =
default_print_array_index,
cp_pass_by_reference,
c_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
@@ -1025,6 +1027,7 @@ const struct language_defn asm_language_defn =
default_print_array_index,
default_pass_by_reference,
c_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
@@ -1070,6 +1073,7 @@ const struct language_defn minimal_language_defn =
default_print_array_index,
default_pass_by_reference,
c_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
diff --git a/gdb/d-lang.c b/gdb/d-lang.c
index 44ff68a..fb6bacd 100644
--- a/gdb/d-lang.c
+++ b/gdb/d-lang.c
@@ -273,6 +273,7 @@ static const struct language_defn d_language_defn =
default_print_array_index,
default_pass_by_reference,
c_get_string,
+ strcmp_iw_ordered,
NULL,
LANG_MAGIC
};
diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index ba56065..81c4af1 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -2756,11 +2756,12 @@ dw2_map_matching_symbols (const char * name, domain_enum namespace,
}
static void
-dw2_expand_symtabs_matching (struct objfile *objfile,
- int (*file_matcher) (const char *, void *),
- int (*name_matcher) (const char *, void *),
- enum search_domain kind,
- void *data)
+dw2_expand_symtabs_matching
+ (struct objfile *objfile,
+ int (*file_matcher) (const char *, void *),
+ int (*name_matcher) (const struct language_defn *, const char *, void *),
+ enum search_domain kind,
+ void *data)
{
int i;
offset_type iter;
@@ -2812,7 +2813,7 @@ dw2_expand_symtabs_matching (struct objfile *objfile,
name = index->constant_pool + MAYBE_SWAP (index->symbol_table[idx]);
- if (! (*name_matcher) (name, data))
+ if (! (*name_matcher) (current_language, name, data))
continue;
/* The name was matched, now expand corresponding CUs that were
diff --git a/gdb/f-lang.c b/gdb/f-lang.c
index 7a8516d..00926fb 100644
--- a/gdb/f-lang.c
+++ b/gdb/f-lang.c
@@ -309,6 +309,7 @@ const struct language_defn f_language_defn =
default_print_array_index,
default_pass_by_reference,
default_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
diff --git a/gdb/jv-lang.c b/gdb/jv-lang.c
index ce65a2d..80fe4a1 100644
--- a/gdb/jv-lang.c
+++ b/gdb/jv-lang.c
@@ -1197,6 +1197,7 @@ const struct language_defn java_language_defn =
default_print_array_index,
default_pass_by_reference,
default_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
diff --git a/gdb/language.c b/gdb/language.c
index 701c479..1bbbfba 100644
--- a/gdb/language.c
+++ b/gdb/language.c
@@ -1200,6 +1200,7 @@ const struct language_defn unknown_language_defn =
default_print_array_index,
default_pass_by_reference,
default_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
@@ -1242,6 +1243,7 @@ const struct language_defn auto_language_defn =
default_print_array_index,
default_pass_by_reference,
default_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
@@ -1282,6 +1284,7 @@ const struct language_defn local_language_defn =
default_print_array_index,
default_pass_by_reference,
default_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
diff --git a/gdb/language.h b/gdb/language.h
index 82e6a88..65d55db 100644
--- a/gdb/language.h
+++ b/gdb/language.h
@@ -318,6 +318,20 @@ struct language_defn
void (*la_get_string) (struct value *value, gdb_byte **buffer, int *length,
struct type **chartype, const char **charset);
+ /* Compare two symbol names according to language rules. For
+ instance, in C++, we might want to ignore whitespaces in
+ the symbol name. Or some case-insensitive language might
+ want to ignore casing during the match.
+
+ Both STR1 and STR2 are expected to be demangled name, except
+ for Ada, where STR1 and STR2 are expected to be encoded names.
+ The latter is because searches are performed using the encoded
+ name in Ada.
+
+ The return value follows the same spirit as strcmp. */
+
+ int (*la_symbol_name_compare) (const char *str1, const char *str2);
+
/* Find all symbols in the current program space matching NAME in
DOMAIN, according to this language's rules.
diff --git a/gdb/linespec.c b/gdb/linespec.c
index 445dbf1..adbefff 100644
--- a/gdb/linespec.c
+++ b/gdb/linespec.c
@@ -44,6 +44,7 @@
#include <ctype.h>
#include "cli/cli-utils.h"
#include "filenames.h"
+#include "ada-lang.h"
typedef struct symtab *symtab_p;
DEF_VEC_P (symtab_p);
@@ -325,11 +326,12 @@ cplusplus_error (const char *name, const char *fmt, ...)
callback to the expand_symtabs_matching method. */
static int
-iterate_name_matcher (const char *name, void *d)
+iterate_name_matcher (const struct language_defn *language,
+ const char *name, void *d)
{
const char **dname = d;
- if (strcmp_iw (name, *dname) == 0)
+ if (language->la_symbol_name_compare (name, *dname) == 0)
return 1;
return 0;
}
@@ -2827,6 +2829,13 @@ decode_variable (struct linespec_state *self, char *copy)
cleanup = demangle_for_lookup (copy, current_language->la_language,
&lookup_name);
+ if (current_language->la_language == language_ada)
+ {
+ /* In Ada, the symbol lookups are performed using the encoded
+ name rather than the demangled name. */
+ lookup_name = ada_name_for_lookup (copy);
+ make_cleanup (xfree, (void *) lookup_name);
+ }
canon = cp_canonicalize_string_no_typedefs (copy);
if (canon != NULL)
diff --git a/gdb/m2-lang.c b/gdb/m2-lang.c
index 106dfcf..c850b1f 100644
--- a/gdb/m2-lang.c
+++ b/gdb/m2-lang.c
@@ -401,6 +401,7 @@ const struct language_defn m2_language_defn =
default_print_array_index,
default_pass_by_reference,
default_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
diff --git a/gdb/objc-lang.c b/gdb/objc-lang.c
index 1da92a5..cb7fa0e 100644
--- a/gdb/objc-lang.c
+++ b/gdb/objc-lang.c
@@ -541,6 +541,7 @@ const struct language_defn objc_language_defn = {
default_print_array_index,
default_pass_by_reference,
default_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
diff --git a/gdb/opencl-lang.c b/gdb/opencl-lang.c
index 895a51c..1deb9d8 100644
--- a/gdb/opencl-lang.c
+++ b/gdb/opencl-lang.c
@@ -1024,6 +1024,7 @@ const struct language_defn opencl_language_defn =
default_print_array_index,
default_pass_by_reference,
c_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
diff --git a/gdb/p-lang.c b/gdb/p-lang.c
index 30b890f..f66d471 100644
--- a/gdb/p-lang.c
+++ b/gdb/p-lang.c
@@ -459,6 +459,7 @@ const struct language_defn pascal_language_defn =
default_print_array_index,
default_pass_by_reference,
default_get_string,
+ strcmp_iw_ordered,
iterate_over_symbols,
LANG_MAGIC
};
diff --git a/gdb/psymtab.c b/gdb/psymtab.c
index c20fd78..946a511 100644
--- a/gdb/psymtab.c
+++ b/gdb/psymtab.c
@@ -1244,13 +1244,12 @@ map_matching_symbols_psymtab (const char *name, domain_enum namespace,
}
static void
-expand_symtabs_matching_via_partial (struct objfile *objfile,
- int (*file_matcher) (const char *,
- void *),
- int (*name_matcher) (const char *,
- void *),
- enum search_domain kind,
- void *data)
+expand_symtabs_matching_via_partial
+ (struct objfile *objfile,
+ int (*file_matcher) (const char *, void *),
+ int (*name_matcher) (const struct language_defn *, const char *, void *),
+ enum search_domain kind,
+ void *data)
{
struct partial_symtab *ps;
@@ -1300,7 +1299,8 @@ expand_symtabs_matching_via_partial (struct objfile *objfile,
&& SYMBOL_CLASS (*psym) == LOC_BLOCK)
|| (kind == TYPES_DOMAIN
&& SYMBOL_CLASS (*psym) == LOC_TYPEDEF))
- && (*name_matcher) (SYMBOL_NATURAL_NAME (*psym), data))
+ && (*name_matcher) (current_language,
+ SYMBOL_NATURAL_NAME (*psym), data))
{
PSYMTAB_TO_SYMTAB (ps);
keep_going = 0;
@@ -1937,7 +1937,9 @@ maintenance_check_symtabs (char *ignore, int from_tty)
void
-expand_partial_symbol_names (int (*fun) (const char *, void *), void *data)
+expand_partial_symbol_names (int (*fun) (const struct language_defn *,
+ const char *, void *),
+ void *data)
{
struct objfile *objfile;
diff --git a/gdb/psymtab.h b/gdb/psymtab.h
index de292c5..940b537 100644
--- a/gdb/psymtab.h
+++ b/gdb/psymtab.h
@@ -30,7 +30,8 @@ extern struct psymbol_bcache *psymbol_bcache_init (void);
extern void psymbol_bcache_free (struct psymbol_bcache *);
extern struct bcache *psymbol_bcache_get_bcache (struct psymbol_bcache *);
-void expand_partial_symbol_names (int (*fun) (const char *, void *),
+void expand_partial_symbol_names (int (*fun) (const struct language_defn *,
+ const char *, void *),
void *data);
void map_partial_symbol_filenames (symbol_filename_ftype *fun, void *data,
diff --git a/gdb/symfile.h b/gdb/symfile.h
index 8b3b705..44f0c01 100644
--- a/gdb/symfile.h
+++ b/gdb/symfile.h
@@ -261,9 +261,12 @@ struct quick_symbol_functions
Otherwise, if KIND does not match this symbol is skipped.
- If even KIND matches, then NAME_MATCHER is called for each symbol defined
- in the file. The symbol's "natural" name and DATA are passed to
- NAME_MATCHER.
