]>
git.rkrishnan.org Git - tahoe-lafs/tahoe-lafs.git/log
Brian Warner [Tue, 16 Oct 2007 19:25:09 +0000 (12:25 -0700)]
add an equally-simple file-verifier
Brian Warner [Tue, 16 Oct 2007 18:00:29 +0000 (11:00 -0700)]
encode.py: update comments, max_segment_size is now 2MiB
Brian Warner [Mon, 15 Oct 2007 23:16:39 +0000 (16:16 -0700)]
add a simple checker, for both files and directories
Zooko O'Whielacronx [Sat, 27 Oct 2007 01:53:21 +0000 (18:53 -0700)]
shush, pyflakes
Zooko O'Whielacronx [Sat, 27 Oct 2007 01:37:40 +0000 (18:37 -0700)]
improve code coverage by removing untested raise RuntimeError and replacing it with tested assert
Zooko O'Whielacronx [Sat, 27 Oct 2007 01:30:44 +0000 (18:30 -0700)]
cli: use urllib.quote() on vdrive-path arguments before passing them through HTTP
Zooko O'Whielacronx [Tue, 23 Oct 2007 00:50:52 +0000 (17:50 -0700)]
update docs about webport (fixes #185)
czooko [Mon, 22 Oct 2007 23:56:30 +0000 (16:56 -0700)]
make the anchor text in the WUI be descriptive of the target -- not "Click Here!"!
czooko [Mon, 22 Oct 2007 23:55:20 +0000 (16:55 -0700)]
trivial change to comment -- sentences start with a capital letter and end with a punctuation mark
I've been volunteering in Irby's first-grade class, and you do NOT get a star
on your Power Sentences worksheet unless you follow this rule.
czooko [Mon, 22 Oct 2007 23:52:55 +0000 (16:52 -0700)]
patch the LogObserver in a more modern, forward-compatible way and update the in-line comments about it
czooko [Tue, 16 Oct 2007 04:35:09 +0000 (21:35 -0700)]
doc: emphasize in our "description" field that we are under a Free Software licence
czooko [Tue, 16 Oct 2007 04:33:25 +0000 (21:33 -0700)]
relnotes.txt: fix bug in relnotes.txt for v0.6.1 -- it incorrectly described v0.6 as having been released in August; it was actually September
Zooko O'Whielacronx [Tue, 16 Oct 2007 04:15:56 +0000 (21:15 -0700)]
TAG allmydata-tahoe-0.6.1
czooko [Tue, 16 Oct 2007 03:55:10 +0000 (20:55 -0700)]
fix the 'license' field of the PyPI db (not the Trove Classifiers that I was changing in recent patches) to describe our licence and link to it
The earlier patches were changing the Trove Classifiers, which is a different thing to this 'license' field.
Zooko O'Whielacronx [Tue, 16 Oct 2007 03:48:09 +0000 (20:48 -0700)]
comment-out the 'license' field because PyPI rejects upload if you have an invalid value therein
This means that we have no machine-readable licence for now. I will make the
human-readable licensing.
Zooko O'Whielacronx [Tue, 16 Oct 2007 03:07:42 +0000 (20:07 -0700)]
tests: make test_encode specify the erasure coding params it wants instead of expecting the defaults to be what it wants
Zooko O'Whielacronx [Tue, 16 Oct 2007 02:57:42 +0000 (19:57 -0700)]
setup: stop claiming that we are under GPL in the "license" field of the PyPI database
Unfortunately, there is no way to claim that we are under a Free Software/Open
Source licence without also claiming to be under a licence that we are not or
claiming to have approval from DFSG or OSI, which we haven't.
Until now, I erred on the side of choosing the licence that is closest to our
from the list (GPL), but that was a bad idea and now I'm erring on the side of
not including a machine-readable licensing claim at all.
Hopefully humans who are interested will quickly find out that we are actually
under a Real Free Software Licence.
But really, this underscores that we need to talk to FSF, edit our licence for
clarity of intent, and submit it to DFSG/OSI.
