netatalk.git
5 years agoSpotlight: use async Tracker SPARQL API master
Ralph Boehme [Tue, 5 Aug 2014 11:14:21 +0000 (13:14 +0200)]
Spotlight: use async Tracker SPARQL API

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agologger: remove flood protection and allocate messages
Ralph Boehme [Tue, 5 Aug 2014 13:30:20 +0000 (15:30 +0200)]
logger: remove flood protection and allocate messages

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoMerge remote-tracking branch 'origin/branch-netatalk-3-1'
Ralph Boehme [Fri, 1 Aug 2014 16:19:10 +0000 (18:19 +0200)]
Merge remote-tracking branch 'origin/branch-netatalk-3-1'

5 years agoafpd: check for modified included config file, FR #95
Ralph Boehme [Fri, 1 Aug 2014 14:09:03 +0000 (16:09 +0200)]
afpd: check for modified  included config file, FR #95

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoMerge remote-tracking branch 'origin/branch-netatalk-3-1'
Ralph Boehme [Fri, 1 Aug 2014 14:19:01 +0000 (16:19 +0200)]
Merge remote-tracking branch 'origin/branch-netatalk-3-1'

5 years agofce: afpd: fix event names array
Ralph Boehme [Mon, 21 Jul 2014 15:31:00 +0000 (17:31 +0200)]
fce: afpd: fix event names array

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agofce: fix event id increment
Ralph Boehme [Mon, 21 Jul 2014 15:30:07 +0000 (17:30 +0200)]
fce: fix event id increment

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agofce: afpd: missing includes and fix SIGCLD handler argument list
Ralph Boehme [Mon, 21 Jul 2014 15:28:10 +0000 (17:28 +0200)]
fce: afpd: missing includes and fix SIGCLD handler argument list

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoman: afp.conf: "fce version" option
Ralph Boehme [Mon, 21 Jul 2014 15:16:45 +0000 (17:16 +0200)]
man: afp.conf: "fce version" option

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agodist: add missing file abigen.sh
Ralph Boehme [Fri, 18 Jul 2014 07:57:14 +0000 (09:57 +0200)]
dist: add missing file abigen.sh

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agofce: FCE version 2 with new event types and new config options
Ralph Boehme [Tue, 15 Jul 2014 23:16:33 +0000 (01:16 +0200)]
fce: FCE version 2 with new event types and new config options

* new events: fmov,dmov,login and logout
* removed event: tmsz
* new options: "fce ignore names" and "fce notify script"

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoWriting metadata xattr on directories with sticky bit set, FR#94
Ralph Boehme [Wed, 25 Jun 2014 09:46:03 +0000 (11:46 +0200)]
Writing metadata xattr on directories with sticky bit set, FR#94

Even though we may have write access to a directory, if the sticky bit
is set only the owner is allowed to write xattrs.

Add an option that enables behaviour to set the xattr as root, using
the fact that we can distinguish between a failing xattr set request
due to insufficient permissions (EACCES) or the sticky bit (EPERM):

  force xattr with sticky bit = yes|no (default: no)

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoconfig: add utility function vdgoption_bool()
Ralph Boehme [Thu, 10 Jul 2014 15:59:53 +0000 (17:59 +0200)]
config: add utility function vdgoption_bool()

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agolibrary versioning: update to 16:0:0
Ralph Boehme [Sat, 5 Jul 2014 06:43:12 +0000 (08:43 +0200)]
library versioning: update to 16:0:0

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoHangs in Netatalk which causes it to stop responding to connections
Ralph Boehme [Sat, 5 Jul 2014 06:14:41 +0000 (08:14 +0200)]
Hangs in Netatalk which causes it to stop responding to connections

Hangs in Netatalk which causes it to stop responding to
connections. The master afpd process gets stuck in a poll loop, being
repeatedly notified that there are connections on its socket, but
never actually doesn anything with them.

Analysis with gdb revealed that the dat astructure dealing with the
main AFP socket and the IPC client sockets was smashed. This could
happen because the function fdset_add_fd() doesn't do bound checking
itself but relied on other parts of the code that enforce a connection
limit. Unfortunately, for low-level AFP connections that don't result
in a full AFP login these checks come too late resulting in a buffer
overflow.

