Re: [avr-libc-dev] Automated testing project

[Joerg Wunsch]

As Peeter Vois wrote:

> As my proposed testsuite uses gdb serial protocol maybe it would be
> possible to set up hardware testplatforms for all / missing devices?

I don't quite understand, how's that going to work?

Normally, I'd assume that simulavr/simulavrxx/avarice are talking
GDB remote protocol.

Could you draw an outline of how the pieces are working together?

> We could improve the simulavr too in this sence if we will have more
> resources.

Forget simulavr, it's unmaintained.  simulavrxx is the designated
successor (it lives in the same savannah project), and has at least
two active maintainers (Klaus Rudolph, being the original author, and
someone who calls himself Bill, and who does the `organizational' work
around it, like CVS integration, bug report handling, auto* script and
all that).  I guess they might accept someone else to work in the
project, in particular the auto* scripts have some rough edges, and
don't work out of the box on my FreeBSD (I tried to quickly create a
FreeBSD port for simulavrxx, but failed in that so far).  I'd hope you
could also work on things like supporting new devices so the simulator
would be useful as a test bed for adding new device support to the
library.  So far, my only verification for new devices is to wait
until I've got a sample of them I can test against (which includes the
risk that this sample might easily be trashed during programming,
BTDT).

> 1. Import the MUL modifications into CVS

Fine.  I think this could become your `introduction' to the avr-libc
team, and if all looks fine, I'm personally more than happy to add you
as a new team member then.

> 2. go on with the log() optimisations
> 3. improve the testsuite to have statistical module for results
> 4. Import the testsuite into CVS somewhere

If it fits into the framework of avr-libc (i.e. if it could be used to
prove the library being correct), be invited.  If it's going to be a
completely different beast, you might consider creating your own
project on savannah as well.

> 5. I remember that I had some year ago porblem with constant + (nan) I
> would check this too

There's a number of bug reports open for the floating point stuff, and
the current FP library is seriously limited in some respects.  We
really need an FP maintainer.  If you want to become that, the better.

> Peeter

> Ühel ilusal kenal päeval Tue, 18 Oct 2005 23:36:39 +0200
> Joerg Wunsch <address@hidden> kirjutas:

Please don't write full quotes below your message.  It looks ugly,
wastes Internet bandwidth, serves no purpose, and is hard to read.

-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)