Zooko O'Whielacronx [Tue, 16 Oct 2007 02:53:59 +0000 (19:53 -0700)]
update a few documents, comments, and defaults to mention 3-of-10 instead of 25-of-100
Brian Warner [Mon, 15 Oct 2007 22:01:59 +0000 (15:01 -0700)]
Makefile: define TRIALCMD with '=' not ':=', to fix make-clean test. Closes #180
Zooko O'Whielacronx [Mon, 15 Oct 2007 21:56:02 +0000 (14:56 -0700)]
relnote.txt: update relnotes.txt for the v0.6.1 release
Zooko O'Whielacronx [Mon, 15 Oct 2007 21:55:53 +0000 (14:55 -0700)]
README: give a link to the TestGrid page
Zooko O'Whielacronx [Mon, 15 Oct 2007 19:29:02 +0000 (12:29 -0700)]
webapi.txt: update webapi.txt to reflect the security fix from #98
Zooko O'Whielacronx [Mon, 15 Oct 2007 18:52:26 +0000 (11:52 -0700)]
setup: use the new find_exe module to find trial
Zooko O'Whielacronx [Mon, 15 Oct 2007 17:25:15 +0000 (10:25 -0700)]
startstop_node: use the new find_exe module
Zooko O'Whielacronx [Mon, 15 Oct 2007 17:25:04 +0000 (10:25 -0700)]
setup: generalize the kludge of finding an executable (i.e. trial or twistd) when there might be only a .py script version of it available
Zooko O'Whielacronx [Mon, 15 Oct 2007 16:08:41 +0000 (09:08 -0700)]
setup: split off README.win32 from README and paste in Mike Booker's notes about building OpenSSL
Zooko O'Whielacronx [Mon, 15 Oct 2007 15:32:21 +0000 (08:32 -0700)]
remove unused imports (thanks, pyflakes)
Zooko O'Whielacronx [Mon, 15 Oct 2007 04:22:01 +0000 (21:22 -0700)]
setup: setuptools_darcs_plugin is boring
Zooko O'Whielacronx [Mon, 15 Oct 2007 03:57:59 +0000 (20:57 -0700)]
remove a file accidentally left over from the experiment in using sqlite to maintain lease information
Zooko O'Whielacronx [Mon, 15 Oct 2007 03:46:51 +0000 (20:46 -0700)]
a slightly nicer method of computing our timestamp format
Zooko O'Whielacronx [Mon, 15 Oct 2007 03:43:11 +0000 (20:43 -0700)]
install our custom timestamp formats in a less disruptive way
The unit tests on Windows fail because trial is attempting to remove its own
log observer during teardown. This patch customizes the extant log observer
object by replacing its formatTime method with our own.
I first tried the approach of storing their log observer object and putting it
back during teardown, but it didn't work (perhaps because our node object
doesn't get a chance to do its deferred stopService behavior in time), and
anyway I generally prefer the "fail-safe", or "crash-only" design.
Zooko O'Whielacronx [Mon, 15 Oct 2007 03:03:32 +0000 (20:03 -0700)]
setup: remove an explanation from the top of the Makefile now that the topic is addresses in the README
Zooko O'Whielacronx [Mon, 15 Oct 2007 02:13:12 +0000 (19:13 -0700)]
CREDITS: Nathan Wilcox ++
Brian Warner [Sat, 13 Oct 2007 23:06:39 +0000 (16:06 -0700)]
test_runner.RunNode: pass an explicit webport, to avoid using 8123 (which might be in used by a running node). Closes #175.
Zooko O'Whielacronx [Sat, 13 Oct 2007 20:38:18 +0000 (13:38 -0700)]
setup: setup_requires setuptools_darcs_plugin. Without it the "./setup.py sdist upload" will silently upload the wrong package contents.
Zooko O'Whielacronx [Sat, 13 Oct 2007 07:38:16 +0000 (00:38 -0700)]
filter out "0.0.0.0" from detected IP addresses
Brian Warner [Fri, 12 Oct 2007 05:29:23 +0000 (22:29 -0700)]
cli: simplify code by using stdlib's httplib module
Brian Warner [Fri, 12 Oct 2007 03:31:48 +0000 (20:31 -0700)]
cli: implement 'mv'. Closes #162.
