This is the mail archive of the cygwin mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Repositories for Cygwin packages.

On 09/10/2015 05:20 PM, David A Cobb wrote:

>>> I am looking at possible work within *COREUTILS*.  Obviously, there are
>>> significant deltas /versus/ GNU Upstream.
>>> Can you point me to the active repo for coreutils?
> Yeah, Marco.  Thanks.
> Actually, the repo is <git://>.
> Are you saying that is your direct upstream and your sources only differ
> by the patchfiles installed by Cygwin-Setup??

Yes, the cygwin build of coreutils is made by taking the upstream
tarball release (which is created from the upstream coreutils.git at
labeled points in time) and then adding additional patches which are
distributed (as required by the GPL) in the source package that you can
download using setup.exe.

> I  should have phrased the question differently, I guess.  My question
> is related to dependencies of the 'coreutils.' Cloning GNU Coreutils
> comes in with sub-module 'gnulib'.
> Suppose I wanted to propose a patch to Coreutils, but being stuck on a
> Windows platform I use Coreutils only through Cygwin64 and MSYS2.

Not a problem. My first patch to upstream coreutils was done exactly in
that manner.

> And, suppose for the moment, some of the changes are only relevant to
> the Windows platform.  I don't (yet) know how much GNU (i.e. RMS) really
> gives a flying bird about making Windows play nice.  So, to whom do I
> propose the changes?  I really, really don't want to create a private
> fork.  If I didn't think my ideas are worthy of pushing up the food
> chain, I should just go back to bed.

Depends on how invasive your changes are. If it is truly
windows-specific and hard to maintain, then upstream probably won't pay
attention (which is why I maintain some cygwin-specific patches, such as
.exe magic manipulations, downstream-only). But if it fixes a bad
upstream assumption (such as "function foo would never do that", except
that on cygwin function foo DOES do that, and it is feasible that some
other system would do likewise), then upstream is the right place. (For
example, my very first patch to upstream coreutils is dated 2005-01-11,
where I fixed to deal with $(EXEEXT) - and more than just
cygwin creates binaries with .exe suffix so it is relevant upstream).

If you're unsure whether a proposed patch is worth posting upstream or
downstream, pick one place, and I'm more than willing to help you
redirect it to the other place if it wasn't appropriate.  (Picking
upstream first is generally a nicer policy).

Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library

Attachment: signature.asc
Description: OpenPGP digital signature

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]