sdjournal: add test for incorrect Read() length

sdjournal: fix incomplete reads in JournalReader.Read()

sdjournal: fix Writer call

sdjournal: Break the loop if error happens during Write().

This patch fixes two issues in JournalReader.Read():

1. The return value of copy() was ignored and JournalReader.Read() was
returning the wrong amount of bytes copied.

2. JournalReader implements the Reader interface so it must be able to read a
specific amount of bytes. Journald on the other hand only allows us to read
full entries of arbitrary size (without byte granularity). JournalReader
therefore needs to internally buffer entries that don't fit in the read buffer.
Callers should keep calling until 0 and/or an error is returned.

Fixes https://github.com/coreos/go-systemd/issues/172

sdjournal: don't close libsystemd on Close()

sdjournal: add GetEntry method to retrieve all fields

As GetEntry outputs every field in a string keyed map, it eases access to a concrete field without hardcoding it.

Ref: https://www.freedesktop.org/software/systemd/man/systemd.journal-fields.html

Implements the same semantics as the JOURNAL_FOREACH_DATA_RETVAL macro in journal-internal.h and produces a *JournalEntry with every field with its corresponding data in a map[string]string. Also it includes address fields like realtime (__REALTIME_TIMESTAMP), monotonic (__MONOTONIC_TIMESTAMP and cursor (__CURSOR).

It mimics more or less `journalctl -o json`.

Fixes #142
Supersedes #136

There is a symbol cache (map from function names to their addresses) so
we don't need to call dlsym everytime we call a function from libsystemd
if it was already called before.

Every time we open a new Journal, we call dlopen(), and every time we
close it, we call dlclose(). Since the map was global and we weren't
clearing it up after closing Journal, we were reusing the addresses from
the previous invocation. This is wrong because libsystemd can end up in
a different address the second time we dlopen() it.

Instead, just don't call dlclose() when we close the Journal so
libsystemd remains in the process address space for the whole life of
it. This means the symbol cache will be shared between different Journal
instances.

According to dlopen(3):

 If the same shared object is loaded again with dlopen(), the
 same object handle is returned. The dynamic linker maintains reference
 counts for object handles, so a dynamically loaded shared object is
 not deallocated until dlclose() has been called on it as many times
 as dlopen() has succeeded on it.

So we don't worry about calling dlopen() on an already opened library.

sdjournal: add test for NewJournalFromDir()

dbus: add support for new ListUnits* methods

* Added getUnitStatus and getUnitFile functions for test
* Added t.Skip for methods which are not available
* Increased verbosity to show skipped tests

Simple test to exercise main code path and error handling of
`NewJournalFromDir()`.

implemented new dbus methods:
* ListUnitsFiltered
* ListUnitsByPatterns
* ListUnitsByNames
* ListUnitFilesByPatterns

refactored legacy ListUnits and ListUnitFiles functions

sdjournal/read: fix comment related to mutually exclusive options

sdjournal: add support for cursors (get&seek&test)

sdjournal: export SYSLOG_IDENTIFIER

rkt needs this because there's a race where short-lived non-root
processes make the journal unable to get all their information (i.e. the
SYSTEMD_UNIT). We work around it by using SYSLOG_IDENTIFIER.

By providing a callback to NewConnection, users can now initialise a session over a transport of their choosing

sdjournal: add ability to rewind JournalReader

This allows users of the JournalReader to rewind to the beginning of the
Journal.

Also add a missing docstring to JournalReader.Close

Simple wrapper around `sd_journal_seek_head`

Add test for dbus/methods.go:GetServiceProperty function

Add method to get service property

sdjournal: dlopen libsystemd instead of linking against it

Add // +build ignore to examples

This means that go get github.com/coreos/go-systemd/... can succeed.

Fixes #156

add _TRANSPORT as a filter field

We don't want to leak it.

Maps are not safe for concurrent use in go.

Systemd functions return negative errno-like errors, wrap those with
syscall.Errno() so we can get descriptive error messages instead of
numbers.

This allows not having to link users of the library to libsystemd,
detect at runtime if it exists or not and react accordingly.