Re: netbsd-10, merging device-streams (xstreamtodev, rdbinfo)

Hi David,

I updated the README and version strings, rebuilt with current
bebbo/amiga-gcc (both gcc and vbcc) and made a new release.

https://github.com/rvalles/device-streams/releases

I also downloaded your test build's binaries and ran them just to
ensure they're the new (now old...) versions; They are.

The new release's binaries have been run and tested by writing into my
swap partition above 4GB.

Please update with the new gcc-built binaries and README.md, and we're
good for actually pulling into netbsd.

Thanks


Roc

On Tue, 21 Mar 2023 at 01:04, David Brownlee <[email protected]> wrote:
>
> On Mon, 20 Mar 2023 at 13:19, Roc Vall=C3=A8s <[email protected]> wrote=
:
> >
> > Hi,
> >
> > The cut issue didn't seem right to me, so I looked at it again. Cut is
> > not the culprit, both cuts behave as expected.
> >
> > Sed is the actual culprit, Expression sed -e 's/[[:blank:]]\+/ /g'
> > seems to at least match something if the \+ is removed from the
> > expression, thus it doesn't like the \+ to match one or more.
> >
> > So I did some digging, and GNU is the odd one, whereas other
> > implementations that follow POSIX.2 do not do this. I understand it
> > isn't in the spec in the first place.
>
> GNU has a habit of adding many additional useful extensions to
> libraries tools... which is great if you are only using GNU tools, but
> a pain for musl libc, BSD and other users :)
>
> On a related note - test build of NetBSD-current with the latest
> device-streams is at:
> https://sync.absd.org/amiga/current/installation/misc/ (Full build
> https://sync.absd.org/amiga/current/ )
>
> I'll hold off for you to dig through the gnu sed gordian knot :) and
> maybe an a version with an updated README, and then commit to the tree
> for real and get a pullup to at least netbsd-10 (and maybe netbsd-9)
>
> Thanks
>
> David