]> git.rkrishnan.org Git - tahoe-lafs/tahoe-lafs.git/log
tahoe-lafs/tahoe-lafs.git
17 years agoadd comment
Zooko O'Whielacronx [Wed, 24 Jan 2007 22:21:45 +0000 (15:21 -0700)]
add comment

17 years agonicer API -- you don't have to shuffle the shares into place before calling decode_all()
Zooko O'Whielacronx [Wed, 24 Jan 2007 22:20:53 +0000 (15:20 -0700)]
nicer API -- you don't have to shuffle the shares into place before calling decode_all()
Instead it does that shuffling in C inside fecmodule.

17 years agostricter typing -- using unsigned char for indexes into shares
Zooko O'Whielacronx [Wed, 24 Jan 2007 22:17:51 +0000 (15:17 -0700)]
stricter typing -- using unsigned char for indexes into shares

17 years agohush pyflakes
Brian Warner [Wed, 24 Jan 2007 22:29:22 +0000 (15:29 -0700)]
hush pyflakes

17 years agofiletree: make delete() work
Brian Warner [Wed, 24 Jan 2007 22:10:53 +0000 (15:10 -0700)]
filetree: make delete() work

17 years agopyfec v0.9
Zooko O'Whielacronx [Tue, 23 Jan 2007 00:17:31 +0000 (17:17 -0700)]
pyfec v0.9

Here is the change history from the first darcs era, in reverse chronological
order:

Mon Jan 22 16:12:56 MST 2007  "Zooko O'Whielacronx <zooko@zooko.com>"
  * move everything into a subdirectory so that I can merge this darcs repo with the tahoe darcs repo

     ./fec -> ./pyfec/fec
     ./setup.py -> ./pyfec/setup.py
    A ./pyfec/

Mon Jan 22 16:10:17 MST 2007  "Zooko O'Whielacronx <zooko@zooko.com>"
  * clean up and minimize fec.c
   * strip out unused code
   * hard-code GF_BITS to 8
   * reindent and reformat curly bracket placement

    M ./fec/fec.c -655 +324
    M ./fec/fec.h -25

Mon Jan 22 14:24:32 MST 2007  "Zooko O'Whielacronx <zooko@zooko.com>"
  * change API to allow a subset of the shares to be produced, and to just pass back pointers to primary shares instead of copying them

    M ./fec/fec.c -24 +40
    M ./fec/fec.h -5 +17
    M ./fec/fecmodule.c -63 +144
    M ./fec/test/test_pyfec.py -16 +25
    M ./setup.py -2 +27

Tue Jan 16 23:01:44 MST 2007  "Zooko O'Whielacronx <zooko@zooko.com>"
  * split encoder from decoder

    M ./fec/fecmodule.c -48 +161
    M ./fec/test/test_pyfec.py -3 +4

Tue Jan 16 14:35:25 MST 2007  "Zooko O'Whielacronx <zooko@zooko.com>"
  * it compiles now!

     ./fec.c -> ./pyfec/fec.c
     ./fec.h -> ./pyfec/fec.h
     ./fecmodule.c -> ./pyfec/fecmodule.c
     ./pyfec -> ./fec
    M ./fec/fec.c -109 +85 r13
    M ./fec/fec.h -3 +2 r13
    M ./fec/fecmodule.c -23 +241 r13
    A ./fec/test/
    A ./fec/test/test_pyfec.py
    A ./pyfec/
    A ./setup.py

Tue Jan  9 10:47:58 MST 2007  zooko@zooko.com
  * start of new fecmodule.c

    A ./fecmodule.c

Mon Jan  1 15:00:04 MST 2007  zooko@zooko.com
  * tidy up error handling

    M ./fec.c -26 +16

Mon Jan  1 14:06:30 MST 2007  zooko@zooko.com
  * remove the on-the-fly encoding option
  We don't currently need it.

