Hmm all right.  That sounds a little tight since we probably won't be able
to properly test everything with a new compiler; as I understood it we were
supposed to release by the end of the month.  Sounds like we better start
moving to the uniarch for our own development and maybe do the release with
it marked as "experimental," and also make it clear that this will be the
last release where 3.2.3 is supported.

Thanks for the insights,
Steve

On Tue, Jun 7, 2011 at 6:32 PM, Peter Bigot <[email protected]> wrote:

> Depends on when 2.1.2 is anticipated.  When uniarch moves to 4.5.3 it will
> be supported, with bug-fix only releases.  New features (including 20-bit
> support) will be prototyped in the 4.6.x branches.  Using
> uniarch-mspgcc-4.5.3-20110701 (or whatever) would be the way I would
> recommend, rather than hitching your wagon to the dead mspgcc4 horse.
>
> Peter
>
>
> On Tue, Jun 7, 2011 at 8:15 PM, Stephen Dawson-Haggerty <
> [email protected]> wrote:
>
>> I'll give uniarch a try, thanks for the status update.
>>
>> I guess the real question I'm asking is which, if any, of these should we
>> release with?  I think it's worth having a supported compiler version since
>> we do hit bugs given we're compiling generated code...
>>
>> Steve
>>
>>
>> On Tue, Jun 7, 2011 at 6:09 PM, Peter Bigot <[email protected]> wrote:
>>
>>> uniarch mspgcc is the new mspgcc4; there are no plans to release another
>>> mspgcc4 version, and support and code development has returned to the mspgcc
>>> project on sourceforge.  We may be within a week or two of having uniarch
>>> available in the experimental Debian package area, for those who don't want
>>> to build it by hand (the tinyos community is on the list to notify once that
>>> happens).  In fact, everything's there already except for msp430-libc, where
>>> we're still cleaning up copyright issues.
>>>
>>> All that's good in mspgcc4 is better in uniarch.  The current master
>>> branches are based on gcc 4.5.2, but the official first release will be
>>> based on 4.5.3.
>>>
>>> Peter
>>>
>>> On Tue, Jun 7, 2011 at 4:10 PM, Stephen Dawson-Haggerty <
>>> [email protected]> wrote:
>>>
>>>> I haven't seen any recent discussion about moving to mspgcc4 for 2.1.2,
>>>> but I know people are using it and it has significantly better support for
>>>> upcoming msp revisions.  My question is if this is something we can shift 
>>>> to
>>>> recommending for the release, since it generates 5-10% smaller code with
>>>> more coming down the line, and the tree currently is known to work with it
>>>> at least on telosb and epic.
>>>>
>>>> Perhaps we could "recommend" mspgcc4 (some specific version which we
>>>> decide) and "deprecate" 3.2.3 for this release, but still test with both
>>>> compilers with the expectation we will move to the newer one only for 
>>>> future
>>>> releases?
>>>>
>>>> Steve
>>>>
>>>> --
>>>> stephen dawson-haggerty
>>>> http://cs.berkeley.edu/~stevedh
>>>> uc berkeley wireless and embedded systems lab
>>>> berkeley, ca 94720
>>>>
>>>> _______________________________________________
>>>> Tinyos-devel mailing list
>>>> [email protected]
>>>>
>>>> https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-devel
>>>>
>>>>
>>>
>>
>>
>> --
>> stephen dawson-haggerty
>> http://cs.berkeley.edu/~stevedh
>> uc berkeley wireless and embedded systems lab
>> berkeley, ca 94720
>>
>
>


-- 
stephen dawson-haggerty
http://cs.berkeley.edu/~stevedh
uc berkeley wireless and embedded systems lab
berkeley, ca 94720
_______________________________________________
Tinyos-devel mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-devel

Reply via email to