[avr-libc-dev] [patch #2651] signal and UART reisters name consistency

[Theodore A. Roth]

This mail is an automated notification from the patch tracker
 of the project: AVR C Runtime Library.

/**************************************************************************/
[patch #2651] Latest Modifications:

Changes by: 
                Theodore A. Roth <address@hidden>
'Date: 
                Tue 04/20/04 at 23:42 (US/Pacific)

            What     | Removed                   | Added
---------------------------------------------------------------------------
          Resolution | None                      | Fixed
              Status | Open                      | Closed


------------------ Additional Follow-up Comments ----------------------------
The register parts of this patch for the mega169 where done in the fix for bug 
#8452 (in cvs now).

The SIG_USART[01]_* changes for the mega162 will not be applied (this part is a 
WONT_FIX).

The SIG_USART_* changes for mega169 will be in cvs by the time  this patch is 
closed out.

I have posted some patches to the avr-libc dev list which allow gcc to generate 
an error if a signal name is misspelled which make the "_UART_SIGNAME_ERR +" 
hack suggested un-needed (assuming my patches go in).






/**************************************************************************/
[patch #2651] Full Item Snapshot:

URL: <http://savannah.nongnu.org/patch/?func=detailitem&item_id=2651>
Project: AVR C Runtime Library
Submitted by: 0
On: Fri 02/20/04 at 12:27

Category:  None
Priority:  3 - Low
Resolution:  Fixed
Assigned to:  troth
Originator Email:  address@hidden
Status:  Closed


Summary:  signal and UART reisters name consistency

Original Submission:  Beacuse m162 has only one UART then it is a little messy 
to use 0 in macro names. 
In the avr-libc (except of m169.h) there are used UART prefixes not USART.

Maybe it is also worth to prohibit users against using old names of macros, 
adding something like:

#define _UART_SIGNAME_ERR       +
#define SIG_UART0_RECV  _UART_SIGNAME_ERR
#define SIG_UART0_DATA  _UART_SIGNAME_ERR
#define SIG_UART0_TRANS _UART_SIGNAME_ERR

Follow-up Comments
------------------


-------------------------------------------------------
Date: Tue 04/20/04 at 23:42         By: troth
The register parts of this patch for the mega169 where done in the fix for bug 
#8452 (in cvs now).

The SIG_USART[01]_* changes for the mega162 will not be applied (this part is a 
WONT_FIX).

The SIG_USART_* changes for mega169 will be in cvs by the time  this patch is 
closed out.

I have posted some patches to the avr-libc dev list which allow gcc to generate 
an error if a signal name is misspelled which make the "_UART_SIGNAME_ERR +" 
hack suggested un-needed (assuming my patches go in).

-------------------------------------------------------
Date: Wed 04/07/04 at 16:58         By: troth
Actually, more than the m169 use USART. If the data sheet says USART, I've been 
told to use USART.

The current m169 header uses SIG_UART0_*. That should probably be changed to 
SIG_USART_*.

The m162 SIG_USART[01]_* probably shouldn't be changed.

Personally, I'd rather see all the signal names for all devices reduced down to 
common and consistent names. We are already divergent from the datasheets with 
respect to interrupt vector names and simplification would reduce the chance of 
using the wrong SIG_* name (which gcc can't even catch and will generate the 
wrong code).

I'd like to hear some more comments on this patch before we take action.






File Attachments
-------------------

-------------------------------------------------------
Date: Fri 02/20/04 at 12:27  Name: iom162_169_h.patch  Size: 3KB   By: None
patch for iom162.h and iom169.h
http://savannah.nongnu.org/patch/download.php?item_id=2651&amp;item_file_id=2755






For detailed info, follow this link:
<http://savannah.nongnu.org/patch/?func=detailitem&item_id=2651>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.nongnu.org/