    M ./fec.c -68
    M ./fec.h -22

Mon Jan  1 13:53:28 MST 2007  zooko@zooko.com
  * original import from Mnet project

    A ./fec.c
    A ./fec.h

17 years agoMakefile: add target to count lines of code and TODOs
Brian Warner [Mon, 22 Jan 2007 08:17:39 +0000 (01:17 -0700)]
Makefile: add target to count lines of code and TODOs

17 years agofiletree: change the way addpath works, now we add workqueue steps for all involved...
Brian Warner [Mon, 22 Jan 2007 08:06:09 +0000 (01:06 -0700)]
filetree: change the way addpath works, now we add workqueue steps for all involved subtrees at about the same time, rather than letting one step add the next when it runs. Finally add a (passing) test for uploading files to CHK-based directories

17 years agofiletree: add vdrive upload/download test, change workqueue relative-filename semantics
Brian Warner [Sun, 21 Jan 2007 23:03:15 +0000 (16:03 -0700)]
filetree: add vdrive upload/download test, change workqueue relative-filename semantics

17 years agoqueen.sendOnly: ignore DeadReferenceError too
Brian Warner [Sun, 21 Jan 2007 23:01:09 +0000 (16:01 -0700)]
queen.sendOnly: ignore DeadReferenceError too

17 years agomove IWorkQueue into allmydata.interfaces, give VirtualDrive an uploader
Brian Warner [Sun, 21 Jan 2007 22:15:31 +0000 (15:15 -0700)]
move IWorkQueue into allmydata.interfaces, give VirtualDrive an uploader

17 years agomove upload/download interfaces to allmydata.interfaces, let SubTreeMaker assert...
Brian Warner [Sun, 21 Jan 2007 22:01:34 +0000 (15:01 -0700)]
move upload/download interfaces to allmydata.interfaces, let SubTreeMaker assert IDownloader-ness of its 'downloader' argument

17 years agotest_filetree: more vdrive._get_file_uri() coverage
Brian Warner [Sun, 21 Jan 2007 20:39:40 +0000 (13:39 -0700)]
test_filetree: more vdrive._get_file_uri() coverage

17 years agofiletree: mark leaf nodes by adding is_leaf_subtree(), stop traversing when we hit...
Brian Warner [Sun, 21 Jan 2007 20:31:16 +0000 (13:31 -0700)]
filetree: mark leaf nodes by adding is_leaf_subtree(), stop traversing when we hit them, to make vdrive._get_file_uri() work

17 years agofiletree: test NoSuchDirectoryError in vdrive.list()
Brian Warner [Sun, 21 Jan 2007 11:50:37 +0000 (04:50 -0700)]
filetree: test NoSuchDirectoryError in vdrive.list()

17 years agomore filetree, workqueue-boxes now hold serialized Nodes, move NodeMaker out to a...
Brian Warner [Sun, 21 Jan 2007 11:18:54 +0000 (04:18 -0700)]
more filetree, workqueue-boxes now hold serialized Nodes, move NodeMaker out to a separate module

17 years agofiletree: start testing IVirtualDrive, beginning with list()
Brian Warner [Sun, 21 Jan 2007 04:14:41 +0000 (21:14 -0700)]
filetree: start testing IVirtualDrive, beginning with list()

17 years agofiletree: put SubTreeMaker and NodeMaker in separate classes
Brian Warner [Sun, 21 Jan 2007 00:04:56 +0000 (17:04 -0700)]
filetree: put SubTreeMaker and NodeMaker in separate classes

17 years agomore filetree work, more tests now pass
Brian Warner [Sat, 20 Jan 2007 22:50:21 +0000 (15:50 -0700)]
more filetree work, more tests now pass

17 years agomore filetree, it's actually starting to make sense now
Brian Warner [Sat, 20 Jan 2007 21:41:51 +0000 (14:41 -0700)]
more filetree, it's actually starting to make sense now

17 years agofiletable: oops, os.listdir() does not guarantee sorted results
Brian Warner [Sat, 20 Jan 2007 20:02:26 +0000 (13:02 -0700)]
filetable: oops, os.listdir() does not guarantee sorted results

