On Wed, 2009-04-22 at 08:04 +0200, Øyvind Harboe wrote:
> Great posting! Can I suggest that you create this as a patch against
> svn so we can match up the todo.txt list against an actual version of
> the OpenOCD svn history?

Can't I simply claim that it's on The List? ;)   I definitely agree that
I should be providing the revision number, but I have evolved the idea
that TODO files inside repositories get treated the same way as other
files do: people fear change.  They become monumental testimonies to the
good intentions of the ambitious.  I know; I've written my fair share.

Instead, I am hoping that it can serve more to guide contributors into
regular project management discussions.  Patches would hide all but the
latest activity, and I want people to be reminded and energized by the
web of goals that lurk behind the daily grind.  The List will give a
complete picture of the activities reported and pending; it only needs
be reposted after sufficient changes have taken place.

Ideally, folks will start to volunteer to cherry-pick tasks off of it.
I think we will see more of that if it gets posted regularly; it should
clearly be a fresh artifact and contain items that we really intend to
carry forward.  Since I am volunteering to take on this challenge of
trying to herd cats in this manner, I hope the community is willing to
embrace this idea for a time (or give me a better alternative).

Project management can be tricky.  I am still trying to get it right,
and -- though I may screw up often -- I hope my effort is appreciated.
In that vein, thanks to you and all others that have been supportive.
I live, I learn, and I try to improve.

> W.r.t. churn, I think it would be helpful, as you do, to try to get
> a lot of those "warning fixes" out of the way so that at least the
> overall codebase is a bit more stable and then hopefully the
> "churn" is localized to a set of files at any one time.

I think that we have passed the "warning fixes" stage at this point,
with -Werror surviving intact.  I might be wrong and some corner cases
remain to be reported, but I believe the community has chimed in on the
problems for all of the major ones.  Fixes are in tree for all of them,
as far as I am aware.  [[New bugs are another thing, as the recent post
from Andy Chenee shows.  Sorry about that one.]]

At this point, the use of -Werror will keep us from having to go through
that particular flavor of pain again, so things should only get better.

> > * Architectural Upgrades
> >  - Allow N:M:P mapping of servers, targets, and interfaces
> 
> - About CFI:
> 
> I'd like to see if the ecosflash driver could not be used to take
> a CFI driver from e.g. eCos and use that in OpenOCD. The
> current OpenOCD CFI driver is fine, such as it is, but e.g.
> eCos CFI drivers are getting a *LOT* of testing from a much
> bigger community.

Added, though you will need to check that I interpreted you correctly.

> - New targets?

Any that you have in mind?  Coldfire and Cortex are on thanks to Nico.
MC1322x from Jeff.  I have added an "insert your target here" item.

Cheers,

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

Reply via email to