Commit Graph

105 Commits

Author SHA1 Message Date
Jakob Borg
cebe21a3af Don't buffer large files a long time while scanning 2015-04-17 16:40:09 +09:00
Jakob Borg
6d84443db8 Homogenize the LocalIndexUpdated event
It had two different formats, and we use "items" instead of "numFiles"
in other places.

(Discovered while documenting :)
2015-04-17 14:22:06 +09:00
Jakob Borg
102a2db1f3 Work around broken Lstat on Android 2015-04-14 19:53:49 +09:00
Jakob Borg
aa803ce2ff Move folder errors to state
The "Invalid" config attribute is retained for errors discovered during
config loading (empty path, duplicate ID). This can only be set or
cleared at config loading time.

Errors discovered during runtime (I/O problems, etc) are now in the
folder state instead. Changes to these are sent as any other folder
state change.
2015-04-13 07:43:45 +09:00
Jakob Borg
936c76119d Index reset should generate file conflicts (fixes #1613) 2015-04-09 13:06:09 +02:00
Audrius Butkevicius
df6dbc5fa4 Only run deadlock detection if asked or non-release/beta (fixes #1536) 2015-04-08 13:40:05 +01:00
Jakob Borg
b2a79855ae Handle long filenames on Windows (fixes #1295) 2015-04-08 14:05:39 +02:00
Audrius Butkevicius
8f05b8f982 Change /rest/system/connections output (fixes #1487) 2015-04-07 13:21:03 +01:00
Jakob Borg
a17333d73e Override needs to twiddle the version a bit more (fixes #1564) 2015-04-06 20:55:40 +02:00
Jakob Borg
ecadf30fe7 model: Use separate db commit routine (fixes #1558) 2015-04-05 16:19:14 +02:00
Lode Hoste
ab287ebf40 Update reset API to reflect new use cases.
/rest/reset clears the entire Syncthing DB and restart the program
/rest/reset&folder=default clears the indexes of the default folder
2015-04-04 22:45:11 +02:00
Audrius Butkevicius
e1757ee726 Fix test 2015-03-30 22:49:16 +01:00
Audrius Butkevicius
7406176fad More graceful handling on folder errors (fixes #762)
Checks health before accepting every scanner batch, also
recovers from errors without having to restart.
2015-03-30 08:27:12 +02:00
Jakob Borg
ba575f55ec Merge pull request #1530 from Zillode/multi-scan
Support multiple scan query strings at the same time
2015-03-29 16:02:45 +02:00
Lode Hoste
2012ce02e8 Support multiple scan query strings at the same time 2015-03-28 22:40:13 +01:00
Audrius Butkevicius
c12265499a Response errors may be protocol defined errors 2015-03-26 22:04:34 +00:00
Audrius Butkevicius
0e341832e0 Handle unknown flags at the model 2015-03-26 22:04:34 +00:00
Audrius Butkevicius
489e2e6ad5 Update Model function signatures 2015-03-26 22:04:33 +00:00
Jakob Borg
6da7f17c4a Implement version vectors 2015-03-25 23:10:34 +01:00
Jakob Borg
8311162be3 Automatically fix file name normalization errors (fixes #430) 2015-03-19 00:21:48 +01:00
Jakob Borg
bdbca75dfa Refactor state tracking (...)
Move state tracking into the puller/scanner objects. This is a first
step towards resolving #1391.

Rename Puller and Scanner to roFolder and rwFolder as they have more
duties than just pulling and scanning, and don't need to be exported.
2015-03-17 19:37:06 +01:00
Audrius Butkevicius
de38b46392 Fix build 2015-03-17 17:54:25 +00:00
Jakob Borg
32425c5561 MPLv2 2015-03-17 16:02:27 +01:00
Pascal Jungblut
49bc74e7a0 Use lowerCamelCase for the JSON API (fixes #1338)
Replace the current mix of UpperCamelCase und lowerCamelCase with
consistent lowerCamelCase keys for the JSON API. Also adapt the frontend
so it works with the changed API.

Attention: this will break existing consumers of the API.
2015-03-16 10:05:01 +01:00
Audrius Butkevicius
57fc0eb5b1 Make sure we start scanning at an indexed location (fixes #1399) 2015-03-08 19:45:47 +00:00
Audrius Butkevicius
9d1e2d9f46 Add /rest/tree API call 2015-03-04 23:39:27 +00:00
Jakob Borg
c25107eff3 Handle weird Lstat() returns for disappeared items (ref #1373) 2015-03-01 10:55:43 +01:00
Karol Różycki
ec160f1f0a Button to rescan all folders, fixes #1151 2015-02-12 21:03:35 +01:00
Jakob Borg
6bbd74adcd Merge pull request #1321 from AudriusButkevicius/bitcheck
Refuse files with unknown bits set (fixes #1276)
2015-02-10 20:27:14 +01:00
Audrius Butkevicius
b87ed97402 Refuse files with unknown bits set (fixes #1276) 2015-02-09 23:32:33 +00:00
Audrius Butkevicius
380d5dfa6d Remove fd cache (ref #1308) 2015-02-01 18:59:24 +00:00
Jakob Borg
2c8b627008 Integer type policy
Integers are for numbers, enabling arithmetic like subtractions and for
loops without getting shot in the foot. Unsigneds are for bitfields.