17 years agofiletree: more tests, still very early
Brian Warner [Sat, 20 Jan 2007 12:13:15 +0000 (05:13 -0700)]
filetree: more tests, still very early

17 years agofiletree: refactor INode serialization, start on tests
Brian Warner [Sat, 20 Jan 2007 11:52:53 +0000 (04:52 -0700)]
filetree: refactor INode serialization, start on tests

17 years agomore filetree hacking, still too early to test
Brian Warner [Sat, 20 Jan 2007 11:25:20 +0000 (04:25 -0700)]
more filetree hacking, still too early to test

17 years agocheckpont more filetree stuff
Brian Warner [Sat, 20 Jan 2007 06:22:39 +0000 (23:22 -0700)]
checkpont more filetree stuff

17 years agosnapshot filetree work: fix pyflakes complaints
Brian Warner [Fri, 19 Jan 2007 09:35:36 +0000 (02:35 -0700)]
snapshot filetree work: fix pyflakes complaints

17 years agosnapshot filetree work: it's getting close
Brian Warner [Fri, 19 Jan 2007 09:23:03 +0000 (02:23 -0700)]
snapshot filetree work: it's getting close

17 years agorename bin/allmydata to bin/allmydata-tahoe, so the package can co-exist with the...
Brian Warner [Sat, 20 Jan 2007 03:05:33 +0000 (20:05 -0700)]
rename bin/allmydata to bin/allmydata-tahoe, so the package can co-exist with the regular (amdlib) package

17 years agodownload.py: fix IDownloader to take a URI
Brian Warner [Fri, 19 Jan 2007 09:17:48 +0000 (02:17 -0700)]
download.py: fix IDownloader to take a URI

17 years agoupdate roadmap
Brian Warner [Fri, 19 Jan 2007 09:17:24 +0000 (02:17 -0700)]
update roadmap

17 years agofiletree.interfaces: remove not-really-code to improve coverage stats
Brian Warner [Thu, 18 Jan 2007 00:00:47 +0000 (17:00 -0700)]
filetree.interfaces: remove not-really-code to improve coverage stats

17 years agofigleaf.el: add code to auto-enable the annotation mode for all source files
Brian Warner [Thu, 18 Jan 2007 08:00:11 +0000 (01:00 -0700)]
figleaf.el: add code to auto-enable the annotation mode for all source files

17 years agoimprove test coverage a bit
Brian Warner [Wed, 17 Jan 2007 22:34:29 +0000 (15:34 -0700)]
improve test coverage a bit

17 years agowebish: present real URI in directory listing, not an unnecessarily b2a'ed form
Brian Warner [Wed, 17 Jan 2007 21:46:02 +0000 (14:46 -0700)]
webish: present real URI in directory listing, not an unnecessarily b2a'ed form

17 years agosplit filetree_new.py up into smaller pieces, in a new subpackage
Brian Warner [Wed, 17 Jan 2007 20:54:38 +0000 (13:54 -0700)]
split filetree_new.py up into smaller pieces, in a new subpackage

17 years agohush pyflakes warnings
Brian Warner [Wed, 17 Jan 2007 04:34:34 +0000 (21:34 -0700)]
hush pyflakes warnings

17 years agoupdate URI format, include codec name
Brian Warner [Wed, 17 Jan 2007 04:29:59 +0000 (21:29 -0700)]
update URI format, include codec name

17 years agoMakefile: use absolute paths when setting PYTHONPATH
Brian Warner [Wed, 17 Jan 2007 04:02:00 +0000 (21:02 -0700)]
Makefile: use absolute paths when setting PYTHONPATH

17 years agowebish: add PBURL to the all-peers table on the welcome page
Brian Warner [Wed, 17 Jan 2007 04:01:18 +0000 (21:01 -0700)]
webish: add PBURL to the all-peers table on the welcome page

17 years agoencode_new.py: recent Foolscap accepts 'None' as a constraint
Brian Warner [Wed, 17 Jan 2007 02:56:30 +0000 (19:56 -0700)]
encode_new.py: recent Foolscap accepts 'None' as a constraint

