summaryrefslogtreecommitdiffstats
path: root/src/ByteBuffer.cpp (follow)
Commit message (Collapse)AuthorAgeFilesLines
* Overhaul endian handling in ByteBuffer and FastNBT (#5543)Ash2 days1-38/+40
| | | | | | | | | | | | | | | * Overhaul endian handling in ByteBuffer and FastNBT Rather than juggling "swapped" and "unswapped" versions of integers, different library functions, #defines, etc., simply always read everything byte-by-byte. This works regardless of host CPU endian, got optimised down to either a normal load or a byteswap on every compiler I tested - only 1 instruction on most CPU architectures. This commit introduces a "Bytes" array type to keep endian-sensitive data seperate from host data, alongside the needed C++ template machinery for it to work seamlessly. This approach is a little bit safer as well since you get length- and type-checking for most callsites. * Remove remaining references to old-style endianness conversion, remove functions themselves. --------- Co-authored-by: Alexander Harkness <me@bearbin.net>
* Protocol: Use correct calculation for delta movements (#5455)Tiger Wang2022-11-081-0/+18
| | | * Protocol: Use correct calculation for delta movements
* Remove redundant initialisationsTiger Wang2022-06-021-1/+1
| | | | The static buffers here are immediately passed in to be overwritten.
* Updated protocol functions to Vector3xx12xx12x2022-04-201-0/+18
| | | | also added support in included functions
* fix various compiler warnings (and some unwanted case statement fallthroughs)hle02022-04-121-1/+1
|
* ByteBuffer: fully defined Position handlingTiger Wang2022-01-161-12/+12
| | | Cast input ints into unsigned to make them 2s complement guaranā˜•d.
* Fix WriteXZYPosition64 and add testsBond_0092022-01-141-1/+1
|
* Fix compilation with Clang 13.0.0Bond_0092022-01-131-3/+7
| | | | | | | | Clang 13 changes: * `-Wreserved-identifier` and `-Wunused-but-set-variable` got added * the `-Wreturn-std-move-in-c++11` option was removed Builds still fail when native optimizations are enabled
* Prepare ChunkData for BlockState storage (#5105)Tiger Wang2021-03-051-0/+28
| | | | | | | | | | | | | | | | | | | | | | | * Rename ChunkData Creatable test * Add missing Y-check in RedstoneWireHandler * Remove ChunkDef.h dependency in Scoreboard * Prepare ChunkData for BlockState storage + Split chunk block, meta, block & sky light storage + Load the height map from disk - Reduce duplicated code in ChunkData - Remove saving MCSBiomes, there aren't any - Remove the allocation pool, ref #4315, #3864 * fixed build * fixed test * fixed the debug compile Co-authored-by: 12xx12 <44411062+12xx12@users.noreply.github.com>
* Fix debug macro situation (#5114)Tiger Wang2021-01-261-1/+1
| | | Use the standard NDEBUG.
* zlib -> libdeflate (#5085)Tiger Wang2021-01-111-11/+26
| | | | | | + Use libdeflate + Use std::byte * Fix passing temporary to string_view + Emulate make_unique_for_overwrite
* 1.14 connection supportTiger Wang2020-07-261-3/+43
|
* Fix building with clang 8.0 (#4346)Bond-0092019-08-111-2/+2
|
* CheckBasicStyle: Check number of empty lines between functions (#4267)peterbell102018-07-261-1/+0
| | | | Add check for number of empty lines between functions and fix the corresponding failures
* Prefer static_cast to reinterpret_cast (#4223)peterbell102018-05-021-3/+3
| | | | | | | * Change reinterpret_cast -> static_cast wherever possible * Remove more unnecessary `const_cast`s. reinterpret_casts should be avoided for the same reason as c-style casts - they don't do any type-checking. reinterpret_cast was mainly being used for down-casting in inheritance hierarchies but static_cast works just as well while also making sure that there is actually an inheritance relationship there.
* Add cUUID class (#3871)peterbell102017-08-251-19/+5
|
* Use CMake's Android generators to crosscompileTiger Wang2016-12-121-37/+0
|
* Spectators added (#2852)bibo382016-10-121-0/+33
|
* Fixed RasPi builds of unit tests.Mattes D2016-08-041-1/+3
| | | | | On RasPi with gcc 4.8.2, the asserts wouldn't compile when tests were enabled. Enforced the assumption that ASSERT code is generated only in Debug builds.
