[avr-gcc-list] AVR port status

[Weddington, Eric]

Hi All,

The list of known AVR GCC bugs is here:
<http://www.nongnu.org/avr-libc/bugs.html>

62 Total Reported Bugs
17 UNCONFIRMED Bugs

There are a number of these bugs that are reported for old versions. If you 
have the ability to build GCC from a snapshot, then please help by building the 
latest GCC 4.5 snapshot and test some of these bugs to see if they are still 
valid for recent versions. If they are no longer valid, we would like to close 
the bug, or if they are still valid, we would like to update the information.

The priorities for bugs:
1. wrong code bugs
2. internal compiler error (ICE) bugs
3. missed optimizations

We know that there are a lot of missed optimization bugs. And while these bugs 
have a lower priority, we would still like to get their information updated. 
There are a number of old missed optimization bugs that we need to know if they 
are still a problem on newer versions.

Also note that C language bugs will have a higher priority over C++ (or other 
language) bugs.

A Very Special Thanks to Andy Hutchinson! He has been very diligent recently in 
testing with the GCC Regression Test Suite, discovering and reporting new bugs, 
and fixing them! :-)


The latest GCC 4.5 Status Report can be found here:
<http://gcc.gnu.org/ml/gcc/2010-01/msg00398.html>
Please note the sentence: "There will be a release candidate made available 
when there are no remaining P1 regressions on the trunk." There are currently 
19 of these P1 regressions (affecting other areas or targets). Which means that 
a GCC 4.5.0 Release Candidate could possibly be made sometime in the next 1-3 
weeks (depending on how fast these regressions get fixed). According to the GCC 
4.5 Status Report, the trunk is currently in "regression and documentation 
fixes only mode". This means that we probably won't be able to get any more 
bugs fixed for the AVR port before 4.5.0 is released. However, we need to know 
where the AVR port stands regarding existing bugs so we can get them fixed over 
future versions.

Thanks,
Eric Weddington