17 years agowebish: more verifierid-to-uri transition
Brian Warner [Wed, 17 Jan 2007 02:55:53 +0000 (19:55 -0700)]
webish: more verifierid-to-uri transition

17 years agovdrive/webish: finish verifierid/uri transition
Brian Warner [Wed, 17 Jan 2007 02:43:13 +0000 (19:43 -0700)]
vdrive/webish: finish verifierid/uri transition

17 years agoupload: fix typo in debug messages
Brian Warner [Wed, 17 Jan 2007 02:42:28 +0000 (19:42 -0700)]
upload: fix typo in debug messages

17 years agoupload: rearrange peer-selection code to be more readable, and fix a silly bug
Brian Warner [Wed, 17 Jan 2007 02:35:12 +0000 (19:35 -0700)]
upload: rearrange peer-selection code to be more readable, and fix a silly bug

17 years agooops, fix that NotEnoughPeersError instrumentation
Brian Warner [Wed, 17 Jan 2007 02:07:23 +0000 (19:07 -0700)]
oops, fix that NotEnoughPeersError instrumentation

17 years agoadd better error reporting to NotEnoughPeersError
Brian Warner [Wed, 17 Jan 2007 01:47:52 +0000 (18:47 -0700)]
add better error reporting to NotEnoughPeersError

17 years agochange upload to push 2 shares instead of 3
Brian Warner [Tue, 16 Jan 2007 22:15:26 +0000 (15:15 -0700)]
change upload to push 2 shares instead of 3
Now that peers can talk to themselves, the 5-node system test won't fail
just because one of the shares was hosted on the downloader (and thus
inaccessible until recently). The 3-share push was there to avoid this
problem.

17 years agonow that foolscap-0.0.7 accepts connect-to-self, allow peers to know about themselves...
Brian Warner [Tue, 16 Jan 2007 22:12:49 +0000 (15:12 -0700)]
now that foolscap-0.0.7 accepts connect-to-self, allow peers to know about themselves. We now require foolscap-0.0.7

17 years agohush pyflakes warnings
Brian Warner [Tue, 16 Jan 2007 04:31:10 +0000 (21:31 -0700)]
hush pyflakes warnings

17 years agorearrange encode/upload, add URIs, switch to ReplicatingEncoder
Brian Warner [Tue, 16 Jan 2007 04:22:22 +0000 (21:22 -0700)]
rearrange encode/upload, add URIs, switch to ReplicatingEncoder

Added metadata to the bucket store, which is used to hold the share number
(but the bucket doesn't know that, it just gets a string).

Modified the codec interfaces a bit.

Try to pass around URIs to/from download/upload instead of verifierids.
URI format is still in flux.

Change the current (primitive) file encoder to use a ReplicatingEncoder
because it provides ICodecEncoder. We will be moving to the (less primitive)
file encoder (currently in allmydata.encode_new) eventually, but for now
this change lets us test out PyRS or zooko's upcoming C-based RS codec in
something larger than a single unit test. This primitive file encoder only
uses a single segment, and has no merkle trees.

Also added allmydata.util.deferredutil for a DeferredList wrapper that
errbacks (but only when all component Deferreds have fired) if there were
any errors, which unfortunately is not a behavior available from the standard
DeferredList.

17 years agoadd simple metadata (a single string) to the storage protocol
Brian Warner [Mon, 15 Jan 2007 21:01:22 +0000 (14:01 -0700)]
add simple metadata (a single string) to the storage protocol

17 years agochange IEncoder to ICodecEncoder, to match the previous change
Brian Warner [Fri, 12 Jan 2007 03:57:14 +0000 (20:57 -0700)]
change IEncoder to ICodecEncoder, to match the previous change

17 years agouse the word 'codec' for erasure coding, for now. 'encode' is used for file-level...
Brian Warner [Fri, 12 Jan 2007 03:51:27 +0000 (20:51 -0700)]
use the word 'codec' for erasure coding, for now. 'encode' is used for file-level segmentation/hashing

