setup: the .tac files created by create_node.py call pkg_resources.require() so that they can load tahoe and twisted packages which were installed with setuptools multi-version mode
Also the create_node.py script itself uses pkg_resources.require() for the same reason.
Brian Warner [Wed, 29 Oct 2008 00:56:18 +0000 (17:56 -0700)]
#527: respond to GETs with early ranges quickly, without waiting for the whole file to download. Fixes the alacrity problems with the earlier code. Still needs cache expiration.
Brian Warner [Tue, 28 Oct 2008 20:41:04 +0000 (13:41 -0700)]
#527: support HTTP 'Range:' requests, using a cachefile. Adds filenode.read(consumer, offset, size) method. Still needs: cache expiration, reduced alacrity.
Brian Warner [Tue, 28 Oct 2008 20:36:46 +0000 (13:36 -0700)]
iputil.py: avoid a DNS lookup at startup (which may timeout tests when run on a partially-offline host) by using 198.41.0.4 instead of A.ROOT-SERVERS.NET
Brian Warner [Fri, 24 Oct 2008 18:52:48 +0000 (11:52 -0700)]
storage: add remote_advise_corrupt_share, for clients to tell storage servers about share corruption that they've discovered. The server logs the report.
Brian Warner [Fri, 24 Oct 2008 04:00:24 +0000 (21:00 -0700)]
mutable/servermap.py: fix needs_merge(), it was incorrectly claiming that mixed shares with distinct seqnums needed a merge, causing repair(force=False) to fail
setup: bundle an SVN snapshot of setuptools instead of the most recent stable release of setuptools
This SVN snapshot fixes a problem that prevents the setting up of nevow:
http://bugs.python.org/setuptools/issue20
setup: reorder dependencies to be sort of increasing order of how much they depend on other stuff
Not that the order makes any different to how it gets installed, as far as I can tell.
Brian Warner [Wed, 22 Oct 2008 08:42:37 +0000 (01:42 -0700)]
immutable/filenode.py: add TODO note about the #514 monitor to check(), rather than going through the checker/verifier code and adding it, since Zooko is currently working on that code
Brian Warner [Wed, 22 Oct 2008 00:03:07 +0000 (17:03 -0700)]
Change deep-size/stats/check/manifest to a start+poll model instead of a single long-running synchronous operation. No cancel or handle-expiration yet. #514.
setup: remove the developer note about doing without GNU make (the GNU make requirement is about to hurt Peter if he tries to follow this doc, by the way)
add classifiers showing with which versions of Python it is known to work.
robk-tahoe [Tue, 21 Oct 2008 00:29:02 +0000 (17:29 -0700)]
* fuse/runtests: added --catch-up-pause option
On linux, write tests are failing because data written to fuse isn't showing
up in tahoe by the time it's checked. it's not clear where this is originating,
since the fuse implementation [should be] waiting for completion of tahoe
operations before returning from its calls. This adds an option to control the
duration of a pause between the fuse write and the check of tahoe, which is by
default set to 2s on linux, which - somewhat inexplicably - seems to 'fix' the
problem, in as far as it allows tests to complete.
previously, upon opening a file for reading, the open() call would block
while the entire file was retrieved from tahoe into the cache directory.
This change adds a DownloaderWithReadQueue class, and associated plumbing,
such that an open() will return promptly with the download initiated 'in
the background'. Subsequent read() operations will block until enough
data has been downloaded to satisfy that request. This provides a behaviour
similar to streaming, i.e. the client application will be able to read
data from the fuse interface while the remainder of the file is still being
downloaded.
robk-tahoe [Mon, 20 Oct 2008 23:24:27 +0000 (16:24 -0700)]
fuse/runtests: added 'read_in_random_order' test
this test uploads a test file to tahoe, and then reads the file from fuse,
but reads the blocks of the file in a random order; this is designed to
exercise the asynchronous download feature of blackmatch - where the file
is downloaded from tahoe asynchronously, and rather than blocking open()
for the entirety of the download, instead individual read() calls are
blocked until enough of the file has been downloaded to satisfy them