Brian Warner [Fri, 12 Oct 2007 02:47:40 +0000 (19:47 -0700)]
tahoe_get.py: remove unused import
Brian Warner [Fri, 12 Oct 2007 02:20:41 +0000 (19:20 -0700)]
cli: add test coverage
Brian Warner [Fri, 12 Oct 2007 00:30:07 +0000 (17:30 -0700)]
node.py: fix timestamps (add ms and Z) by replacing the FileLogObserver. #171.
Zooko O'Whielacronx [Sat, 13 Oct 2007 06:21:53 +0000 (23:21 -0700)]
README: edit to clarify that you can't use "make" if you installed it the easy_install way
Zooko O'Whielacronx [Sat, 13 Oct 2007 05:59:37 +0000 (22:59 -0700)]
setup: make ez_setup.py work to upgrade setuptools even if there is already a setuptools installed which is too old
This works only if setup.py is invoked as "./setup.py install" (or
"python ./setup.py install" or whatever). It doesn't work if it is invoked by
easy_install. On the other hand, I don't know why easy_install would execute
ez_setup.py anyway -- I thought that it didn't execute the setup.py scripts.
See this mailing list thread for details:
http://mail.python.org/pipermail/distutils-sig/2007-October/008339.html
Zooko O'Whielacronx [Thu, 11 Oct 2007 23:27:52 +0000 (16:27 -0700)]
setup: upgrade zfec from 1.0.4 to 1.1
Zooko O'Whielacronx [Thu, 11 Oct 2007 23:12:13 +0000 (16:12 -0700)]
change another example to use port 8123
Brian Warner [Thu, 11 Oct 2007 22:37:52 +0000 (15:37 -0700)]
cli: fix usage to refer to 'tahoe', not 'allmydata'. Closes #154.
Brian Warner [Thu, 11 Oct 2007 21:55:23 +0000 (14:55 -0700)]
add public testnet .furls to docs/testnet/, and copy into .deb . Closes #157.
Brian Warner [Thu, 11 Oct 2007 21:37:29 +0000 (14:37 -0700)]
deb: add docs/* to the debian package
Brian Warner [Thu, 11 Oct 2007 20:19:11 +0000 (13:19 -0700)]
docs/webapi.txt: mention that we default to a --webport of 8123
Brian Warner [Thu, 11 Oct 2007 20:17:33 +0000 (13:17 -0700)]
docs: change our default HTTP port to 8123
Brian Warner [Thu, 11 Oct 2007 20:17:23 +0000 (13:17 -0700)]
change our default HTTP port to 8123
Brian Warner [Thu, 11 Oct 2007 10:39:29 +0000 (03:39 -0700)]
bin/tahoe: rename 'allmydata-tahoe' in some comments
Brian Warner [Thu, 11 Oct 2007 10:38:24 +0000 (03:38 -0700)]
rename bin/allmydata-tahoe to bin/tahoe. Closes #155.
Brian Warner [Thu, 11 Oct 2007 09:48:06 +0000 (02:48 -0700)]
rename client.tac to tahoe-client.tac, so that 'ps ax|grep tahoe' works. Closes #156.
