GSoC 2019 | POSIX Compliance - RSB patch generated ndbm library (lib_a-ndbm.o) successfully in RTEMS Toolchain

[Vaibhav Gupta]

On Mon, Jul 15, 2019 at 10:21 PM Gedare Bloom <gedare at rtems.org> wrote:

> On Thu, Jul 11, 2019 at 9:01 PM Vaibhav Gupta <vaibhavgupta40 at gmail.com>
> wrote:
> >
> > Sounds good to me.
> >
> > On Thu, Jul 11, 2019 at 6:47 PM Joel Sherrill <joel at rtems.org> wrote:
> >>
> >>
> >>
> >> On Thu, Jul 11, 2019 at 2:12 AM Vaibhav Gupta <vaibhavgupta40 at gmail.com>
> wrote:
> >>>
> >>> Hello,
> >>> After Joel pointed out in an offlist discussion,
> >>> I made a new patch for ndbm port.
> >>> .
> >>> To send the changes to Newlib, i had to place `ndbm.h` , `ndbm.c` in
> their respective places and make  changes in Makefile.am.
> >>> Before, I applied same patch to RSB hence ndbm library was not
> generated.
> >>> .
> >>
> >> In my local build yesterday, I saw the symbols in the installed libc.a.
> I have not run the tests.
> >>
> >>>
> >>> .
> >>> This time I also added files generated by `autoreconf -fvi` in the
> patch.
> >>> .
> >>> This patch is 10MB in size hence cannot be send in raw format on
> mailing list.
> >>
> >>
> >> The person committing is supposed to do the autoreconf and commit that.
> >>
> >> No one has answered if it is OK to commit. That was the last message in
> the thread.
> >
> > I will ping on that thread again for confirmation.
> >>
> >>
> >>>
> >>> .
> >>> This patch worked with RSB and ndbm library (lib_a-ndbm.o) was
> generated
> >>> successfully in RTEMS Toolchain.
> >>
> >>
> >> I'm hoping we can avoid this by pushing the patch to newlib, then
> bumping the hash for
> >> newlib in the RSB, then adding your ndbm test patch to RTEMS.
> >
> > Yeah, meanwhile testsuite can be verified.
> > .
> > Also, please look at the sources I send on devel for fenv. Should I
> ignore architectures
> > which are not having FreeBSD source? or Should i pick from NetBSD and
> FreeBSD
> > both?
> Pick from both, figuring out how to translate the NetBSD
> implementation to work with how the FreeBSD ones are implemented, if
> necessary. Mostly, they define very similar data structures,
> macros/functions, etc., so I think it should be possible to unify
> them.
>
Sure, then I will work in this direction. Also, SPARC has only NetBSD
source,
so I guess it can be ported directly from it.

-Vaibhav Gupta

>
> Gedare
>
> >>
> >>
> >> --joel
> >>
> >>>
> >>>
> >>>
> >>> Thank you
> >>> Vaibhav Gupta
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20190715/1f78da9e/attachment-0002.html>