- "int" for numbers that will always be laughably smaller than four
  billion, and where we don't care about the serialization format.

- "int32" for numbers that will always be laughably smaller than four
  billion, and will be serialized to four bytes.

- "int64" for numbers that may approach four billion or will be
  serialized to eight bytes.

- "uint32" and "uint64" for bitfields, depending on required number of
  bits and serialization format. Likewise "uint8" and "uint16", although
  rare in this project since they don't exist in XDR.

- "int8", "int16" and plain "uint" are almost never useful.
2015-01-19 10:34:36 -08:00
Jakob Borg
221e3eddd5 Remove leveldb panic workaround
Haven't seen this triggered for a long time...
2015-01-19 10:23:00 -08:00
Jakob Borg
53cd289b90 Verify folder<->device permission in Request
Requests from valid devices for valid folders should be rejected if the
folder is not shared with that device.
2015-01-16 12:50:51 +01:00
Jakob Borg
eef1aebe8c Refactor out protocol and luhn (protocol dependency) packages 2015-01-13 13:22:56 +01:00
Jakob Borg
e6c9baf6ef Rename db.Set to db.FileSet 2015-01-12 20:57:39 +01:00
Jakob Borg
8d6db7be31 Rename package internal/files to internal/db 2015-01-12 20:57:22 +01:00
Jakob Borg
571cf7d490 Merge pull request #1182 from AudriusButkevicius/autoauto
Connecting to a newer node triggers autoupgrade check (fixes #1177)
2015-01-11 09:16:12 +01:00
Audrius Butkevicius
5034a41c08 Connecting to a newer node triggers autoupgrade check (fixes #1177) 2015-01-10 17:05:19 +00:00
Audrius Butkevicius
5827a686b8 Make parallel hasher configurable, remove finisher setting (fixes #1199) 2015-01-10 14:32:20 +00:00
Jakob Borg
d46ce5003c Implement GetGlobalTruncated 2015-01-09 08:41:02 +01:00
Jakob Borg
4c4143d9be Move FileInfoTruncated to files package
This is where it's used, and it clarifies that it's never used over the
wire.
2015-01-09 08:28:24 +01:00
Jakob Borg
8bc7d259f4 Move FileIntf to files package, expose Iterator type
This is where FileIntf is used, so it should be defined here (it's not
a protocol thing, really).
2015-01-09 08:18:42 +01:00
Jakob Borg
75d0dc251e Use comma-ok idiom to signal files missing in database (fixes #1186)
Prevents us from doing stupid things to the folder root (empty file
path) when nodes disconnect...
2015-01-06 22:40:20 +01:00
Audrius Butkevicius
5ac01a3af4 Hash blocks after receipt, try multiple peers (fixes #1166) 2015-01-03 23:21:57 +00:00
Audrius Butkevicius
56ccb5b2ab New device, folder prompts (fixes #120, fixes #330) 2015-01-03 23:06:41 +00:00
Jakob Borg
5143c09bcf Refactor / cleanup 2015-01-02 15:54:04 +01:00
Jakob Borg
2496185629 Only buffer file names, not full &FileInfo 2015-01-02 15:33:39 +01:00
Audrius Butkevicius
fd0a147ae6 Add job queue (fixes #629)
Request to terminate currently ongoing downloads and jump to the bumped file
incoming in 3, 2, 1.

Also, has a slightly strange effect where we pop a job off the queue, but
the copyChannel is still busy and blocks, though it gets moved to the
progress slice in the jobqueue, and looks like it's in progress which it isn't
as it's waiting to be picked up from the copyChan.

As a result, the progress emitter doesn't register on the task, and hence the file
doesn't have a progress bar, but cannot be replaced by a bump.

I guess I can fix progress bar issue by moving the progressEmiter.Register just
before passing the file to the copyChan, but then we are back to the initial
problem of a file with a progress bar, but no progress happening as it's stuck
 on write to copyChan

I checked if there is a way to check for channel writeability (before popping)
but got struck by lightning just for bringing the idea up in #go-nuts.

My ideal scenario would be to check if copyChan is writeable, pop job from the
queue and shove it down handleFile. This way jobs would stay in the queue while
they cannot be handled, meaning that the `Bump` could bring your file up higher.
2015-01-02 15:33:39 +01:00
Jakob Borg
22f39be197 Exit before attempting to use nil variables on scanning nonexistent folder 2014-12-23 14:14:05 +01:00