Brian Warner [Thu, 11 Oct 2007 09:24:17 +0000 (02:24 -0700)]
node.py: don't append 'Z' to the timestamp, since it's really localtime. We need deeper changes to make it be UTC
Brian Warner [Thu, 11 Oct 2007 09:20:20 +0000 (02:20 -0700)]
check_memory.py: don't use self.keepalive_file until it's been initialized
Brian Warner [Thu, 11 Oct 2007 09:19:59 +0000 (02:19 -0700)]
create_node.py: allow config['webport'] to be missing, for check_memory
Brian Warner [Thu, 11 Oct 2007 09:13:05 +0000 (02:13 -0700)]
node.py: set logging timestamp to '2007-10-11 02:11:14.000Z', per ticket #171. No milliseconds yet, though
Brian Warner [Thu, 11 Oct 2007 09:01:23 +0000 (02:01 -0700)]
create_node: use a webport by default, on localhost:8011
Brian Warner [Thu, 11 Oct 2007 08:55:29 +0000 (01:55 -0700)]
test_cli.py: hush pyflakes with a dummy usage, until we get some real CLI tests
Brian Warner [Thu, 11 Oct 2007 08:54:23 +0000 (01:54 -0700)]
runner: make most commands use ~/.tahoe by default (create-client, start/stop/restart, all CLI tools, but *not* create-introducer
Brian Warner [Thu, 11 Oct 2007 08:38:04 +0000 (01:38 -0700)]
webish: write node.url, for the benefit of CLI tools
Brian Warner [Thu, 11 Oct 2007 08:34:44 +0000 (01:34 -0700)]
cli: use urllib.escape on all URIs
Brian Warner [Thu, 11 Oct 2007 07:30:36 +0000 (00:30 -0700)]
cli: add --node-directory and --root-uri to all commands
Brian Warner [Thu, 11 Oct 2007 13:58:08 +0000 (06:58 -0700)]
webapi.txt: reinstate documentation of the unpleasant URI-escaping needed for
slashes in dirnode URIs, to be resolved some day by #102.
Brian Warner [Thu, 11 Oct 2007 03:10:03 +0000 (20:10 -0700)]
'tahoe restart': re-enable --force option, which was accidentally lost
Zooko O'Whielacronx [Thu, 4 Oct 2007 21:14:48 +0000 (14:14 -0700)]
setup: we needn't depend on make-version targets because setup.py always attempts to make a version whenever it is executed
Zooko O'Whielacronx [Thu, 4 Oct 2007 21:11:16 +0000 (14:11 -0700)]
setup: attempt to work-around the problem that paths might end with trailing back-slashes (on Windows) by appending a PATHSEP (i.e. ":" or ";") instead of an OSSEP (i.e. "/" or "\")
I don't know what will happen if the path ends up with something like
"C:\Programs and Files\Whatever\;" on Windows, and then that gets passed to
cygwin bash. This reminds me of Brian's suggestion to use Python helper
scripts (c.f. misc/find-dep-eggs.py) instead of writing this stuff in the
GNUmake language. And *that* reminds me of the idea of writing the whole damn
thing in Python instead of in GNUmake, i.e. make all of our build tools be
plugins for setuptools instead of being GNUmake targets.
Zooko O'Whielacronx [Thu, 4 Oct 2007 20:09:20 +0000 (13:09 -0700)]
setup: patch to fix bug in our latest ez_setup.py if pkg_resources can't be imported
Zooko O'Whielacronx [Thu, 4 Oct 2007 19:37:21 +0000 (12:37 -0700)]
README: add note about how to build on Windows-native using gcc, and a bit of editing
Zooko O'Whielacronx [Thu, 4 Oct 2007 18:23:20 +0000 (11:23 -0700)]
setup: bundle simplejson-1.7.3 in misc/dependencies
Zooko O'Whielacronx [Thu, 4 Oct 2007 18:18:46 +0000 (11:18 -0700)]
setup: import bugfix to ez_setup.py
Zooko O'Whielacronx [Thu, 4 Oct 2007 18:07:46 +0000 (11:07 -0700)]
more information SkipTest for XMLRPC
Zooko O'Whielacronx [Thu, 4 Oct 2007 15:12:53 +0000 (08:12 -0700)]
README: OpenSSL is bundled with pyOpenSSL on Windows-native
Zooko O'Whielacronx [Wed, 3 Oct 2007 22:28:37 +0000 (15:28 -0700)]
setup: remove MANIFEST.in in favor of the new setuptools_darcs_plugin
MANIFEST.in is a script where you express all the files that you want to be
included (on top of the ones that setuptools figured out for itself).
setuptools_darcs_plugin is a setuptools plugin that makes all files which are
under darcs revision control get included.
Zooko O'Whielacronx [Wed, 3 Oct 2007 22:14:14 +0000 (15:14 -0700)]
setup: use ez_setup.py without a "download delay"
Zooko O'Whielacronx [Wed, 3 Oct 2007 22:13:19 +0000 (15:13 -0700)]
setup: import the latest version of ez_setup.py with my patches
Zooko O'Whielacronx [Tue, 2 Oct 2007 20:19:07 +0000 (13:19 -0700)]
README: refactor README
fixes #163
hopefully fixes #148 -- but somebody else will have to try it (maybe Mike Booker) to find out!
