From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 19346 invoked by alias); 8 May 2002 02:19:08 -0000 Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sources.redhat.com Received: (qmail 19338 invoked from network); 8 May 2002 02:19:07 -0000 Received: from unknown (HELO cygnus.com) (205.180.83.203) by sources.redhat.com with SMTP; 8 May 2002 02:19:07 -0000 Received: from redhat.com (reddwarf.sfbay.redhat.com [172.16.24.50]) by runyon.cygnus.com (8.8.7-cygnus/8.8.7) with ESMTP id TAA11958; Tue, 7 May 2002 19:18:53 -0700 (PDT) Message-ID: <3CD887FF.1B553507@redhat.com> Date: Tue, 07 May 2002 19:19:00 -0000 From: Michael Snyder Organization: Red Hat, Inc. X-Accept-Language: en MIME-Version: 1.0 To: Elena Zannoni CC: Michael Snyder , gdb-patches@sources.redhat.com, jimb@redhat.com Subject: Re: [RFA] Add vector type flag to stabsread.c References: <200205062331.g46NV4k26237@reddwarf.sfbay.redhat.com> <15576.35004.495092.847403@localhost.redhat.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-SW-Source: 2002-05/txt/msg00195.txt.bz2 Elena Zannoni wrote: > > Michael Snyder writes: > > > > OK. > Should the stabs doc be updated? Who controls that? > > Elena Are you standing in for Eli today? ;-) OK, I see where it should go, I'll add it tomorrow. > > This patch does to stabs what Elena has recently done for dwarf -- > > add an attribute extension that allows an array type to be tagged > > as a vector. Knowing that it is a vector, GDB can then modify the > > way the object is passed as a function argument. > > > > The syntax for stabs will be to add the string "@V;" in front of > > the "ar" declaration. The "@" attribute extension already exists > > in stabs -- we're just adding a new instance of it. > > > > The GCC patch for this extension should be coming out today. > > > > 2002-05-06 Michael Snyder > > > > * stabsread.c (read_type): Add recognition for new attribute: > > "@V;" means that an array type is actually a vector. > > This is analogous to the vector flag that's been added to dwarf2. > > > > Index: stabsread.c > > =================================================================== > > RCS file: /cvs/src/src/gdb/stabsread.c,v > > retrieving revision 1.32 > > diff -p -r1.32 stabsread.c > > *** stabsread.c 4 May 2002 00:21:09 -0000 1.32 > > --- stabsread.c 6 May 2002 23:39:00 -0000 > > *************** define_symbol (CORE_ADDR valu, char *str > > *** 1904,1909 **** > > --- 1904,1910 ---- > > break; > > > > case 't': > > + /* Typedef */ > > SYMBOL_TYPE (sym) = read_type (&p, objfile); > > > > /* For a nameless type, we don't want a create a symbol, thus we > > *************** read_type (register char **pp, struct ob > > *** 2359,2364 **** > > --- 2360,2368 ---- > > /* Used to distinguish string and bitstring from char-array and set. */ > > int is_string = 0; > > > > + /* Used to distinguish vector from array. */ > > + int is_vector = 0; > > + > > /* Read type number if present. The type number may be omitted. > > for instance in a two-dimensional array declared with type > > "ar1;1;10;ar1;1;10;4". */ > > *************** again: > > *** 2574,2580 **** > > forward-referenced), and we must change it to a pointer, function, > > reference, or whatever, *in-place*. */ > > > > ! case '*': > > type1 = read_type (pp, objfile); > > type = make_pointer_type (type1, dbx_lookup_type (typenums)); > > break; > > --- 2578,2584 ---- > > forward-referenced), and we must change it to a pointer, function, > > reference, or whatever, *in-place*. */ > > > > ! case '*': /* Pointer to another type */ > > type1 = read_type (pp, objfile); > > type = make_pointer_type (type1, dbx_lookup_type (typenums)); > > break; > > *************** again: > > *** 2732,2747 **** > > > > switch (*attr) > > { > > ! case 's': > > type_size = atoi (attr + 1); > > if (type_size <= 0) > > type_size = -1; > > break; > > > > ! case 'S': > > is_string = 1; > > break; > > > > default: > > /* Ignore unrecognized type attributes, so future compilers > > can invent new ones. */ > > --- 2736,2757 ---- > > > > switch (*attr) > > { > > ! case 's': /* Size attribute */ > > type_size = atoi (attr + 1); > > if (type_size <= 0) > > type_size = -1; > > break; > > > > ! case 'S': /* String attribute */ > > ! /* FIXME: check to see if following type is array? */ > > is_string = 1; > > break; > > > > + case 'V': /* Vector attribute */ > > + /* FIXME: check to see if following type is array? */ > > + is_vector = 1; > > + break; > > + > > default: > > /* Ignore unrecognized type attributes, so future compilers > > can invent new ones. */ > > *************** again: > > *** 2844,2852 **** > > type = read_array_type (pp, type, objfile); > > if (is_string) > > TYPE_CODE (type) = TYPE_CODE_STRING; > > break; > > > > ! case 'S': > > type1 = read_type (pp, objfile); > > type = create_set_type ((struct type *) NULL, type1); > > if (is_string) > > --- 2854,2864 ---- > > type = read_array_type (pp, type, objfile); > > if (is_string) > > TYPE_CODE (type) = TYPE_CODE_STRING; > > + if (is_vector) > > + TYPE_FLAGS (type) |= TYPE_FLAG_VECTOR; > > break; > > > > ! case 'S': /* Set or bitstring type */ > > type1 = read_type (pp, objfile); > > type = create_set_type ((struct type *) NULL, type1); > > if (is_string)