Should place the DLLs correctly on Windows for QtRO standalone builds.

The original DLLDESTDIR stays at 'lib' when prefix_build is true; not
sure why, though.

Change-Id: I9772f2442bb6347ccb1806542b313754fa6e48e6
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Change-Id: Ib78770703e0d134793b51ec260194edb0ce9a73b
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Change-Id: I33c393a2b714cd7032411555125f38f8dcc045d2
Reviewed-by: Allen Winter <allen.winter@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Change-Id: Iba200b971055cd03a35903a4a233bfffb7b05163
Reviewed-by: Allen Winter <allen.winter@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

This actually fixes behavior when CONFIG+=force_independent is used.

Fixes:
- Now generates headers in the build dir when requested
- Also generates version/module header now

Change-Id: I6999561d25f1d9fa178c560170c63e480b4ff518
Reviewed-by: Allen Winter <allen.winter@kdab.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>

Change POD (in)equality operators from member function
to non-member, non-friend.

Change-Id: I7e8893fcbe1734f4ff97026fecd99d93b8ea4e55
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>

Change-Id: I70c60da4b8fa0736756d6f93ea0369944a41ea51
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>

POD types (before this change) didn't include equality operators,
so using them as a rep property would fail for SimpleSource, since
the auto-generated setter had "if (x != _x)" as a check.

Change-Id: I07cc8fc212a6a0206830f140fbb4bfc6d5cba137
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>

I don't know if this is the right fix, but startRow has the previous rowCount.

Change-Id: I4e00fd5a2f618162c6d76ec26430234a2992c328
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

 - rename qregexparser.h -> qrepregexparser.h
 - install parser.g
 - add a small .prf file to allow easily usage of reparser

Change-Id: Iefbe3581f4da56c290e9ba05a8923b6a0e7917bb
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Allow comments at beginning of file as well as multiple
comment lines.

Change-Id: I821b26068da5133213c357840df536c735a28d2b
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>

This will allow enums/Q_ENUM (or Q_ENUMS, pre-Qt 5.5)
to be generated by REPC so enumeration types can be
easily passed over the QtRO network.  Enums are put
in their own Q_GADGET class.

Change-Id: Ic93ed5533c9c3c6516be4c9a1b04ce188803e014
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>

In some cases (mostly for testing) it is beneficial
to only create one header file, with "extra" types
(PODs, or the upcoming Enums) generated only once.

This changed allows you to use REPC_MERGED in your
.pro files to create such a merged header.

Change-Id: I4b7995ddcc0116443377467c1e6f2d2713d64894
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>

This change moves the repc parser from hand-coded to
a QLALR parser.  It uses a special* lexer/parser header
file that combines regexes** with QLALR to make
simplified grammars.  That is, regular expressions
are the building blocks of the grammar, not more basic
tokens.  It doesn't need lex or a compatible lexer
generator.  It should support full conversion to a
complete (non-regex based) language if necessary, but
incrementally.

The goal is to make repc easier to extend by moving to
a grammar.

*qregexparser.h
**As of Qt 5.5, bootstrap (which is required for repc,
since it is a host tool) includes QRegExp, but not
QRegularExpressions.  So this version of the parser
uses QRegularExpressions in non-bootstrap mode, and
includes a bunch of #defines to work with QRegExp in
bootstrap mode.  THIS IS FRAGILE.  The regex syntax
is different for each library.  Ideally QRE will move
to bootstrap, and these hacks can be removed.

Change-Id: I86f4e764fd22702771d6b00358a0bf2da72658c0
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>

The user can use "tcp://[ip

 address]:0" address to register the server
on any interface on any free port, so we need to m_originalUrl fields
after the server listens in order to get the server port & address.

Change-Id: Iefedeb2bfb03f8bd28ea2ed00503abf2bc6c5af2
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

This reverts commit 9f2c918c

, which
was committed before it passed all unit tests.  Reverting to address
those tests.

Change-Id: If0034716e317bd6f4b1b7ed2563e7981b05cf71f
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Allen Winter <allen.winter@kdab.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>

The user can use "tcp://:0" address to register the server on any
interface on any free port, so we need to m_originalUrl fields after the
server listens in order to get the server port & address.

Change-Id: I421254928e1140dba2b050bcf7e25d45ee2bceba
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>

If the address in not supported m_factory.createServer return 0.

Change-Id: I424824f59b1a0dc89c2a2605c879676d4c4c37e0
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

This will allow QtRO to compile with Qt versions
earlier than 5.5 (the Q_GADGET change
44b3a933

).

