Hi Michal,
I've cc'ed -upsdev, since some info could be interesting for everybody.
2011/5/27 Michal Soltys <soltys at ziu.info>
> On 11-05-25 09:35, Arnaud Quette wrote:
> > Hi Michal,
> >
> > 2011/5/25 Michal Soltys <soltys at ziu.info <mailto:soltys at
ziu.info>>
> >(...)
> >
> > I'm preparing larger patchset with everything discussed
earlier:
> >
> > - icanon
> > - commands taking arguments
> > - variable issues (to avoid ups.firmware.old - like tricks)
> >
> >
> > thanks for this update. no pb, since there is 2.6.1 planned for a few
> > days from now.
> > there will probably be a 2.6.2 somewhere during the summer.
>
> It should be ready then. I think much earlier, I need some more testing
> though (and there's still documentation to be added at the end).
>
do you think you'll have more commits for 2.6.1.
depending on that, and the remaining patches to be merged, this should be
out either this week end (also depending on the weather!) and early next
week...
> >
> > side question: have you worked with icanon on win32?
>
> Haven't really peeked over win32 port too much. How do the other upses
> (canonical mode or not) behave so far ? In that monster patch from
> mailing list, I saw a compatibility layer for stuff not exisiting in mingw
> (I think) - such as tcsetattr, etc. How it works in practice is of course
> entirely different matter, but it's not like there's some difficult
magic
> being
> done with serial ports either (be it apcsmart or other serial drivers).
>
> Either way, from what I can see - function I currently use for reading
> from ups should be icanon-agnostic, with major difference being billion
> selects vs. 1
>
> Overall, the code changes on my side so far:
>
> 01:12 > git diff --stat ^99ea29b3d43a5c HEAD
> drivers/Makefile.am | 2 +-
> drivers/apcsmart.c | 1399
> ++++++++++++++++++++++++++++++++---------------
> drivers/apcsmart.h | 365 ++++---------
> drivers/apcsmart_tabs.c | 149 +++++
> drivers/apcsmart_tabs.h | 100 ++++
> 5 files changed, 1297 insertions(+), 718 deletions(-)
>
> apcsmart: [temporary]
> apcsmart: cleanups, split into two files
> apcsmart: next changes to instcmd stuff
> apcsmart: simplify query_ups()
> apcsmart: add apc_write_rep() and apc_write_long(), initial changes to
> instcmd stuff
> apcsmart: allow tearing down/setting up serial connection during normal
> activity; changes should fix #535583 as well
> apcsmart: trivial changes
> apcsmart: enable icanon
> apcsmart: handle 1:n vars (nut:apc) variables, so we don't need tricks
such
> as ups.firmware.old
> apcsmart: reintroduce ignored set in protocol_verify() for not exposed
> cmds/vars (=> a bit more quiet logs)
> apcsmart: introduce icanon upsread() and upsflush() (functions will be used
> after 'icanon' commit)
> apcsmart: add quick reference about alerts and to-be-removed defines.
> Introduce SER_* defines what will be used in new functions.
> apcsmart: add command '@' to compatibility tables.
> apcsmart: trivial changes
>
>
> Not a small part in the above is code shuffling with relatively small
> changes. But
> there's a good chunk of new things as well (+ old commits on my behalf)
>
> Out of curiosity - I wonder, have you considered migrating to git at some
> point in the future ?
>
seriously, for more than a year.
even more that I'm using it for my debs, and now that it's available on
Alioth.
The thing is that I'm considering the general needs, so not only the (D)VCS,
but more generally the tools that goes along, to track roadmap, features,
bugs, ... And gforge is still not enough agile nor inline with the
state-of-the-art there.
I've not yet had enough time to look at github, but there are interesting
points.
Trac has also some points, as launchpad, Rally (for specifications), but
none of these suits all our needs.
Any ideas / comments / help welcomed.
cheers,
Arnaud
--
Linux / Unix Expert R&D - Eaton - http://powerquality.eaton.com
Network UPS Tools (NUT) Project Leader - http://www.networkupstools.org/
Debian Developer - http://www.debian.org
Free Software Developer - http://arnaud.quette.free.fr/
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.alioth.debian.org/pipermail/nut-upsdev/attachments/20110527/50f9f0f4/attachment.html>
