civetweb, sts work

civetweb, sts work

[Marcus Watts]

I've just pushed out two new ceph branches, and I pushed out
a revised copy of civetweb that these both use.

I've pushed my git branches to git@github.com:ceph/ceph.git
They are:
wip-openssl-2
	( based on recentish master. )
	ssl patches (should be the same as what's already in master)
	+ "build out of tree" am changes (as described previously)
	+ civetweb "ng" changes.
		I've managed to build this with both cmake and configure.
		I haven't had time to do any testing past that yet.
		It "should" all work of course...
wip-rgw-sts-3
	everything that's in wip-openssl-2
	+ rgw-sts work.
	includes polymorphic rgw error object
	split out of some parts of rgw rest interface into objstore
		+ "generic" bits. (sts isn't an object store)
	basic sts encode/decode stuff.
		I've managed to build this with both cmake and configure.
		I haven't had time to do any testing past that yet.
		All the basic bits should work.
		The radosgw parts should work if I haven't broken them.
		The sts parts are still missing a lot of functionality.

I pushed my civitweb stuff to https://github.com/mdw-at-linuxbox/civetweb
Probably should should go to ceph/civetweb but I haven't had time
to follow up on that yet.
wip-easy1
	( based on recent upstream civetweb release)
	+ cmake build changes (builds w/ cmake on linux)
	+ modifications to support all of the functionality used by ceph,
		hopefully in a form that can be accepted by the civetweb
		folks upstream.
		I've built this standalone, and gotten it to pass all
		of its self-test logic.  I haven't done any other testing yet.

I imagine I'll find out "tomorrow" what gitbuilder thought
of all this.
	{ Hm.  rocksdb might catch me up.  I'm building with "rocksdb-devel",
		but I haven't touched ceph.spec.in ... }

Next steps on the roadmap,
fix whatever gitbuilder finds
auth_mellon-like functionality for sts
radosgw token changes
radosgw userdb changes

other bits along the side,
push rocksdb changes to rocksdb upstream
push civetweb changes to civetweb upstream
document using openssl w/ civetweb
{ and figure out how to hack gitbuilder rocksdb rpm/deb stuff... }

						-Marcus Watts

Re: civetweb, sts work

[Matt Benjamin]

Hi Marcus,

Can you pull up wip-rgw-sts-3 (it's last commit from upstream looks to be Feb. 26).

Thanks!

Matt

----- Original Message -----
> From: "Marcus Watts" <mwatts@redhat.com>
> To: ceph-devel@vger.kernel.org
> Sent: Wednesday, March 9, 2016 3:35:04 AM
> Subject: civetweb, sts work
> 
> I've just pushed out two new ceph branches, and I pushed out
> a revised copy of civetweb that these both use.
> 
> I've pushed my git branches to git@github.com:ceph/ceph.git
> They are:
> wip-openssl-2
> 	( based on recentish master. )
> 	ssl patches (should be the same as what's already in master)
> 	+ "build out of tree" am changes (as described previously)
> 	+ civetweb "ng" changes.
> 		I've managed to build this with both cmake and configure.
> 		I haven't had time to do any testing past that yet.
> 		It "should" all work of course...
> wip-rgw-sts-3
> 	everything that's in wip-openssl-2
> 	+ rgw-sts work.
> 	includes polymorphic rgw error object
> 	split out of some parts of rgw rest interface into objstore
> 		+ "generic" bits. (sts isn't an object store)
> 	basic sts encode/decode stuff.
> 		I've managed to build this with both cmake and configure.
> 		I haven't had time to do any testing past that yet.
> 		All the basic bits should work.
> 		The radosgw parts should work if I haven't broken them.
> 		The sts parts are still missing a lot of functionality.
> 

-- 
Matt Benjamin
Red Hat, Inc.
315 West Huron Street, Suite 140A
Ann Arbor, Michigan 48103

http://www.redhat.com/en/technologies/storage

tel.  734-707-0660
fax.  734-769-8938
cel.  734-216-5309

Re: civetweb, sts work

[Marcus Watts]

On Wed, Mar 09, 2016 at 10:32:09AM -0500, Matt Benjamin wrote:
> Hi Marcus,
> 
> Can you pull up wip-rgw-sts-3 (it's last commit from upstream looks to be Feb. 26).
> 
> Thanks!
> 
> Matt

Sure, I can rebase it on the yet later master.

(It *was* the latest master when I started pulling things up.)

				-Marcus