Re: [ptxdist] [PATCH 3/3] host-openocd: add new package

[Andreas Pretzsch <apr@cn-eng.de>]

On Mo, 2014-06-23 at 16:09 +0200, Michael Olbrich wrote:
> On Sun, Jun 22, 2014 at 09:41:09AM +0200, Michael Grzeschik wrote:
> > Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de>
> > ---
> >  rules/host-openocd.in   | 184 ++++++++++++++++++++++++++++++++++++++++++++++++
> >  rules/host-openocd.make |  71 +++++++++++++++++++
> >  2 files changed, 255 insertions(+)
> >  create mode 100644 rules/host-openocd.in
> >  create mode 100644 rules/host-openocd.make
> > 
> > diff --git a/rules/host-openocd.in b/rules/host-openocd.in
> > new file mode 100644
> > index 0000000..8071f3c
> > --- /dev/null
> > +++ b/rules/host-openocd.in
> > @@ -0,0 +1,184 @@
> > +## SECTION=hosttools
> > +
> > +menuconfig HOST_OOCD
> > +	tristate
> > +	prompt "openocd"
> > +	select HOST_LIBUSB
> > +	select HOST_LIBUSB_COMPAT
> > +	select HOST_LIBFTDI1
> > +	help
> > +	  The "Open On-Chip Debugger" provides JTAG/SWD access from GDB
> > +	  (or directly with TCL scripts) to processors with ARM and MIPS
> > +	  based cores.
> > +
> > +if HOST_OOCD
> > +
> 
> This is a host tool, so we don't have any real space constraints. And no
> extra dependencies. Does it take a lot longer, when everything is enabled?
> If not, then I'd suggest to just enable everything and drop the options.

Not my patch, but I just send a similar one to the list, so I'll comment
on this (dead) discussion, too.

First, while there is some kind of default set selected by OpenOCD
configure depending on the present libs on the system, I doubt it will
be suitable for everyone. Either things are missing or (won't harm) too
broad. And, I wouldn't bet everything will always compile on all hosts.

Second, in the end, it's a pure development tool, which is probably
externally compiled by many users already. So integrating this as host
tool into ptxdist is more like a convenience thing, if at all.

Coming to Michael Grzeschiks patchset (thanks for that), I intended to
polish it a bit for my usage here (OpenOCD on the target).
Unfortunately, there are quite a few typos in the options, some are only
in the Kconfig, some only in the makefile. And others were missing.
Looks like the usual result of a patch living off-tree for quite some
version interations :-)

Anyway, due to this, I just rewrote the whole thing from scratch.
Mainly for OpenOCD on target, the host part is more a by-product, albeit
working.
In both cases, there are no options selected at all by default.
For the target, this won't make sense anyway, for the host it is
debatable, see above. Left as an exercise to the reader ;-)

In this respect I also ignored his patches for host-libftdi1 and
host-libusb-compat, which did not went mainline. I'd suppose anybody
needing them will pick them up, but that won't be me today.

Regarding my patch and comments to details, please see the mail
"[ptxdist] [PATCH] openocd: new package for openocd-0.8.0 (host,
target)" from "Mon, 29 Sep 2014 19:12:11 +0200" on the list.

-- 

carpe noctem engineering
Ingenieurbuero fuer Hard- & Software-Entwicklung Andreas Pretzsch
Dipl.-Ing. (FH) Andreas Pretzsch          Tel. +49-(0)731-5521572
Hahnengasse 3                             Fax: +49-(0)731-5521571
89073 Ulm, Germany                        email: apr@cn-eng.de


-- 
ptxdist mailing list
ptxdist@pengutronix.de