Add a bound check. While we're at it, rewrite the fdset code to use a
full blown data structure 'struct asev' encapsultating the
implementation details.

Fixes bug #572.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoversion: 3.1.4dev
Ralph Boehme [Thu, 3 Jul 2014 11:00:20 +0000 (13:00 +0200)]
version: 3.1.4dev

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoRelease: version 3.1.3
Ralph Boehme [Thu, 3 Jul 2014 07:59:12 +0000 (09:59 +0200)]
Release: version 3.1.3

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoabigen: fix for gdb syms error
Ralph Boehme [Thu, 3 Jul 2014 07:57:42 +0000 (09:57 +0200)]
abigen: fix for gdb syms error

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agogitignore: test-driver
Ralph Boehme [Thu, 3 Jul 2014 05:54:20 +0000 (07:54 +0200)]
gitignore: test-driver

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agomanpages: generate from XML
Ralph Boehme [Thu, 3 Jul 2014 05:46:47 +0000 (07:46 +0200)]
manpages: generate from XML

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoAdd advanced option "chmod request" controlling ACLs
Ralph Boehme [Tue, 24 Jun 2014 16:25:15 +0000 (18:25 +0200)]
Add advanced option "chmod request" controlling ACLs

chmod request = ignore | preserve (default) | simple

* ignore = UNIX chmod() requests are completely ignored
* preserve = preserve ZFS ACEs for named users and groups or POSIX ACL
  group mask
* simple = just to a chmod() as requested without any extra steps

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoUse realpath_safe() for getting the absolute path
Ralph Boehme [Tue, 24 Jun 2014 08:44:47 +0000 (10:44 +0200)]
Use realpath_safe() for getting the absolute path

Authored-by: gaaronk
Reviewed-by: Ralph Boehme <rb@sernet.de>
5 years agoad: path detection in getvolbypath(), bug #570
Ralph Boehme [Fri, 20 Jun 2014 09:54:34 +0000 (11:54 +0200)]
ad: path detection in getvolbypath(), bug #570

The fix for bug #563 broke ad file utilities, bug #570.

Authored-by: gaaronk
Reviewed-by: Ralph Boehme <rb@sernet.de>
5 years agoResource fork conversion from AppleDouble v2 broken, bug #568
Ralph Boehme [Sun, 15 Jun 2014 04:07:10 +0000 (06:07 +0200)]
Resource fork conversion from AppleDouble v2 broken, bug #568

Converting from AppleDouble v2 to ea may corrupt the resource fork. In
some circumstances an offset calculation is wrong resulting in corrupt
resource forks after the conversion.

Ensure the resource fork AppleDouble entry offset is always set up and
we don't rely on the manual offset calculation.

Skip ADEID_RFORK in ad_copy_header() and allow copying of AppleDouble
entry that have an initial zero size in the destination adouble.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoMerge branch 'branch-netatalk-3-1' of netafp.com:git/netatalk into branch-netatalk-3-1
Ralph Boehme [Wed, 18 Jun 2014 15:00:36 +0000 (17:00 +0200)]
Merge branch 'branch-netatalk-3-1' of netafp.com:git/netatalk into branch-netatalk-3-1

Conflicts:
NEWS

5 years agoFix a crash when accessing ._ files, bug #564
Ralph Boehme [Wed, 18 Jun 2014 14:54:24 +0000 (16:54 +0200)]
Fix a crash when accessing ._ files, bug #564

At least FreeBSD is picky in that it requires PROT_READ if we're going
to read a mmap()ed memory region mapped, only specifying PROT_WRITE
doesn't imply PROT_READ. Other OSen seem to handle this differently.

Also add a missing pointer reinitialisation to the buf pointer.

Reported-by: Sean Eric Fagan
Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoNEWS: Fix FR number
Ralph Boehme [Wed, 18 Jun 2014 14:26:51 +0000 (16:26 +0200)]
NEWS: Fix FR number

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoDelete 0 byte size resource fork ._ files
Ralph Boehme [Thu, 12 Jun 2014 20:40:23 +0000 (22:40 +0200)]
Delete 0 byte size resource fork ._ files

Signed-off-by: Ralph Boehme <rb@sernet.de>
Change-Id: ef5f88f5f2cefa75cf20d0ee447b4e20579a2f12

5 years agoRefactoring: move forkpath computation to function scope
Ralph Boehme [Thu, 12 Jun 2014 20:24:42 +0000 (22:24 +0200)]
Refactoring: move forkpath computation to function scope

Preperation for new feature to delete zero byte sized resource forks.