17 years agosendOnly: oops, I keep forgetting that you can't really use f.trap in a lambda, becau...
Brian Warner [Wed, 10 Jan 2007 04:13:05 +0000 (21:13 -0700)]
sendOnly: oops, I keep forgetting that you can't really use f.trap in a lambda, because it returns the failure it trapped

17 years agotest_system.py: exercise queen.Roster._lost_node too
Brian Warner [Wed, 10 Jan 2007 02:40:36 +0000 (19:40 -0700)]
test_system.py: exercise queen.Roster._lost_node too

17 years agoworkqueue: more tests
Brian Warner [Tue, 9 Jan 2007 08:58:50 +0000 (01:58 -0700)]
workqueue: more tests

17 years agoworkqueue: more tests, coverage now at 63.4%, yay
Brian Warner [Tue, 9 Jan 2007 08:36:12 +0000 (01:36 -0700)]
workqueue: more tests, coverage now at 63.4%, yay

17 years agoworkqueue: more improvements, more tests
Brian Warner [Tue, 9 Jan 2007 05:29:42 +0000 (22:29 -0700)]
workqueue: more improvements, more tests

17 years agoworkqueue: start adding tests
Brian Warner [Tue, 9 Jan 2007 04:29:33 +0000 (21:29 -0700)]
workqueue: start adding tests

17 years agofigleaf.el: when converting to the elisp-readable format, use our in-tree version...
Brian Warner [Tue, 9 Jan 2007 04:29:03 +0000 (21:29 -0700)]
figleaf.el: when converting to the elisp-readable format, use our in-tree version of figleaf rather than one found on the system. Also change the keybinding to toggle annotations to C-cC-a, which is easier to type

17 years agofigleaf: delete coverage file (.figleaf) before each test run, otherwise I suspect...
Brian Warner [Tue, 9 Jan 2007 04:28:08 +0000 (21:28 -0700)]
figleaf: delete coverage file (.figleaf) before each test run, otherwise I suspect coverage data will accumulate from one test run to the next

17 years agorename the new filetable code to 'filetree', since robk astutely pointed out that...
Brian Warner [Mon, 8 Jan 2007 23:29:14 +0000 (16:29 -0700)]
rename the new filetable code to 'filetree', since robk astutely pointed out that 'table' is misleading and implies a flat list of files in a single directory

17 years agofigleaf.el: add some emacs22 comments
Brian Warner [Sun, 7 Jan 2007 20:10:15 +0000 (13:10 -0700)]
figleaf.el: add some emacs22 comments

17 years agoMakefile: add a default target to just build the tree
Brian Warner [Sun, 7 Jan 2007 20:09:37 +0000 (13:09 -0700)]
Makefile: add a default target to just build the tree

17 years agofix pyflakes warning in debugshell, by providing a dummy value for 'app' that will...
Brian Warner [Sat, 6 Jan 2007 04:15:59 +0000 (21:15 -0700)]
fix pyflakes warning in debugshell, by providing a dummy value for 'app' that will be overwritten when the manhole connection is established

17 years agouse absolute import of 'allmydata.Crypto' rather than a relative import of just ...
Brian Warner [Sat, 6 Jan 2007 04:12:26 +0000 (21:12 -0700)]
use absolute import of 'allmydata.Crypto' rather than a relative import of just 'Crypto', to make it clear that we're using our own form rather than relying upon the system version

17 years agoworkqueue: fix pyflakes warnings, code is still quite incomplete
Brian Warner [Sat, 6 Jan 2007 04:11:19 +0000 (21:11 -0700)]
workqueue: fix pyflakes warnings, code is still quite incomplete

17 years agodebian: give incrementing version numbers to .deb packages, using timestamps
Brian Warner [Sat, 6 Jan 2007 03:06:51 +0000 (20:06 -0700)]
debian: give incrementing version numbers to .deb packages, using timestamps

