Re: [patch 7/7] uml: resolve symbols in back-traces

Re: [patch 7/7] uml: resolve symbols in back-traces

[Blaisorblade]

On Thursday 28 October 2004 00:27, akpm@osdl.org wrote:
> From: Chris Wedgwood <cw@f00f.org>

> Resolve symbols in back-traces.

> Signed-off-by: Chris Wedgwood <cw@f00f.org>
> Signed-off-by: Andrew Morton <akpm@osdl.org>
> ---
The idea is right - I already had it in my tree, with some little changes (for 
instance including the symbol address and so on). So I'll send it reviewed a 
bit.

I had not yet sent it because there was more critical stuff and because 
sometimes on panic it does not work yet - this seems to be a missing flush 
from the UML console drivers :-(, a bit long to solve because currently the 
locking is broken (a semaphore is used in process context, nothing in 
interrupt context).

And removing the final Emacs comment is not welcome (I don't care, but Jeff 
does. If that should be removed, that's a separate problem).

Bye
-- 
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729

Re: [patch 7/7] uml: resolve symbols in back-traces

[Chris Wedgwood]

On Thu, Oct 28, 2004 at 08:34:21PM +0200, Blaisorblade wrote:

> And removing the final Emacs comment is not welcome (I don't care,
> but Jeff does. If that should be removed, that's a separate
> problem).

the emacs comments are gratuitous and completely pointless, they serve
no useful purpose.  fwiw in my .emacs i have:

    (defun cw-linux-c-mode ()
      "C mode with adjusted defaults for use with the Linux kernel."
      (interactive)
      (c-mode)
      (c-set-style "linux"))
    (setq auto-mode-alist
	  (append '(("wk/linux.*/.*\\.[ch]$" . cw-linux-c-mode))
		  '(("/usr/src/linux.*/.*\\.[ch]$" . cw-linux-c-mode))
		  auto-mode-alist))

which actually could be cleaned up a bit (it's been hacked over the
years and never cleaned up suitably) but the idea is pretty simple


  --cw

Re: [patch 7/7] uml: resolve symbols in back-traces

[Blaisorblade]

On Thursday 28 October 2004 21:28, Chris Wedgwood wrote:
> On Thu, Oct 28, 2004 at 08:34:21PM +0200, Blaisorblade wrote:
> > And removing the final Emacs comment is not welcome (I don't care,
> > but Jeff does. If that should be removed, that's a separate
> > problem).

> the emacs comments are gratuitous and completely pointless, they serve
> no useful purpose.  fwiw in my .emacs i have:

>     (defun cw-linux-c-mode ()
>       "C mode with adjusted defaults for use with the Linux kernel."
>       (interactive)
>       (c-mode)
>       (c-set-style "linux"))
>     (setq auto-mode-alist
>    (append '(("wk/linux.*/.*\\.[ch]$" . cw-linux-c-mode))
>     '(("/usr/src/linux.*/.*\\.[ch]$" . cw-linux-c-mode))
>     auto-mode-alist))
>
> which actually could be cleaned up a bit (it's been hacked over the
> years and never cleaned up suitably) but the idea is pretty simple

I mostly agree on this, and thought it myself (I *never* use Emacs, only Vim, 
but no flames here).

However, let's be kind with Jeff, since he's the UML maintainer. You can be 
right, but there's also courtesy.

If Jeff says "OK", then someone can submit a separate patch removing all the 
final comments.

Until then, please don't.

And never mix such unrelated changes in a patch.
-- 
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729

Re: [patch 7/7] uml: resolve symbols in back-traces

[Chris Wedgwood]

On Thu, Oct 28, 2004 at 11:02:48PM +0200, Blaisorblade wrote:

> If Jeff says "OK", then someone can submit a separate patch removing
> all the final comments.

he did

> And never mix such unrelated changes in a patch.

idealistically i agree, but doing trivial / obvious cleanups inside
other patches is fine if you ask me

it's not like we are mixing semantically different changes

Re: [patch 7/7] uml: resolve symbols in back-traces

[Jeff Dike]

cw@f00f.org said:
> the emacs comments are gratuitous and completely pointless, they serve
> no useful purpose.  fwiw in my .emacs i have: 

Yeah, that's why I acked that particular change.

They're not completely pointless, they just cater to an individual's development
environment, and that sort of stuff should be in the environment, and not the
code.

				Jeff

Re: [patch 7/7] uml: resolve symbols in back-traces

[Chris Wedgwood]

On Thu, Oct 28, 2004 at 05:32:43PM -0400, Jeff Dike wrote:

> They're not completely pointless, they just cater to an individual's
> development environment, and that sort of stuff should be in the
> environment, and not the code.

unnecessary in the code then, the exception to this perhaps being the
compile-command stuff that was (still is?) in some of the network
drivers as that really is per-file state

Re: [uml-devel] Re: [patch 7/7] uml: resolve symbols in back-traces

[Blaisorblade]

On Thursday 28 October 2004 22:51, Chris Wedgwood wrote:
> On Thu, Oct 28, 2004 at 05:32:43PM -0400, Jeff Dike wrote:
> > They're not completely pointless, they just cater to an individual's
> > development environment, and that sort of stuff should be in the
> > environment, and not the code.
>
> unnecessary in the code then, the exception to this perhaps being the
> compile-command stuff that was (still is?) in some of the network
> drivers as that really is per-file state
??? I don't understand you well. If there are compile-commands for emacs, 
they're broken too - using make namefile.o ARCH=um is the kernel universal 
solution.
-- 
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729

Re: [uml-devel] Re: [patch 7/7] uml: resolve symbols in back-traces

[Chris Wedgwood]

On Fri, Oct 29, 2004 at 01:44:11AM +0200, Blaisorblade wrote:

> ??? I don't understand you well. If there are compile-commands for
> emacs, they're broken too - using make namefile.o ARCH=um is the
> kernel universal solution.

consider a load-able module for uml ... the compile command might be
to build and load the module into a running UML instance...  for some
people that's a pretty nice working module where you dont have to
leave your editor to boot/run test things

Re: [uml-devel] Re: [patch 7/7] uml: resolve symbols in back-traces

[Blaisorblade]

On Thursday 28 October 2004 23:32, Jeff Dike wrote:
> cw@f00f.org said:
> > the emacs comments are gratuitous and completely pointless, they serve
> > no useful purpose.  fwiw in my .emacs i have:
>
> Yeah, that's why I acked that particular change.
>
> They're not completely pointless, they just cater to an individual's
> development environment, and that sort of stuff should be in the
> environment, and not the code.
>
>     Jeff
Ok. In fact I already agreed, and I was a bit unconfortable with that. However 
I'm working on other stuff, so I never worried about that, while seeing you 
adding such comments lately. I'm also concerned about not stepping on other 
people toes.
-- 
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729