Signed-off-by: Ralph Boehme <rb@sernet.de>
Change-Id: ef5f88f5f2cefa75cf20d0ee447b4e20579a2f12

5 years agoRefactoring: move struct dir and dirlookup() to function scope
Ralph Boehme [Thu, 12 Jun 2014 19:53:17 +0000 (21:53 +0200)]
Refactoring: move struct dir and dirlookup() to function scope

Preperation for new feature to delete zero byte sized resource forks.

Signed-off-by: Ralph Boehme <rb@sernet.de>
Change-Id: ef5f88f5f2cefa75cf20d0ee447b4e20579a2f12

5 years agoMemory overflow caused by 'basedir regex', bug #567
Ralph Boehme [Tue, 17 Jun 2014 17:25:45 +0000 (19:25 +0200)]
Memory overflow caused by 'basedir regex', bug #567

Authored-by: Takao Shimayoshi
Reviewed-by: Ralph Boehme <rb@sernet.de>
5 years agoUpdate Unicode support to version 7.0.0
Ralph Boehme [Tue, 17 Jun 2014 16:55:02 +0000 (18:55 +0200)]
Update Unicode support to version 7.0.0

Authored-by: HAT <hat@fa2.so-net.ne.jp>
Reviewed-by: Ralph Boehme <rb@sernet.de>
5 years agoUnarchiving certain ZIP archives fails, bug #569
Ralph Boehme [Tue, 17 Jun 2014 15:02:20 +0000 (17:02 +0200)]
Unarchiving certain ZIP archives fails, bug #569

The current code that checks whether the client is allowed to use
pathnames starting with ._ checks whether the file is an AppleDouble
file, and if it is and the filler does not contain the magic string
"Mac OS X", access to the file is denied. This results in the
following sequence of steps:

* client created empty file "._NAME"
* client wants to stat the newly created file "._NAME"
* server checks whether access to the file is allowed
* check fails

Modify check in ad_valid_header_osx() to allow access to all .-
AppleDouble files where the filler string is not "Netatalk", which is
the magic string Netatalk puts there.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agofix typo
HAT [Thu, 5 Jun 2014 15:40:45 +0000 (00:40 +0900)]
fix typo

5 years agoSpotlight: "spotlight expr = false" broke the default Spotlight query
Ralph Boehme [Wed, 4 Jun 2014 20:42:37 +0000 (22:42 +0200)]
Spotlight: "spotlight expr = false" broke the default Spotlight query

The default Spotlight query a Mac uses is

  *=="test*"cwd || kMDItemTextContent=="test*"cwd

which is of course using a logic expression. The parser has an extra
rule for catching this (MATCH or MATCH) for the purpose of munging
both expression into a single fts:match SPARQL query. Which means
we shouldn't check whether logic expressions are enabled or not.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoSpotlight: new options for controlling query behaviour
Ralph Boehme [Wed, 4 Jun 2014 14:36:58 +0000 (16:36 +0200)]
Spotlight: new options for controlling query behaviour

Add three options that allow fine grainted control over the resulting
SPARQL queries:

"sparql results limit = NUMBER", default is unlimited
Useful for limiting the result set for very large systems

"spotlight attributes = STRING", useful for limiting queries to
attributes that are indexed, many are not.

"spotlight expr = BOOLEAN", default: yes, useful for disabling the use
of complex queries that will take too long to complete and hang Tracker.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoSpotlight: SPARQL query optimisations
Ralph Boehme [Wed, 4 Jun 2014 10:11:40 +0000 (12:11 +0200)]
Spotlight: SPARQL query optimisations

Use tracker:uri-is-descendant FILTER instead of regex for filtering
results on the base path of the volume.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoBump version to 3.1.3dev
Ralph Boehme [Wed, 4 Jun 2014 08:28:19 +0000 (10:28 +0200)]
Bump version to 3.1.3dev

5 years agoRemove uniconv
Ralph Boehme [Tue, 3 Jun 2014 11:56:54 +0000 (13:56 +0200)]
Remove uniconv

5 years agoPrepare 3.1.2 releas
Ralph Boehme [Tue, 3 Jun 2014 11:50:39 +0000 (13:50 +0200)]
Prepare 3.1.2 releas

5 years agoAdd generated files from yacc and flex
Ralph Boehme [Tue, 3 Jun 2014 07:45:26 +0000 (09:45 +0200)]
Add generated files from yacc and flex

5 years agoSpotlight: SPARQL query optimisation
Ralph Boehme [Tue, 3 Jun 2014 07:41:25 +0000 (09:41 +0200)]
Spotlight: SPARQL query optimisation

Avoid DISTINCT in SPARQL queries and add a LIMIT which is arbitrarily
set to 100 for now.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoNew options that control whether dbus and Tracker are started
Ralph Boehme [Thu, 22 May 2014 11:16:02 +0000 (13:16 +0200)]
New options that control whether dbus and Tracker are started

'start dbus' and 'start tracker', both default to yes, FR#91

Signed-off-by: Ralph Boehme <sloowfranklin@gmail.com>
5 years agofd leak when using appledouble = v2, bug #554
Ralph Boehme [Wed, 21 May 2014 14:59:38 +0000 (16:59 +0200)]
fd leak when using appledouble = v2, bug #554

With appledouble = v2 in afp.conf there is an fd leak in certain
circumstances while using the Finder, from Daniel Hazelbaker.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agogetvolbypath returns incorrect volume, bug #563
Ralph Boehme [Wed, 21 May 2014 14:42:58 +0000 (16:42 +0200)]
getvolbypath returns incorrect volume, bug #563

getvolbypath() would match a given path "/foo/barbaz" with an existing
volume path of "/foo/bar" because the strings where compared with
strnlen() where n is the length of "/foo/bar", ie we did a substring
match.

I faintly remember there was a reason for this, but I couldn't come up
with any sensible configuration of usage of dbd that would not also
work by simply using strcmp() instead of strmcmp().

Of course, in order for dbd to work with paths that included a
trailing slash, there's a need to adding some slash stripping code,
because our code (somehow) ensures (hopefully!) that in struct
vol.v_path is always stored with trailinsh slashes stripped.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agodbd: two fixes for new user option
Ralph Boehme [Wed, 30 Apr 2014 12:47:20 +0000 (14:47 +0200)]
dbd: two fixes for new user option

* initialize user variable with NULL
* add -u to option summary line

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoSupport for using $u username variable in AFP volume definitions
Ralph Boehme [Wed, 19 Feb 2014 15:19:42 +0000 (16:19 +0100)]
Support for using $u username variable in AFP volume definitions

* Prepare CNID backend dbd for user variable expansion
* Store a struct vol pointer in the CNID db handles.
* Pass three arguments to the CNID daemon:
  - the afp.conf volume section name
  - the volume path
  - an optional username

* Rename several struct elements by prefixing them with a struct tag,
  helps catching bad or missing assignments.

* Disable uniconv, doesn't use laod_volumes() and thus wont work with
  the new CNID dbd backend.

From FR#90.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoxattrs on BSD, bug #562
Ralph Boehme [Tue, 20 May 2014 16:27:55 +0000 (18:27 +0200)]
xattrs on BSD, bug #562

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoquotas: improve Linux quota behaviour
Ralph Boehme [Wed, 30 Apr 2014 12:33:58 +0000 (14:33 +0200)]
quotas: improve Linux quota behaviour

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoMerge branch 'sharemodes-bug560' into branch-netatalk-3-1
Ralph Boehme [Tue, 20 May 2014 15:50:11 +0000 (17:50 +0200)]
Merge branch 'sharemodes-bug560' into branch-netatalk-3-1

5 years agolocking: put the Solaris share reservation after our locking stuff
Ralph Boehme [Thu, 1 May 2014 09:27:33 +0000 (11:27 +0200)]
locking: put the Solaris share reservation after our locking stuff

Fix for Bug 560: put the Solaris share reservation after our locking stuff.
Note that this still leaves room for a race condition where between placing our own
locks above and putting the Solaris share move below another client puts a lock.
We then end up with set locks from above and return with an error code, the proper
fix requires a sane cleanup function for the error path in this function.

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoSpotlight: add SPAQRL mapping for kMDItemLastUsedDate
Ralph Boehme [Thu, 1 May 2014 15:05:12 +0000 (17:05 +0200)]
Spotlight: add SPAQRL mapping for kMDItemLastUsedDate

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agofix typo
HAT [Thu, 1 May 2014 13:47:16 +0000 (22:47 +0900)]
fix typo

createvol -> creatvol

5 years agodebian initscript: add 0 and 6 to Default-Stop. debian-bug#745520
HAT [Sat, 26 Apr 2014 12:00:42 +0000 (21:00 +0900)]
debian initscript: add 0 and 6 to Default-Stop. debian-bug#745520

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=745520

5 years agoOnly run dbus-daemon if Spotlight is enabled, bug #556
Ralph Boehme [Wed, 23 Apr 2014 11:34:57 +0000 (13:34 +0200)]
Only run dbus-daemon if Spotlight is enabled, bug #556

5 years agoRemove tracker miner code
Ralph Boehme [Tue, 22 Apr 2014 14:25:34 +0000 (16:25 +0200)]
Remove tracker miner code

tracker_miner_manager_new_full() was reported to be potentially
locking up for an unknown reason.

As we diddn't use the miner anyway for other reasons (see comment in
the code), simply #ifdef 0 the whole code.

5 years agoadd pwent->pw_dir to logmessage.
HAT [Thu, 10 Apr 2014 13:36:01 +0000 (22:36 +0900)]
add pwent->pw_dir to logmessage.

inspired by Bug #557

5 years agoRemove misleading logmessage added in previous commit e4e8f3a8
Ralph Boehme [Thu, 10 Apr 2014 09:28:13 +0000 (11:28 +0200)]
Remove misleading logmessage added in previous commit e4e8f3a8

5 years agoImprove reliability of afpd child handler
Ralph Boehme [Tue, 8 Apr 2014 09:55:58 +0000 (11:55 +0200)]
Improve reliability of afpd child handler

5 years agoSpotlight: file modification date, bug #545
Ralph Boehme [Tue, 1 Apr 2014 10:15:32 +0000 (12:15 +0200)]
Spotlight: file modification date, bug #545

The Spotlight module misleadingly updated a files modification date
when the client requested an "kMDItemLastUsedDate" attribute update.

It's questionable whether we're expected at all to map and modify
attributes to filesystem metadata, but my guess is we are.

As a fix, use "kMDItemFSContentChangeDate" attribute instead of
"kMDItemLastUsedDate".

Fixes bug #545.

5 years agoAdd missing parenthesis
Ralph Boehme [Fri, 21 Mar 2014 10:20:35 +0000 (11:20 +0100)]
Add missing parenthesis

5 years agoFix typo and remove unused variables
Ralph Boehme [Tue, 18 Mar 2014 10:45:40 +0000 (11:45 +0100)]
Fix typo and remove unused variables

5 years agoOption "vol dbpath" was broken in 3.1.1
Ralph Boehme [Mon, 17 Mar 2014 10:08:57 +0000 (11:08 +0100)]
Option "vol dbpath" was broken in 3.1.1

Signed-off-by: Ralph Boehme <rb@sernet.de>
5 years agoBump version
Ralph Boehme [Thu, 13 Mar 2014 11:57:56 +0000 (12:57 +0100)]
Bump version

5 years agoUpdate release notes
Ralph Boehme [Thu, 13 Mar 2014 11:44:40 +0000 (12:44 +0100)]
Update release notes

5 years agoUpdate libatalk so version
Ralph Boehme [Thu, 13 Mar 2014 11:42:35 +0000 (12:42 +0100)]
Update libatalk so version

5 years agoBump version to 3.1.1
Ralph Boehme [Thu, 13 Mar 2014 11:06:23 +0000 (12:06 +0100)]
Bump version to 3.1.1

5 years agoDon't flood log with failed sys_set_ea() messages
Ralph Boehme [Sun, 2 Mar 2014 10:21:31 +0000 (11:21 +0100)]
Don't flood log with failed sys_set_ea() messages

5 years agoMerge branch 'branch-netatalk-3-1'
Ralph Boehme [Sat, 1 Mar 2014 06:23:57 +0000 (07:23 +0100)]
Merge branch 'branch-netatalk-3-1'

5 years agoRun dbpath through variable substitution
Ralph Boehme [Sat, 1 Mar 2014 06:00:29 +0000 (07:00 +0100)]
Run dbpath through variable substitution

5 years agoAdd a flag to load_volumes() that controls load behaviour
Ralph Boehme [Fri, 28 Feb 2014 17:56:52 +0000 (18:56 +0100)]
Add a flag to load_volumes() that controls load behaviour

5 years agoDon't print message "Ignoring ._file" for every ._ file, bug #552
Ralph Boehme [Fri, 21 Feb 2014 22:20:18 +0000 (23:20 +0100)]
Don't print message "Ignoring ._file" for every ._ file, bug #552

5 years agoFix a crash in of_closefork(), bug #551
Ralph Boehme [Thu, 20 Feb 2014 16:06:34 +0000 (17:06 +0100)]
Fix a crash in of_closefork(), bug #551

Apparently dirlookup() might return NULL when trying to query the did
of an open fork:

    dir = dirlookup(ofork->of_vol, ofork->of_did);

Add a NULL check before using the result.

5 years agoGenerate manpage from XML
Ralph Boehme [Thu, 20 Feb 2014 10:44:15 +0000 (11:44 +0100)]
Generate manpage from XML

5 years agoMerge remote-tracking branch 'origin/develop'
Ralph Boehme [Sat, 15 Feb 2014 07:22:57 +0000 (08:22 +0100)]
Merge remote-tracking branch 'origin/develop'

5 years agoAutomatic conversion for OS X ._ AppleDouble files, bug #550
Ralph Boehme [Fri, 14 Feb 2014 10:31:57 +0000 (11:31 +0100)]
conversion for OS X ._ AppleDouble files, bug #550

Add automatic on-access conversion for ._ AppleDouble files created by
OS X to Netatalk's AppleDouble format.

OS X .AppleDouble files most of the times contain a FinderInfo
AppleDouble entry with a size larger then 32 bytes, containing packed
xattrs. Netatalk fails to cleanly deal with these and logs them as
"bogus entry id 9".

5 years agodbd: remove orphaned ._ AppleDouble files. Bug #549.
Ralph Boehme [Thu, 13 Feb 2014 11:28:44 +0000 (12:28 +0100)]
dbd: remove orphaned ._ AppleDouble files. Bug #549.

5 years agoveto files: syntax clarification
HAT [Sun, 9 Feb 2014 03:17:15 +0000 (12:17 +0900)]
veto files: syntax clarification

5 years agoAdd file pathname to logmessage parse_entries: bogus eid. FR#87
Ralph Boehme [Sat, 1 Feb 2014 17:20:52 +0000 (18:20 +0100)]
Add file pathname to logmessage parse_entries: bogus eid. FR#87

5 years agoupdate manpage of afp.conf
HAT [Fri, 24 Jan 2014 12:09:49 +0000 (21:09 +0900)]
update manpage of afp.conf

add dbus daemon.
sort.

5 years agomacusers showed root user. Bug #495.
HAT [Tue, 21 Jan 2014 11:32:00 +0000 (20:32 +0900)]
macusers showed root user. Bug #495.

5 years agoRemoving ressource fork AppleDouble fails, bug #542
Ralph Boehme [Wed, 8 Jan 2014 00:30:47 +0000 (01:30 +0100)]
Removing ressource fork AppleDouble fails, bug #542

When the AFP client uses FPSetForkParms to set the ressource fork
length to 0, Netatalk will call unlink() in ad_rtruncate() in
libatalk/adouble/ad_write.c in order to to unlink the files
AppleDouble ressource fork.

The function ad_rtruncate() gets just the filename as path from the
caller, so obviously at this point the caller should have ensured that
it had called chdir() to the enclosing directory of the file.

Fixes bug #542.