17 years agotest_encode_share.py: fix some pyflakes warnings
Brian Warner [Sat, 6 Jan 2007 01:53:22 +0000 (18:53 -0700)]
test_encode_share.py: fix some pyflakes warnings

17 years agoadd some (disabled) encoder benchmarking code
Brian Warner [Sat, 6 Jan 2007 01:46:03 +0000 (18:46 -0700)]
add some (disabled) encoder benchmarking code

17 years agochange Encoder to use py_ecc, now that it works
Brian Warner [Sat, 6 Jan 2007 01:12:45 +0000 (18:12 -0700)]
change Encoder to use py_ecc, now that it works

17 years agodisable figleaf tracing during py_ecc, since it takes *forever*, especially on the...
Brian Warner [Sat, 6 Jan 2007 01:12:04 +0000 (18:12 -0700)]
disable figleaf tracing during py_ecc, since it takes *forever*, especially on the slow buildslave

17 years agoencode.py: add some timing comments
Brian Warner [Fri, 5 Jan 2007 07:48:42 +0000 (00:48 -0700)]
encode.py: add some timing comments

17 years agoonly run a single (short) py_ecc test on slave3, since it is so slow the tests timeout
Brian Warner [Fri, 5 Jan 2007 07:42:52 +0000 (00:42 -0700)]
only run a single (short) py_ecc test on slave3, since it is so slow the tests timeout

17 years agoencoding: fix the last py_ecc problem, tests pass now
Brian Warner [Fri, 5 Jan 2007 07:06:42 +0000 (00:06 -0700)]
encoding: fix the last py_ecc problem, tests pass now

17 years agochange encode_new to use IEncoder
Brian Warner [Fri, 5 Jan 2007 06:51:35 +0000 (23:51 -0700)]
change encode_new to use IEncoder

17 years agofix our use of py_ecc (set log2FieldSize=8 explicitly)
Brian Warner [Fri, 5 Jan 2007 06:50:21 +0000 (23:50 -0700)]
fix our use of py_ecc (set log2FieldSize=8 explicitly)

17 years agoestablish IEncoder/IDecoder, create suitable interfaces for both the simple replicati...
Brian Warner [Fri, 5 Jan 2007 04:52:51 +0000 (21:52 -0700)]
establish IEncoder/IDecoder, create suitable interfaces for both the simple replicating encoder and the py_ecc one, add a (failing) unit test for it

17 years agofigleaf.el: announce annotation/deannotation
Brian Warner [Fri, 5 Jan 2007 04:52:40 +0000 (21:52 -0700)]
figleaf.el: announce annotation/deannotation

17 years agomore pyflakes cleanups
Brian Warner [Fri, 5 Jan 2007 01:06:20 +0000 (18:06 -0700)]
more pyflakes cleanups

17 years agofix pyflakes warnings in figleaf
Brian Warner [Fri, 5 Jan 2007 01:04:43 +0000 (18:04 -0700)]
fix pyflakes warnings in figleaf

17 years agoadd padding to segments and convert type from string to list of integers before passi...
Zooko O'Whielacronx [Fri, 5 Jan 2007 00:58:14 +0000 (17:58 -0700)]
add padding to segments and convert type from string to list of integers before passing to py_ecc
But then I get an error that I don't understand about ECC math...

17 years agoimprove figleaf: fix some line-numbering bugs, add an alphabetically-sorted index...
Brian Warner [Thu, 4 Jan 2007 08:26:43 +0000 (01:26 -0700)]
improve figleaf: fix some line-numbering bugs, add an alphabetically-sorted index HTML page

17 years agomore figleaf emacs work: enable the minor mode, then type C-cA to toggle annotations
Brian Warner [Thu, 4 Jan 2007 08:25:36 +0000 (01:25 -0700)]
more figleaf emacs work: enable the minor mode, then type C-cA to toggle annotations

17 years agomodify figleaf2html to show module names instead of .py filenames, also add a --root...
Brian Warner [Thu, 4 Jan 2007 05:23:25 +0000 (22:23 -0700)]
modify figleaf2html to show module names instead of .py filenames, also add a --root argument to restrict coverage to a specific parent directory