+ If even KIND matches, then NAME_MATCHER is called for each symbol
+ defined in the file. The current language, the symbol name and
+ DATA are passed to NAME_MATCHER. The symbol "natural" name should
+ be passed to NAME_MATCHER for all languages except Ada, where
+ the encoded name is passed instead (see la_symbol_name_compare in
+ struct language_defn for more details on this).
If NAME_MATCHER returns zero, then this symbol is skipped.
@@ -271,11 +274,12 @@ struct quick_symbol_functions
DATA is user data that is passed unmodified to the callback
functions. */
- void (*expand_symtabs_matching) (struct objfile *objfile,
- int (*file_matcher) (const char *, void *),
- int (*name_matcher) (const char *, void *),
- enum search_domain kind,
- void *data);
+ void (*expand_symtabs_matching)
+ (struct objfile *objfile,
+ int (*file_matcher) (const char *, void *),
+ int (*name_matcher) (const struct language_defn *, const char *, void *),
+ enum search_domain kind,
+ void *data);
/* Return the symbol table from OBJFILE that contains PC and
SECTION. Return NULL if there is no such symbol table. This
diff --git a/gdb/symtab.c b/gdb/symtab.c
index 3ae5cf9..1ea4253 100644
--- a/gdb/symtab.c
+++ b/gdb/symtab.c
@@ -1054,7 +1054,21 @@ fixup_symbol_section (struct symbol *sym, struct objfile *objfile)
/* Compute the demangled form of NAME as used by the various symbol
lookup functions. The result is stored in *RESULT_NAME. Returns a
- cleanup which can be used to clean up the result. */
+ cleanup which can be used to clean up the result.
+
+ For Ada, this function just sets *RESULT_NAME to NAME, unmodified.
+ Normally, Ada symbol lookups are performed using the encoded name
+ rather than the demangled name, and so it might seem to make sense
+ for this function to return an encoded version of NAME.
+ Unfortunately, we cannot do this, because this function is used in
+ circumstances where it is not appropriate to try to encode NAME.
+ For instance, when displaying the frame info, we demangle the name
+ of each parameter, and then perform a symbol lookup inside our
+ function using that demangled name. In Ada, certain functions
+ have internally-generated parameters whose name contain uppercase
+ characters. Encoding those name would result in those uppercase
+ characters to become lowercase, and thus cause the symbol lookup
+ to fail. */
struct cleanup *
demangle_for_lookup (const char *name, enum language lang,
@@ -3152,7 +3166,8 @@ search_symbols_file_matches (const char *filename, void *user_data)
/* A callback for expand_symtabs_matching. */
static int
-search_symbols_name_matches (const char *symname, void *user_data)
+search_symbols_name_matches (const struct language_defn *language,
+ const char *symname, void *user_data)
{
struct search_symbols_data *data = user_data;
@@ -3960,7 +3975,8 @@ add_macro_name (const char *name, const struct macro_definition *ignore,
/* A callback for expand_partial_symbol_names. */
static int
-expand_partial_symbol_name (const char *name, void *user_data)
+expand_partial_symbol_name (const struct language_defn *language,
+ const char *name, void *user_data)
{
struct add_name_data *datum = (struct add_name_data *) user_data;
diff --git a/gdb/testsuite/gdb.ada/fullname_bp.exp b/gdb/testsuite/gdb.ada/fullname_bp.exp
index faf63b9..0e4d32f 100644
--- a/gdb/testsuite/gdb.ada/fullname_bp.exp
+++ b/gdb/testsuite/gdb.ada/fullname_bp.exp
@@ -25,6 +25,9 @@ if {[gdb_compile_ada "${srcfile}" "${binfile}" executable [list debug ]] != "" }
return -1
}
+# Note: We restart the debugger before setting each breakpoint, because
+# we want to test the situation where the symtab for our breakpoint
+# has not been created yet.
clean_restart ${testfile}
# Break on "pck.hello" rather than just "hello" to make sure we trigger
@@ -32,3 +35,23 @@ clean_restart ${testfile}
gdb_test "break pck.hello" \
"Breakpoint \[0-9\]+ at 0x\[0-9a-f\]+: file .*pck.adb, line \[0-9\]+."
+# Do the same, but this time using a linespec where the user also
+# provided a filename.
+
+clean_restart ${testfile}
+
+gdb_test "break pck.adb:pck.hello" \
+ "Breakpoint \[0-9\]+ at 0x\[0-9a-f\]+: file .*pck.adb, line \[0-9\]+."
+
+# Same scenarios as above, but with a function name that is spelled
+# with upper-case letters.
+
+clean_restart ${testfile}
+
+gdb_test "break Pck.Hello" \
+ "Breakpoint \[0-9\]+ at 0x\[0-9a-f\]+: file .*pck.adb, line \[0-9\]+."
+
+clean_restart ${testfile}
+
+gdb_test "break pck.adb:Pck.Hello" \
+ "Breakpoint \[0-9\]+ at 0x\[0-9a-f\]+: file .*pck.adb, line \[0-9\]+."
--
1.7.1