Re: [PATCH v2 03/28] Emit some type declarations in .gdb_index

[Tom Tromey <tom@tromey.com>]

>>>>> "Simon" == Simon Marchi <simark@simark.ca> writes:

Simon> On 2025-04-02 19:45, Tom Tromey wrote:
>> If you run struct-decl.exp with the .gdb_index board, you will see
>> that "the_type" is not emitted in the index.  This would cause a
>> failure in this series.  The fix is to ensure that certain necessary
>> type declarations are emitted.
>> 
>> However, a naive fix here will regress stub-array-size.exp, where a
>> type declaration and a type definition are both seen -- but the
>> declaration is seen first and causes a failure.  This is handled by
>> adding some code (including a mild hack) to filter out type
>> declarations when a corresponding type definition is seen.

Simon> I don't really understand the use of GDB_INDEX_SYMBOL_KIND_UNUSED5, what
Simon> does it mean?

I thought it was reasonably well explained in the comments.
Could you say which part needs more text to be clear?


+  /* We sort the indexes in a funny way: GDB_INDEX_SYMBOL_KIND_UNUSED5
+     is always sorted last; then otherwise we sort by numeric value.
+     This ensures that we prefer the definition when both a definition
+     and a declaration (stub type) are seen.  */
...
+  /* Rewrite GDB_INDEX_SYMBOL_KIND_UNUSED5.  This ensures that a type
+     declaration will be deleted by the subsequent squashing step, if
+     warranted.  */
...
+	  /* If we added a type declaration, we want to note this
+	     fact for later, because we don't want a type declaration
+	     to cause the real definition to be omitted from the
+	     index.  GDB_INDEX_SYMBOL_KIND_UNUSED5 is used here, but
+	     rewritten later before the index is written.  */


Tom