This doesn't address the pods tests/examples.

Change-Id: I0584afd5ca03c24e5857970963cfb795378ac978
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>

Give the fetch a timeout, and make sure all of
the data is collected before the timeout.

Change-Id: I439dd2029a95a56023598ede7e884cb6ed789c3b
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>

This is a performance optimization.

Instead of passing the property name over the
wire, we only pass the property index.  Now,
because we need to handle adapters (where the
same property index can from from an adapter
as the Source object, we need a new loopup
function.  Thus the change to repc.

Change-Id: I054dc450a48778274fb475f6487c62637c9f9393
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>

This is a large change, but it is the same thing repeated.

QRemoteObjectPacket was a set of types for (de-)serializing
QtRO changes to send over the wire.  But it required
copying the changes to/from the packet before using
those changes.  This goes directly to/from QDataStream
instead.  I think it also makes the InitPacket/InitDynamic
steps a bit easier to read, and should improve the init
performance.

Change-Id: I62b1776858ff74a8532c8f2682aa94ac7248a803
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>

Remove QRemoteObjectPacket::serialize() functions in
favor of direct serialize<Type>Packet() functions.

This remove the need to copy data into the Packet type
before serializing.

Change-Id: Ic5953a6e63731b64e79ffa5bef38a5b62179dd35
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>

Pull the QRemoteObjectPacketTypeEnum definition out of the
QRemoteObjectPacket class.

This is in preparation for some performance changes.

Change-Id: Iedfa1798f18ec723327dc68a8ad039c6600a694a
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>

Instead of recursively calling onClientRead(),
use a do/while loop.  This prevents helps the
callstack for Model/View QtRO objects, where
there can be a lot of small packets.

Change-Id: Iecd8ded2ec43c713bfe2abc6ed6753a31d5c39c6
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>

Connect() was giving a runtime error on "quit"

Change-Id: I4ec205dc75dd8e9501d26e0dbf88a4ab0c8cf884
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>

We need to write a unittest for remoting with
an adapter, that case is only covered by modelview
tests right now. This needs more time, so i don't add
it here as the fix is more important.

Change-Id: I9841f707ca564ffb7fc2e86a7e0605a87637ed25
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

The buffer size wasn't initialized, which lead to undefined behavior

Change-Id: Id9d7546072d93df67a9b9fc7355f3087bb334d39
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

This avoids more reallocations.

Change-Id: Ifd4593b3481fd44426251729567c179fc577d777
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

In my opinion this should go upstream for all QDataStream
deserialization to avoid deallocation and reallocation of
memory if it is not necessary.

Change-Id: I17df81a79e834c39aabe09cee975d23000243b1b
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Change-Id: Iecee4a12850d9f200fd0505fd14a141f3b380c79
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

The default would create a QByteArray and reading that
from a datastream doesn't reuse the memory in the QByteArray.

Change-Id: Id3bb1b4ff40d5388cd537bec6a03d6f73493e10f
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Established a baseline for QDataStream transmission
and compare that against remote objects.

Change-Id: I42a50aba6864180938c20e578330087c5544315d
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Avoid allocation of QByteArray on the write side of Property change
serialization.

Cache the newed packets, this helps because the memory of the QStrings
in the packets can be reused and doesn't need a new allocation and
we save the packet allocation.

Change-Id: I3848013c222d89508b7011a60f964883c01d7dd6
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

This improves the runtime of the unittests quite
a lot, we need to check that we can get rid of waitForRegistry
too.

Change-Id: I72ca8ec46e3ee725dee729154e228601ddd9dacf
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Reduced the memory overhead by removing the allocation
of a QDataStream QByteArray pair per packet.

Change-Id: I600fdbdd86b89e94ce2a025e95c9188e8b4a1c9c
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

This requires removing the notify signal from POD
Q_PROPERTY values.

Change-Id: I12da6aed4f0babef5bb90f6817da74573070d608
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>

Having no SIGNALS (or no SLOTS) defined in a .rep file
would cause a compilation error on some compilers due
signalArgCount/signalArgTypes not being declared.

Change-Id: Ibe70f96a63775f9cdbb062f4b764b94b0947b67b
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Björn Breitmeyer <bjoern.breitmeyer@kdab.com>

If there is no connection and QtRO is build in debug mode it will assert,
if it build in release mode it will crash.

Change-Id: Icc528de233aa912c3b40acb84ac72ee9a24d62fb
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Change-Id: I5ff3c30c58e2e44e850ec560de717bcd415c4592
Reviewed-by: BogDan Vatra <bogdan@kde.org>