Instructions to send a Patch

Vaibhav Gupta vaibhavgupta40 at gmail.com
Wed May 29 02:23:13 UTC 2019


On Tue, May 28, 2019 at 11:47 AM Aditya Upadhyay <aadit0402 at gmail.com>
wrote:

> On Mon, May 27, 2019 at 10:42 PM Vaibhav Gupta <vaibhavgupta40 at gmail.com>
> wrote:
> >
> > Hello, okay I will keep that in mind.
> > .
> > .
> > .
> > .
> > But there is one problem with this patch then.
> > .
> > .
> > If i just copy-paste ndbm.h, db.h(needed by ndbm.h) and ndbm.c and make
> the entry of ndbm.c  in Makefile.am.
> > Then run 'autoreconf' to add the entry for ndbm.c in Makefile.in.  And
> then run `make` command. There will be compilation error.
> > .
>
> See Instructions here about regenerating the configuration files in newlib.
> https://sourceware.org/newlib/README
> Newlib uses a specific version of autoconf. You can use `autoreconf
> -fvi` command to regenerate that
> Makefile.in file.
>
Okay

>
> > Header files needed by libc/posic/*.c files are present in libc/include
> directory. It is default path.
> > .
> > But ndbm.c also requires "hash.h" which is present inside "libc/search".
> > .
> > To over come it, i had to modify the entry of ndbm.c in Makefile.in and
> add "-I (topr_srcdir)/search" to specify the compiler for alternative
> source path.
>
> If you specify any header file inclusion in Makefile.am, then you will
> need to add the things manually
> and when you will run autoreconf command, this .in file will get updated.
>
> I guess I will need to modify configure.in file. If I am not wrong that
is responsible for what autoreconf is doing.

> > .
> > .
> > The problem is, to modify Makefile.in there should be entry of ndbm.c .
> And for that i need to run autoreconf command. Because of that multiple
> files get modified and hence the patch size got bigger..
> > .
> > .
>
> I think, you have the problem with your git cloning of newlib-repo.
> Otherwise It would not have generated this much
> bigger patch.
>
No, Actually inclusion of any .c file affects almost 300 + files. They get
modified when autoreconf is run. 'git diff' gave a rough idea, this is why
patch got bigger.

>
> >
> > One solution can be that I should not use autoreconf and make manual
> entries for ndbm.c in Makefile.in . Should I try that?
> > Because there is need to mention alternative source path for header
> files, to compile ndbm.c.
> >
> You can do this manually, but it would be a wise approach to use
> autoreconf.
>
> > .
> > Thankyou
> > Vaibhav Gupta
> >
> >
> > On Mon, 27 May, 2019, 7:38 PM Aditya Upadhyay, <aadit0402 at gmail.com>
> wrote:
> >>
> >> Hi Vaibhav,
> >>
> >> You have sent a completely wrong patch. You do not need to include all
> >> these files what are existed there. You should generate a patch for
> >> your work like addition or modification or if you have added any other
> >> file.
> >>
> >> This patch should go to newlib mailing list. For your entire GSoC
> >> period, You will send patch to Newlib-Cygwin and For testing
> >> contributed methods, you will be adding testcases in rtems/testsuite.
> >>
> >> For testing those methods, you will have to apply patches to
> >> rtems-source-builder, see instructions here:
> >>
> https://docs.rtems.org/releases/rtemsdocs-4.10.2/share/rtems/html/started/started00036.html
> >>
> >> If you have not introduced yourself on newlib mailing list, then
> >> Introduce yourself there. Point your plan what you are going to do
> >> this GSoC Period. Though these are parts of Community bonding period.
> >>
> >> I appreciate your effort. This is just an initial phase of GSoC. You
> >> should take care of these things.
> >>
> >> Thanks,
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20190529/57086c47/attachment-0002.html>


More information about the devel mailing list