On Monday 18 May 2009, Zach Welch wrote:
> The tactical goal of each release should to be focus on delivering a
> product with steadily increasing stability and device support (whether
> host, target, or interface).

I can't quite imagine what other goals might be.  ;)

So ... releases every three months or so, to make sure bugfixes
and new capabilities don't sit too long in the kitchen before
being served as entrees or appetizers?

There's also the issue of how to get proper "user feedback"
into the cycle.  Requiring folk to subscribe to the developer
list, or get a berlios account to submit a bug report, rather
minimizes the amount of such feedback.


>   Once we have machine-generated records of 
> certain combinations of hardware having worked with a certain version of
> OpenOCD, it will be virtually impossible for us to miss regressions with
> particular hardware combinations.

No, it'll still be really easy.  Especially at this level of
system development, hardware comes and goes.  Someone may
test with a board but need to return it, so they can't retry
with a later release.  Or it may break, or become obsolete
so it goes into a box and doesn't get retested.

The way to get reproducibility is to have dedicated test
hardware, and dedicated test resources ... Øyvind was working
on some of that.

But also, expect that goofs will happen, and make it easy
for "users" to report problems ... they will have access to
a *LOT* of hardware the main developers don't.

_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development

Reply via email to