This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [python] [patch] PR python/12656 (API for special blocks)
- From: Tom Tromey <tromey at redhat dot com>
- To: pmuldoon at redhat dot com
- Cc: gdb-patches at sourceware dot org
- Date: Fri, 07 Oct 2011 14:41:24 -0600
- Subject: Re: [python] [patch] PR python/12656 (API for special blocks)
- References: <m339f4envs.fsf@redhat.com>
>>>>> "Phil" == Phil Muldoon <pmuldoon@redhat.com> writes:
Phil> This patch addresses PR python/12656. It allows the user to determine if
Phil> a given gdb.Block is the static block, or the global block.
Phil> Additionally I added two new APIs to actually get the global/static
Phil> blocks on behalf of the user.
Thanks.
Phil> I also, because the global/static GDB fetch methods use const struct
Phil> *block, changed the Python code to use consts. This required some minor
Phil> refactoring both inside and outside the Python API code.
Nice. Looks pretty good.
Phil> +@defun Block.is_global ()
Phil> +@defun Block.is_static ()
Why are these methods rather than attributes?
Phil> diff --git a/gdb/python/py-block.c b/gdb/python/py-block.c
Phil> index 08d4455..5ef7d2e 100644
Phil> --- a/gdb/python/py-block.c
Phil> +++ b/gdb/python/py-block.c
Phil> @@ -28,7 +28,7 @@
Phil> typedef struct blpy_block_object {
Phil> PyObject_HEAD
Phil> /* The GDB block structure that represents a frame's code block. */
Phil> - struct block *block;
Phil> + const struct block *block;
Phil> /* The backing object file. There is no direct relationship in GDB
Phil> between a block and an object file. When a block is created also
Phil> store a pointer to the object file for later use. */
Phil> @@ -50,7 +50,7 @@ typedef struct {
Phil> /* Pointer back to the original source block object. Needed to
Phil> check if the block is still valid, and has not been invalidated
Phil> when an object file has been freed. */
Phil> - struct blpy_block_object *source;
Phil> + const struct blpy_block_object *source;
Why make this one const?
It seems wrong given that we have to decref it when the iterator is
destroyed.
Phil> + Py_INCREF (Py_None);
Phil> + return Py_None;
I think it is more normal to use Py_RETURN_NONE here.
Phil> gdb_test "python print block_iter.is_valid()" "False" \
Phil> "Check block validity"
Phil> +
Gratuitous newline addition.
Phil> -extern struct value *value_of_variable (struct symbol *var, struct block *b);
Phil> +extern struct value *value_of_variable (struct symbol *var, const struct block *b);
This should probably wrap after the ",".
Tom