Brian Warner [Thu, 27 Sep 2007 01:24:51 +0000 (18:24 -0700)]
upgrade to foolscap-0.1.7
Brian Warner [Thu, 27 Sep 2007 01:16:15 +0000 (18:16 -0700)]
check_speed: average multiple pings when measuring RTT
Brian Warner [Thu, 27 Sep 2007 01:15:13 +0000 (18:15 -0700)]
debian: oops, set debian/changelog version to 0.6.0-1+, to be less than everything the buildbot is creating
Brian Warner [Thu, 27 Sep 2007 00:58:42 +0000 (17:58 -0700)]
debian: put a version string of '0.6.0+' in the debian changelog
Brian Warner [Thu, 27 Sep 2007 00:55:24 +0000 (17:55 -0700)]
debian/sid: copy rules from feisty, the same can be used for both
Brian Warner [Thu, 27 Sep 2007 00:52:30 +0000 (17:52 -0700)]
debian/feisty: use our original bin/allmydata-tahoe instead of setuptools's
because the setuptools "entry points" form asserts that there are
setuptools-visible packages like nevow/zope.interface (i.e. they have .egg-info
metadata). Until very recently, most debian systems did not install this
metadata. Instead, we rely upon the usual debian dependency checking as
expressed in debian/control .
Zooko O'Whielacronx [Sat, 29 Sep 2007 18:05:25 +0000 (11:05 -0700)]
README: add dependency on OpenSSL
Zooko O'Whielacronx [Sat, 29 Sep 2007 18:05:20 +0000 (11:05 -0700)]
README: edits
Zooko O'Whielacronx [Thu, 27 Sep 2007 22:06:17 +0000 (15:06 -0700)]
setup: don't include zope.interface in our automatically-satisfiable dependencies for now
Brian Warner [Wed, 26 Sep 2007 22:00:59 +0000 (15:00 -0700)]
dump-share: clarify the label on the size of the original file
Brian Warner [Wed, 26 Sep 2007 20:07:33 +0000 (13:07 -0700)]
check_speed: measure RTT, report per-file times as a multiple of RTT
Brian Warner [Wed, 26 Sep 2007 19:21:15 +0000 (12:21 -0700)]
control: add measure_peer_response_time(), to estimate RTT for the mesh
Brian Warner [Wed, 26 Sep 2007 19:20:48 +0000 (12:20 -0700)]
client.py: add the missing remote_get_nodeid() method claimed in interfaces.py
Brian Warner [Wed, 26 Sep 2007 19:07:37 +0000 (12:07 -0700)]
test_system.py minor typo
Brian Warner [Wed, 26 Sep 2007 19:06:55 +0000 (12:06 -0700)]
test_system.py: add coverage for allmydata.control
Brian Warner [Wed, 26 Sep 2007 04:03:54 +0000 (21:03 -0700)]
test_client.py: add TODO to test the hotline-timeout file
Brian Warner [Wed, 26 Sep 2007 03:07:03 +0000 (20:07 -0700)]
check_speed.py: minor comment
Brian Warner [Wed, 26 Sep 2007 02:22:33 +0000 (19:22 -0700)]
client.py: increase hotline timeout, the check_memory test is failing
Brian Warner [Wed, 26 Sep 2007 01:57:36 +0000 (18:57 -0700)]
check_speed.py: use more small-file tests to improve accuracy of per-file time
Brian Warner [Wed, 26 Sep 2007 01:47:48 +0000 (18:47 -0700)]
check_memory: more keepalive-file debug messages
Brian Warner [Wed, 26 Sep 2007 01:26:54 +0000 (18:26 -0700)]
check_memory: put framework log in a stable place, add debug code to figure
out why the keepalive-file isn't doing its job
Brian Warner [Tue, 25 Sep 2007 01:12:37 +0000 (18:12 -0700)]
test_system.py: do one upload, then test debug scripts, then do other uploads