5 years agoafpd -V show Spotlight support
HAT [Tue, 14 Jan 2014 15:03:38 +0000 (00:03 +0900)]
afpd -V show Spotlight support

5 years agoRegistering with mDNS crashed, bug #540
Ralph Boehme [Sat, 4 Jan 2014 16:08:20 +0000 (17:08 +0100)]
Registering with mDNS crashed, bug #540

5 years agoFix for incomplete AFP 3.4 commit 70e3fcd8744752d6de0b93d089f41c28dca30bde
Ralph Boehme [Thu, 9 Jan 2014 10:45:14 +0000 (11:45 +0100)]
Fix for incomplete AFP 3.4 commit 70e3fcd8744752d6de0b93d089f41c28dca30bde

Return "no such item" error when non-existing EAs are requested.

5 years agoMerge pull request #12 from hat001/avahi-workaround-3-1
slowfranklin [Thu, 19 Dec 2013 08:33:42 +0000 (00:33 -0800)]
Merge pull request #12 from hat001/avahi-workaround-3-1

Workaround for a problem which cannot be advertized by Avahi. Bug #541.

5 years agoWorkaround for a problem which cannot be advertized by Avahi. Bug #541.
HAT [Wed, 18 Dec 2013 13:50:25 +0000 (22:50 +0900)]
Workaround for a problem which cannot be advertized by Avahi. Bug #541.

5 years agoAdd support for AFP 3.4
Ralph Boehme [Thu, 7 Nov 2013 11:29:57 +0000 (12:29 +0100)]
Add support for AFP 3.4

5 years agoUpdate NEWS
Ralph Boehme [Tue, 5 Nov 2013 09:59:16 +0000 (10:59 +0100)]
Update NEWS

5 years agoFix a bug where only the first configured UAM was loaded
Ralph Boehme [Tue, 5 Nov 2013 09:57:17 +0000 (10:57 +0100)]
Fix a bug where only the first configured UAM was loaded

Caused by strtok() not being async signal safe, use strtok_r() instead.
Fixes bug #537.

5 years agoMerge branch 'branch-netatalk-3-1'
Ralph Boehme [Mon, 28 Oct 2013 15:46:51 +0000 (16:46 +0100)]
Merge branch 'branch-netatalk-3-1'

5 years agoAdd libatalk ABI for 3.1.0 and bump library version leaving room for 3.0 series
Frank Lahm [Mon, 28 Oct 2013 15:32:59 +0000 (16:32 +0100)]
Add libatalk ABI for 3.1.0 and bump library version leaving room for 3.0 series

5 years agoUpdate version and release notes for 3.1.0 release
Ralph Boehme [Mon, 28 Oct 2013 14:53:28 +0000 (15:53 +0100)]
Update version and release notes for 3.1.0 release

5 years agoAdd exit code
Frank Lahm [Mon, 28 Oct 2013 11:31:47 +0000 (12:31 +0100)]
Add exit code

Use dedicated exit code for AFP connections that were dropped
by the client right after the TCP handshake.

5 years agoMerge remote-tracking branch 'origin/branch-netatalk-3-0' into develop
Frank Lahm [Mon, 28 Oct 2013 09:34:25 +0000 (10:34 +0100)]
Merge remote-tracking branch 'origin/branch-netatalk-3-0' into develop

Conflicts:
NEWS
VERSION
doc/www/ReleaseNotes

5 years agoBuild fixes for the Kerberos UAM
Frank Lahm [Mon, 28 Oct 2013 09:06:37 +0000 (10:06 +0100)]
Build fixes for the Kerberos UAM

o missing Kerberos includes and libraries
o conditional code that handled the case when Kerberos is not available
  was c/p from etc/afp/status.c and wasn't updated for the changes introduced
  in the bugfix for #531

5 years agoUpdate version and news for 3.0.7
Frank Lahm [Mon, 28 Oct 2013 09:04:43 +0000 (10:04 +0100)]
Update version and news for 3.0.7

5 years agoMerge branch 'release-3.0.6'
Ralph Boehme [Fri, 25 Oct 2013 15:30:54 +0000 (17:30 +0200)]
Merge branch 'release-3.0.6'

5 years agoMerge branch 'branch-netatalk-3-0' into release-3.0.6
Ralph Boehme [Fri, 25 Oct 2013 15:30:18 +0000 (17:30 +0200)]
Merge branch 'branch-netatalk-3-0' into release-3.0.6

5 years agoMerge branch 'release-3.0.6'
Ralph Boehme [Fri, 25 Oct 2013 15:28:09 +0000 (17:28 +0200)]
Merge branch 'release-3.0.6'