openstack/swift - swift - OpenDev: Free Software Needs Free Tools

In the interest of keeping the core Swift code repository less
complex, but still offering a quick way to find associated projects
that enhance or use Swift, I've added this new Associated Projects
page prominently to the Swift documentation.
This will allow much less barrier to entry on enhancing Swift as
other projects can work independently and will only need to wait on
the core Swift project for approval of minimal tweaks to the core
Swift code base.
It will also allow an easy central place to find cool associated
projects that otherwise might go unnoticed or even duplicated.
The plan is to move non-essential projects that are currently
embedded in the Swift repository out into their own projects with
links to them on this new page. This would include items such as
(just what I can think of right now): bin/swift command line tool and
clients, swift-bench, swift-dispersion, TempURL, FormPost, StaticWeb,
Rate Limiting, Swift3, Domain Remap, and CNAME Lookup.
After all that is done, those projects will be able to move forward
much more quickly and new developers for Swift itself will have much
less to learn and get confused about.
Change-Id: Ib8447d8bd480f0a3d8f0413137ccdba73a11dd91

`swift-ring-builder X.builder [add|set_weight|set_info|remove]` can
now work on multiple devices at once.
This allows for faster bulk operations; for example, if an
administrator has to add 48 new devices to the ring, then this lets
them do it in one command invocation. Since each invocation of
swift-ring-builder follows a deserialize-operate-serialize pattern,
using the bulk form of the commands on N devices will save N-1
[de]serializations.
Change-Id: Ibf77528f28026866af3c94d0370181c8013d5bc5

bug 701248
bug 819303
This change makes syslog optional, by providing a unit test
config with the ability to cause test/unit/__init__.py to
replace SysLogHandler with a fake logging class. The
default behavior is unchanged.
FakeLogger now inherits directly from Handler and mocks out
its API - this was a backward-compatibility issue.
Change-Id: I653dec148c4f6cf81759de03964c6a3290c1a290

The listing of containers was inclusive whereas the listing of
objects was exclusive. I searched and searched and only found
end_marker usage with listings of objects so I changed the listings
of containers to match that.
Change-Id: Ia8eb66a7ceaef28c698e2d4e19dd5dfd9197a911

Change-Id: I41356ee250c9088a2387b0d493586dd990a04ac3

Change-Id: Idf2ea227b0078026eb2693ff7c0a24ec926bb6f3

On a relatively recent build of swift, all my scripts blew up with
"No tenant specified". It was caused by the fix to add --os_tenant_name,
commit ID: 208b8e85a8
This patch restores the old behavior. I tested it to work with old
swauth-based Swift, new Keystone-based Swift, in the latter case using
both -U and --os_tenant_name arguments.
Note that this patch permits to use a literal colon in the user name
with the new syntax, as long as tenant is specified. Empty tenant
names are not allowed with either syntax.
Change-Id: I7785e6981a9d6281d0421c43875ee19d61d5ff43
Bug: 982909

Replace HTTP status code with constants like HTTP_OK, HTTP_NOT_FOUND, etc.
Change-Id: I76cd8bf880e7269454192a3bc4b027a297741247

Change-Id: I78dbb390cf0afe03d4bcf1aa9ffc9497392579dd

Change-Id: I766b2da0a3769869b5b6c5565106e9032fea8c3e

In the side bar, links to past versions and docs.openstack.org
On the index.rst, links to wiki and docs.openstack.org
Change-Id: Icf33c6f396e1ab016fd86a56e61df3e063a1bae2

This makes it a lot easier for CI jobs to be setup and return the true
status of the testsuite.
Change-Id: I0c2b0d6b17f560a889cd2f22ab98a68361f6d738

Fixes bug 984042
Change-Id: I18a6eac497b24338af4d82a568484b0fc9d97382

The test was failing silently. To catch these errors the response
from controller.PUT() is now checked.
Fixes to the TestObjectController.test_PUT_auto_content_type():
* Check for the correct 'Content-Type' header in the fake connect() method
* Pass in the correct number of responses to fake_http_connect()
* Pass a 'transfer-encoding' header to PUT() so that
 constraints.check_object_creation() passes
* Pass the correct number of elements in the 'expected' iter to the
 test_content_type() method
Change-Id: I8dced7b2f3c7e1b1d292c6995b1faf46f5e863fe

Change-Id: I31a64614231391b52855b88220319b2359bea142

Change-Id: I16a8d1f5543d62a6d968bf1b9f52f319902437cb

Change-Id: I974f729da60e5ab9453daf9e52466b3e1af5c69b

Change-Id: Id68231ced070b27d1f7fb5a230d345694632e7dd

Corrected its/it's mistakes, harmonized line wrapping within some docs
and clarified doc wording in several places.
Change-Id: Ib9ac6d5e859f770a702e1fad6de8d4abe0390b47

Change-Id: Iaf5f7011cf41f503b7ccdea671cd31a0466f5248

Admittedly, this is ridiculously trivial, but the coloring of
"hash" always peeved me. Not so much that it existed, but that
I did not understand what caused it. After consulting our Sphinx
experts, it turned out that the code block was considered code
in Python and "hash" is a reserved word in Python, so it was
colored.
One option is to declare the code block "ini", which is what
it is, but that causes the block stand out just as badly, unless
we highlight every one of them. So, for now use "none".
Change-Id: Icee021fc30839eeb02cd3e232ff6be92ed32df52

In the process of wrapping for profiling, a trivial wrap with def main()
ended in a ton of these:
 NameError: global name 'parser' is not defined
A closer investigation revealed an apparent confusion in argument names.
The problem was introduced by commit 985968f765
of November 2010. Note that Greg fixed st_delete() correctly in that one,
but apparently forgot the others.
Unit tests pass for both current (bad) code, and the patched code.
By the magic of Python scoping, current code works, just impossible
to wrap for a quick hack.
Change-Id: I18f2494ce2a201d04d49f99ea259d141939dcb6a

Somehow I forgot the account and container replicators with the last
db_preallocation patch.
Change-Id: I1a42ef000d60b59aedb113d91485319d6919bc36

Change-Id: Idb84bd6800c047ce1080039434b7957aa7d69b4d

If you do specify --os_auth_url rather than using the environment variable
OS_AUTH_URL, you'll end up with different behaviour.
Change-Id: I7430c2a619487375427b1c377ce00245eff107ff

Change-Id: I5a837bea8699b78484265dc2434b190d0cf15de2

Change-Id: I1c571951f25a6e724cdd4699eb94baad5e47eb95

Change-Id: I6e58eed72aa3d1835b1a9ef99014d27a7dfd810e