17 years agomove figleaf2html and figleaf_htmlizer.py into our tree, for easier customization
Brian Warner [Thu, 4 Jan 2007 05:06:51 +0000 (22:06 -0700)]
move figleaf2html and figleaf_htmlizer.py into our tree, for easier customization

17 years agofigleaf: move a copy into allmydata.util.figleaf, update Makefile/trial stuff
Brian Warner [Thu, 4 Jan 2007 04:38:29 +0000 (21:38 -0700)]
figleaf: move a copy into allmydata.util.figleaf, update Makefile/trial stuff

17 years agoadd some experimental emacs test-coverage-annotation tools, still in development
Brian Warner [Tue, 2 Jan 2007 06:48:42 +0000 (23:48 -0700)]
add some experimental emacs test-coverage-annotation tools, still in development

17 years agocheckpoint work-in-progress for WorkQueue, a disk-persistent list of work to be done
Brian Warner [Tue, 2 Jan 2007 06:47:16 +0000 (23:47 -0700)]
checkpoint work-in-progress for WorkQueue, a disk-persistent list of work to be done

17 years agofiletable_new: import py_ecc, make the tests get a little bit further
Brian Warner [Tue, 2 Jan 2007 06:44:12 +0000 (23:44 -0700)]
filetable_new: import py_ecc, make the tests get a little bit further

17 years agoadd __init__ and setup.py glue for py_ecc, so we can import it
Brian Warner [Tue, 2 Jan 2007 06:43:24 +0000 (23:43 -0700)]
add __init__ and setup.py glue for py_ecc, so we can import it

17 years agomake encode_new use py_ecc for real live erasure coding
Zooko O'Whielacronx [Fri, 29 Dec 2006 20:50:53 +0000 (13:50 -0700)]
make encode_new use py_ecc for real live erasure coding
(This patch is not tested -- I'm working on a Mac which doesn't have gcc
installed...  (gcc is necessary for the crypto module.)  I will now attempt to
connect to a better development computer over my mom's staticy, failure-prone,
14.4 K modem...)

17 years agoimport mathutil from pyutil
Zooko O'Whielacronx [Fri, 29 Dec 2006 20:50:42 +0000 (13:50 -0700)]
import mathutil from pyutil

17 years agoimport py_ecc, a pure python fec by Emin Martinian, which is under a permissive licence
Zooko O'Whielacronx [Fri, 29 Dec 2006 20:46:45 +0000 (13:46 -0700)]
import py_ecc, a pure python fec by Emin Martinian, which is under a permissive licence
It is too slow for a real product, but is a quick way to get a working prototype, and also is freely redistributable by us...

17 years agoencode_new: fix a bug in subshare hashes and tidy-up a couple of things
Zooko O'Whielacronx [Fri, 29 Dec 2006 19:40:10 +0000 (12:40 -0700)]
encode_new: fix a bug in subshare hashes and tidy-up a couple of things

17 years agomore filetable_new tests
Brian Warner [Mon, 25 Dec 2006 07:56:18 +0000 (00:56 -0700)]
more filetable_new tests

17 years agofiletable_new: fix the test
Brian Warner [Sun, 24 Dec 2006 19:47:03 +0000 (12:47 -0700)]
filetable_new: fix the test

17 years agocheckpointing new filetable work.. tests don't pass yet
Brian Warner [Sun, 24 Dec 2006 19:39:24 +0000 (12:39 -0700)]
checkpointing new filetable work.. tests don't pass yet

18 years agofiletable: shuffle lines a bit to appease figleaf's confusion
Brian Warner [Thu, 14 Dec 2006 11:33:57 +0000 (04:33 -0700)]
filetable: shuffle lines a bit to appease figleaf's confusion

18 years agoencode: start to fix a few problems, still a lot of work left to go
Brian Warner [Thu, 14 Dec 2006 11:31:17 +0000 (04:31 -0700)]
encode: start to fix a few problems, still a lot of work left to go