[PATCHv3,1/2] gdb: have mdict_size always return a symbol count

Message ID ca11ca3d286b77ef56ed6fc935e05aeee33e3de1.1684529933.git.aburgess@redhat.com
State New
Headers
Series Add __repr__() implementation to a few Python types |

Commit Message

Andrew Burgess May 19, 2023, 9:27 p.m. UTC
  In the next commit we would like to have mdict_size return the number
of symbols in the dictionary, currently mdict_size is just a
heuristic, sometimes it returns the number of symbols, and sometimes
the number of buckets in a hashing dictionary (see size_hashed in
dictionary.c).

Currently this vague notion of size is good enough, the only place
mdict_size is used is in a maintenance command in order to print a
message containing the size of the dictionary ... so we don't really
care that the value isn't correct.

However, in the next commit we do want the size returned to be the
number of symbols in the dictionary, so this commit makes mdict_size
return the symbol count in all cases.

The new use is still not on a hot path -- it's going to be a Python
__repr__ method, so all I do in this commit is have size_hashed walk
the dictionary and count the entries, obviously this could be slow if
we have a large number of symbols, but for now I'm not worrying about
that case.  We could always store the symbol count if we wanted, but
that would increase the size of every dictionary for a use case that
isn't going to be hit that often.

I've updated the text in 'maint print symbols' so that we don't talk
about the size being 'syms/buckets', but just 'symbols' now.
---
 gdb/dictionary.c | 13 ++++++++++++-
 gdb/dictionary.h |  3 +--
 gdb/symmisc.c    |  2 +-
 3 files changed, 14 insertions(+), 4 deletions(-)
  

Patch

diff --git a/gdb/dictionary.c b/gdb/dictionary.c
index 4f8df240a3e..f3dfe092830 100644
--- a/gdb/dictionary.c
+++ b/gdb/dictionary.c
@@ -650,7 +650,18 @@  insert_symbol_hashed (struct dictionary *dict,
 static int
 size_hashed (const struct dictionary *dict)
 {
-  return DICT_HASHED_NBUCKETS (dict);
+  int nbuckets = DICT_HASHED_NBUCKETS (dict);
+  int total = 0;
+
+  for (int i = 0; i < nbuckets; ++i)
+    {
+      for (struct symbol *sym = DICT_HASHED_BUCKET (dict, i);
+	   sym != nullptr;
+	   sym = sym->hash_next)
+	total++;
+    }
+
+  return total;
 }
 
 /* Functions only for DICT_HASHED_EXPANDABLE.  */
diff --git a/gdb/dictionary.h b/gdb/dictionary.h
index d982396cb31..44bd075fcf2 100644
--- a/gdb/dictionary.h
+++ b/gdb/dictionary.h
@@ -159,8 +159,7 @@  extern struct symbol *
 extern struct symbol *mdict_iter_match_next (const lookup_name_info &name,
 					     struct mdict_iterator *miterator);
 
-/* Return some notion of the size of the multidictionary: the number of
-   symbols if we have that, the number of hash buckets otherwise.  */
+/* Return the number of symbols in multidictionary MDICT.  */
 
 extern int mdict_size (const struct multidictionary *mdict);
 
diff --git a/gdb/symmisc.c b/gdb/symmisc.c
index ff7f31f885f..e414dea7bf6 100644
--- a/gdb/symmisc.c
+++ b/gdb/symmisc.c
@@ -286,7 +286,7 @@  dump_symtab_1 (struct symtab *symtab, struct ui_file *outfile)
 	  /* drow/2002-07-10: We could save the total symbols count
 	     even if we're using a hashtable, but nothing else but this message
 	     wants it.  */
-	  gdb_printf (outfile, ", %d syms/buckets in ",
+	  gdb_printf (outfile, ", %d symbols in ",
 		      mdict_size (b->multidict ()));
 	  gdb_puts (paddress (gdbarch, b->start ()), outfile);
 	  gdb_printf (outfile, "..");