Mirror of the gdb mailing list
 help / color / mirror / Atom feed
From: Geoffrey Keating <geoffk@geoffk.org>
To: Devang Patel <dpatel@apple.com>
Cc: gcc-patches@gcc.gnu.org, gdb@sources.redhat.com
Subject: Re: [PATCH] Debugging Vector Types
Date: Sat, 23 Apr 2005 00:13:00 -0000	[thread overview]
Message-ID: <m2d5smgxio.fsf@greed.local> (raw)
In-Reply-To: <547BC9E3-4C03-4724-8BB5-BB1A99BB3625@apple.com>

Devang Patel <dpatel@apple.com> writes:

> GCC encodes Vector Types as an array of N elements (at least, in
> STABS format). This means, GDB is not able to distinguish it from a
> normal Array. And these may lead to confusing output from GDB. One of
> our user reported that for
> 
>      vector unsigned char a = (vector unsigned char)  (
> 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16 );
> 
> gdb prints
> 
>      (gdb) print a
>      $1 = "\001\002\003\004\005\006\a\b\t\n\v\f\r\016\017\020"
> 
> If following is used
> 
>      vector unsigned char a = (vector unsigned char)  (
> 'a','b','c','d','e','f','g','h','i','j','k','l','m','n','o','p' );
> 
> then gdb prints
> 
>      (gdb) print a
>      $1 = "abcdefghijklmnop"
> 
> but, it's not vector!
> 
> If GCC provides GDB some kind of hints then GDB may be able to
> display something meaningful. This patch attaches "attribute vector"
> to the stabs string used to represent vector types. I am told that
> GDB ignores attributes that it does not recognize. So we can update
> GCC independently. I tested small test cases and our darwin GDB does
> not get offended by this additional attribute.
> 
>          * dbxout.c (dbxout_type): Emit attribute vector.
>          * config/darwin.h (DBX_USE_VECTOR_ATTRIBUTE): Define.
>          * doc/tm.texi (DBX_USE_VECTOR_ATTRIBUTE): Document.
> 
>          * gcc.dg/stabs-attrib-vect.c: New test.
> 
> Bootstrapped and tested on powerpc-darwin.
> OK  for mainline ?

- I don't think this should be Darwin-specific.  Shouldn't everyone have it?
  It's not like it can break Solaris dbx or something, unless people are
  using vector types, in which case IMO they should just use GDB.

- You need to update stabs.texi, which I think lives in GDB.  (It's
  the only stabs documentation we have, so keeping it up-to-date is
  important!)

Other than that the patch seems reasonable enough to me.

The GDB people may have some more comments, so I've CCed them.

GDB people: As I understand the patch, what it does is add '@V;' to
the end to the stab for the type, so it looks like "vi:(0,16)=@V;".


       reply	other threads:[~2005-04-23  0:13 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <547BC9E3-4C03-4724-8BB5-BB1A99BB3625@apple.com>
2005-04-23  0:13 ` Geoffrey Keating [this message]
2005-04-23  0:36   ` Devang Patel
2005-04-23  0:51     ` Geoff Keating
2005-04-24 22:35       ` Daniel Jacobowitz
2005-04-25  2:19         ` Devang Patel

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=m2d5smgxio.fsf@greed.local \
    --to=geoffk@geoffk.org \
    --cc=dpatel@apple.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=gdb@sources.redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox