Re: [Openocd-development] New patch to review for openocd: e043913 bugfixes: numerous bugs in error propagation found by clang

2011-10-27 Thread Peter Stuge
ger...@openocd.zylin.com wrote:
 -DumpTargets:
 +DumpTargets:;

Hm?


//Peter
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development


Re: [Openocd-development] New patch to review for openocd: e043913 bugfixes: numerous bugs in error propagation found by clang

2011-10-27 Thread Øyvind Harboe
2011/10/28 Peter Stuge pe...@stuge.se:
 ger...@openocd.zylin.com wrote:
 -DumpTargets:
 +DumpTargets:;

 Hm?

Syntactically a declaration can't follow a label, so add an empty statement.


-- 
Øyvind Harboe - Can Zylin Consulting help on your project?
US toll free 1-866-980-3434
http://www.zylin.com/
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development


Re: [Openocd-development] New patch to review for openocd: e043913 bugfixes: numerous bugs in error propagation found by clang

2011-10-27 Thread Peter Stuge
Øyvind Harboe wrote:
 2011/10/28 Peter Stuge pe...@stuge.se:
  ger...@openocd.zylin.com wrote:
  -DumpTargets:
  +DumpTargets:;
 
  Hm?
 
 Syntactically a declaration can't follow a label, so add an empty
 statement.

I don't like the change so much. It ends up declaring the exact same
variable twice, and other odd noise, to silence a warning. Is there
no better way?


//Peter
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development


Re: [Openocd-development] New patch to review for openocd: e043913 bugfixes: numerous bugs in error propagation found by clang

2011-10-27 Thread Øyvind Harboe
On Fri, Oct 28, 2011 at 12:23 AM, Peter Stuge pe...@stuge.se wrote:
 Øyvind Harboe wrote:
 2011/10/28 Peter Stuge pe...@stuge.se:
  ger...@openocd.zylin.com wrote:
  -DumpTargets:
  +DumpTargets:;
 
  Hm?

 Syntactically a declaration can't follow a label, so add an empty
 statement.

 I don't like the change so much. It ends up declaring the exact same
 variable twice, and other odd noise, to silence a warning. Is there
 no better way?

Reusing the 'target' variable name here is bad.

They are really two very distinct variable lifetimes.

The scope of the variables are now reduced.

The code is now more easily re-factored as multiple functions
because the same variable isn't being reused over and over for
difference purposes.


-- 
Øyvind Harboe - Can Zylin Consulting help on your project?
US toll free 1-866-980-3434
http://www.zylin.com/
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development


Re: [Openocd-development] New patch to review for openocd: e043913 bugfixes: numerous bugs in error propagation found by clang

2011-10-27 Thread Peter Stuge
Øyvind Harboe wrote:
  I don't like the change so much. It ends up declaring the exact same
  variable twice, and other odd noise, to silence a warning. Is there
  no better way?
 
 Reusing the 'target' variable name here is bad.
..
 The code is now more easily re-factored as multiple functions

Ok, can we please do that then, instead of working around static
analysis? Also, I'm not sure the command handler should really
return ERROR_OK if the target is unknown. (Not a new bug, but I
think it's a bug just the same.)


//Peter
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development


Re: [Openocd-development] New patch to review for openocd: e043913 bugfixes: numerous bugs in error propagation found by clang

2011-10-27 Thread Øyvind Harboe
On Fri, Oct 28, 2011 at 12:59 AM, Peter Stuge pe...@stuge.se wrote:
 Øyvind Harboe wrote:
  I don't like the change so much. It ends up declaring the exact same
  variable twice, and other odd noise, to silence a warning. Is there
  no better way?

 Reusing the 'target' variable name here is bad.
 ..
 The code is now more easily re-factored as multiple functions

 Ok, can we please do that then, instead of working around static
 analysis?

In this case, I disagree that it is working around static analysis.

I'll see about having  a peek at the code again for cleaning it up.


-- 
Øyvind Harboe - Can Zylin Consulting help on your project?
US toll free 1-866-980-3434
http://www.zylin.com/
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development