f4280a2b2c
Tweak put tests
...
Add a very basic test, and put everything in the same directory so it
gets cleaned up properly.
2018-03-16 12:59:08 -07:00
5434f3676c
Add the ability for 'put -' to read from stdin
...
Fixes #44 , closes #98
2018-03-16 12:59:08 -07:00
d961456920
Add StatFs and the df [-h] command
2017-03-14 00:47:54 +01:00
a732047294
Allow the CLI to load the NameNodes from the Hadoop configuration
2017-02-13 11:34:53 -07:00
5ca9568680
Add a -v / --version flag
2017-01-23 13:08:15 +01:00
cefa4bc7e7
Make mv behavior match the posix spec a little better
...
- Moving a non-directory on top of a directory should never work
- Replacing an existing file should be the default
2016-10-26 21:05:48 +02:00
8774aa7d4c
Force overwrite in rename and provide option support in mv
2016-10-26 10:15:58 -07:00
b629910101
Format ls -l year correctly
...
Without this fix, ls -l generated Unicode character 2015 for the year
2015 rather than the string "2015"
2016-04-03 16:49:58 -07:00
c43cceb489
Fix ls -l format
...
`ls -l` prints Month Day incorrectly for 2-digit dates. Use the _ empty
padding method to format the Month and Day.
2016-02-13 12:30:09 +01:00
2028fbc05e
Add hdfs put
2016-01-05 00:02:27 +01:00
0e0b705362
Add Username function and also add to paths which only had old logic
2015-12-21 16:33:30 +01:00
a08db18c38
all: run goimports -w -l .
2015-11-23 16:25:37 -08:00
7df08bbd5e
cmd/tail: do not try to read more than file size
2015-11-23 12:22:47 -08:00
885c569282
file: implement io.ReaderAt correctly
...
io.ReaderAt says ReadAt should fill up the passed in slice as
much as there is data from the passed in offset.
2015-11-22 17:18:34 -08:00
d1dc635a3f
fetch getopt from github instead of googlecode
...
`getopt` package is moved to github, so there's no reason to use old
code.google.com link. It also removes mercurial dependency on installing
`hdfs` command.
2015-06-26 14:39:30 +09:00
9519821475
remove completion testing code that snuck in
2015-02-26 16:47:44 +01:00
5ced0621b7
Add GetContentSummary, and rewrite `hdfs du` to use it
...
The current process, of recursively stat'ing files, is way too slow for big
directory trees. GetContentSummary returns the total size for a whole tree, so
lets use that.
2015-02-26 16:39:24 +01:00
1cc0e2711a
divert minicluster output to a file
2015-02-25 12:23:52 +01:00
7306883008
When creating directories for 'get', create them as 0755.
...
Fixes #9 .
2015-02-25 03:13:06 +01:00
93e320c0d9
tail should print to stdout, not to stderr
2015-01-19 23:12:10 +01:00
0b6bc39d66
add du -sh
2015-01-06 12:45:13 +01:00
738b7cc9c9
in ls, don't flush the tabwriter until the very end
...
(also make some formatting tweaks)
2015-01-06 12:40:26 +01:00
979dfc4c44
gofmt
2015-01-06 12:36:14 +01:00
c8bc953e4d
make exit status global
2015-01-06 12:05:17 +01:00
efbd8f615e
add ls -h
2014-12-19 22:54:20 +01:00
3caac15a68
add FileReader.Checksum and 'hdfs checksum'
...
The FileReader.Checksum method provides access to HDFS's built in checksum
ability, which works by checksumming concatenated checksums which are already
stored alongside the data.
This also adds a datanodeFailover type, to abstract over handling datanode
failure in multiple contexts (reading the checksum is the second such context;
reading blocks was the original one, and writing blocks will eventually be
similar).
2014-12-19 02:54:26 +01:00
f2f22c4631
gofmt
2014-11-19 02:50:31 +01:00
e1dcc570c1
vendor mobydick.txt
...
It changed on gutenberg (????), which broke the tests.
2014-11-19 02:46:49 +01:00
7c2faa189c
make sure we're not clobbering the canonical mobydick.txt
2014-11-19 02:34:20 +01:00
b975c54800
more linting
2014-11-07 20:13:50 +01:00
5c5077f5b6
Match unix ls by printing full paths (only when passed them directly)
2014-11-04 20:58:32 +01:00
f9ffd5ccd4
make sure we never expand nonexistent paths from globs
2014-11-04 20:16:22 +01:00
7590866a06
update one test message
2014-11-04 00:40:34 +01:00
1406bb3801
update usage for rm -f
2014-11-04 00:40:15 +01:00
5d6bea147b
Merge pull request #2 from snoble/snoble-rm-f
...
Add -f to rm to suppress errors on non-existence.
2014-11-04 00:39:25 +01:00
dce42f8c73
Add -f to rm to suppress errors on non-existence.
2014-11-03 15:20:59 -05:00
124daea3f1
gomft
2014-11-03 20:20:11 +01:00
5288ee46f3
remove unused import
2014-11-01 19:40:55 +01:00
6c23e98338
add tests for head and tail
2014-11-01 18:58:53 +01:00
f9b2974614
Instead of using a Scanner to read lines in head, look for the offset and copy
...
This fixes two bugs:
- trailing CRs were being stripped from the output, because the ScanLines
implementation does that
- long files without newlines weren't being printed at all
2014-11-01 18:56:56 +01:00
cc34a26dc3
another test fix
2014-11-01 18:36:15 +01:00
5c76a7becb
fix touch test
2014-11-01 18:34:20 +01:00
7916986dd4
fix rm tests
2014-10-31 21:39:07 +01:00
3b3141ba95
add tests for cat
2014-10-31 21:38:12 +01:00
5baa02d9fe
add tests for mkdir and touch
2014-10-31 21:19:37 +01:00
de48a95262
add tests for mv
2014-10-31 21:09:20 +01:00
a0befc4db3
when rming, change the Op for the error to 'rm'
2014-10-31 21:04:46 +01:00
e48cf99b93
add some rm tests
2014-10-31 21:00:04 +01:00
56af180726
refactor out caching in cmd/hdfs
...
It's definitely a premature optimization, and also blows up with huge
directories.
2014-10-31 20:46:22 +01:00
5cd50d3b56
attempt to fix travis
2014-10-31 20:15:04 +01:00
Colin Marc
Shastick
Junjie Qian
Shane Kumpf
Dave Beckett
Ben Kochie
Erik Bartels
Ashish Gandhi
Jihyun Yu
Steven H. Noble