Commit graph

38 commits

Author SHA1 Message Date
jmjatlanta
2051f7714a ipfs name command line options processing 2017-09-21 07:51:18 -05:00
a907f1dd2d
The API changed pthread to scope, so it can load multiple instances. 2017-09-20 14:39:26 -03:00
jmjatlanta
acf506296e Updating tests 2017-09-18 19:04:52 -05:00
John Jones
a9481631df Several memory leak fixes for journal code 2017-09-07 18:45:09 -05:00
John Jones
cb1ea3ceff Sorting a memory leak in datastore/journalstore 2017-09-07 11:05:56 -05:00
jmjatlanta
e4f1c9b39c Making providerstore smarter 2017-07-31 12:50:12 -05:00
jmjatlanta
cb05b249ba Fixed warnings 2017-07-17 16:14:40 -05:00
John Jones
069379acf4 Fix of a few compiler warnings 2017-07-17 13:05:56 -05:00
John Jones
d6ee0f7d5d Fixed small memory leak 2017-05-11 07:13:13 -05:00
John Jones
def5331d4c Better handling of locally stored files 2017-05-11 07:04:54 -05:00
John Jones
3de4b757e4 Bugfixes and unit test fixes 2017-04-27 15:52:20 -05:00
John Jones
e756fdf510 Memory and test fixes 2017-04-27 11:35:26 -05:00
John Jones
03696dd6e7 Major changes to support large file transfer 2017-04-20 17:56:03 -05:00
John Jones
2b0a29a06b Implemented find providers from remote peers
Now, you can ask a known node for a key, and it can pass a list of peers
that are perhaps unknownn to the calling peer.
2017-04-16 23:47:53 -05:00
John Jones
62096ffc1c Handling finding of peer via swarm
Also included is a way to pass port and swarm info to methods that build
the config file. This makes testing easier. Multiple peers can be
started on the same machine easily.
2017-04-13 09:31:58 -05:00
John Jones
94d6005587 Moved datastore interface from ipfs to libp2p 2017-04-06 09:33:28 -05:00
John Jones
e8b8d06f24 refactoring + beginning of file transfer
Now attempting to use kademlia to find a hash, and NodeIO to transfer
the file
2017-03-19 07:47:19 -05:00
John Jones
15a8abff9a Made the stream methods more generic 2017-03-09 10:01:09 -05:00
John Jones
f1aac5d707 Starting to handle get_value request 2017-02-27 12:27:40 -05:00
John Jones
9882c28743 Fixed memory leaks
Several tests had memory leaks. As well found a few leaks within the
importer and resolver areas.
2017-01-02 00:38:09 -05:00
jmjatlanta
61d0adc445 misc tweaks and fixes
Now paying attention to the IPFS_PATH environment variable to determine
where the repository is. Fixed some broken tests. Fixed a bug whereby a
subdirectory within a subdirectory was not displaying correctly when
imported.
2017-01-01 23:48:09 -05:00
jmjatlanta
c2fe60949e Added the ability to retrieve the file using directories
Files can be refered to directly using their hash, or the hash of their
directory and the file name.
2016-12-29 19:05:44 -05:00
jmjatlanta
396dfc6abc Directory hashes match 2016-12-29 04:42:01 -05:00
John Jones
8d2aeab016 Fixed various memory leaks 2016-12-23 10:49:30 -05:00
jmjatlanta
c8fdb084e4 Moving closer to binary compatability with go version of ipfs
The files are now stored in the same format. Now it is necessary to
reverse engineer the directory structure
2016-12-21 06:40:19 -05:00
jmjatlanta
a1166e840a Move of node to merkledag 2016-12-19 09:03:28 -05:00
jmjatlanta
1d63cdb4a1 Persisting large files in chunks 2016-12-15 12:38:08 -05:00
jmjatlanta
34301c286e Storing large files
Files larger than about 200K are split into smaller files and stored in
the ipfs file system in blocks.
2016-12-15 05:40:24 -05:00
John Jones
5168bc87e0 Fixed memory leak in test 2016-12-14 12:24:44 -05:00
jmjatlanta
033dd767b4 More work on persisting data to disk.
Blockstore now storing the data, whereas datastore is storing the key
and filename. The key should be the multihash (currently the sha256, not
the multihash), and the value is the filename (base32).
2016-12-14 12:07:43 -05:00
John Jones
88d177cf4a Fixed minor leak in test and more compiler warnings 2016-12-14 06:39:06 -05:00
jmjatlanta
bf0cbfb412 Removed compiler warnings, starting to import large files 2016-12-14 06:25:09 -05:00
John Jones
876e2dfcf2 Fixed memory leaks around node and allocations 2016-12-12 18:27:46 -05:00
jmjatlanta
a36ba20557 made node links a single linked list instead of array
This will save some complicated memory allocations
2016-12-12 16:58:27 -05:00
jmjatlanta
4fe768c2c5 Working with nodes to add protobuf to persist to db
adding a link to a node does a realloc. I believe this is causing
problems. I am going to replace this and see if that fixes the issue.
2016-12-12 15:06:17 -05:00
jmjatlanta
e0b0552b39 Pushing Cid into protobuf 2016-12-12 06:27:06 -05:00
John Jones
a180a63160 Setting uninitialized values
Uninitialized values should be set to NULL to aid in memory deallocation
at cleanup time.
2016-12-05 18:17:17 -05:00
jmjatlanta
b9b90014a6 Fixed tests 2016-12-05 06:13:20 -05:00