* Fix endianness issue with ODROID-C2 Boards (#3268)Alexander Harkness2016-07-201-1/+1
|
* SelfTests: Moved ByteBuffer test to a separate project.Mattes D2016-06-181-66/+0
|
* 1.9 / 1.9.2 / 1.9.3 / 1.9.4 protocol support (#3135)Pokechu222016-05-141-0/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * Semistable update to 15w31a I'm going through snapshots in a sequential order since it should make things easier, and since protocol version history is written. * Update to 15w34b protocol Also, fix an issue with the Entity Equipment packet from the past version. Clients are able to connect and do stuff! * Partially update to 15w35e Chunk data doesn't work, but the client joins. I'm waiting to do chunk data because chunk data has an incomplete format until 15w36d. * Add '/blk' debug command This command lets one see what block they are looking at, and makes figuring out what's supposed to be where in a highly broken chunk possible. * Fix CRLF normalization in CheckBasicStyle.lua Normally, this doesn't cause an issue, but when running from cygwin, it detects the CR as whitespace and creates thousands of violations for every single line. Lua, when run on windows, will normalize automatically, but when run via cygwin, it won't. The bug was simply that gsub was returning a replaced version, but not changing the parameter, so the replaced version was ignored. * Update to 15w40b This includes chunk serialization. Fully functional chunk serialization for 1.9. I'm not completely happy with the chunk serialization as-is (correct use of palettes would be great), but cuberite also doesn't skip sending empty chunks so this performance optimization should probably come later. The creation of a full buffer is suboptimal, but it's the easiest way to implement this code. * Write long-by-long rather than creating a buffer This is a bit faster and should be equivalent. However, the code still doesn't look too good. * Update to 15w41a protocol This includes the new set passengers packet, which works off of the ridden entity, not the rider. That means, among other things, that information about the previously ridden vehicle is needed when detaching. So a new method with that info was added. * Update to 15w45a * 15w51b protocol * Update to 1.9.0 protocol Closes #3067. There are still a few things that need to be worked out (picking up items, effects, particles, and most importantly inventory), but in general this should work. I'll make a few more changes tomorrow to get the rest of the protocol set up, along with 1.9.1/1.9.2 (which did make a few changes). Chunks, however, _are_ working, along with most other parts of the game (placing/breaking blocks). * Fix item pickup packet not working That was a silly mistake, but at least it was an easy one. * 1.9.2 protocol support * Fix version info found in server list ping Thus, the client reports that it can connect rather than saying that the server is out of date. This required creating separate classes for 1.9.1 and 1.9.2, unfortunately. * Fix build errors generated by clang These didn't happen in MSVC. * Add protocol19x.cpp and protocol19x.h to CMakeLists * Ignore warnings in protocol19x that are ignored in protocol18x * Document BLOCK_FACE and DIG_STATUS constants * Fix BLOCK_FACE links and add separate section for DIG_STATUS * Fix bat animation and object spawning The causes of both of these are explained in #3135, but the gist is that both were typos. * Implement Use Item packet This means that buckets, bows, fishing rods, and several other similar items now work when not looking at a block. * Handle DIG_STATUS_SWAP_ITEM_IN_HAND * Add support for spawn eggs and potions The items are transformed from the 1.9 version to the 1.8 version when reading and transformed back when sending. * Remove spammy potion debug logging * Fix wolf collar color metadata The wrong type was being used, causing several clientside issues (including the screen going black). * Fix 1.9 chunk sending in the nether The nether and the end don't send skylight. * Fix clang build errors * Fix water bottles becoming mundane potions This happened because the can become splash potion bit got set incorrectly. Water bottles and mundane potions are only differentiated by the fact that water bottles have a metadata of 0, so setting that bit made it a mundane potion. Also add missing break statements to the read item NBT switch, which would otherwise break items with custom names and also cause incorrect "Unimplemented NBT data when parsing!" logging. * Copy Protocol18x as Protocol19x Aditionally, method and class names have been swapped to clean up other diffs. This commit is only added to make the following diffs more readable; it doesn't make any other changes (beyond class names). * Make thrown potions use the correct appearence This was caused by potions now using metadata. * Add missing api doc for cSplashPotionEntity::GetItem * Fix compile error in SplashPotionEntity.cpp * Fix fix of cSplashPotionEntity API doc * Temporarilly disable fall damage particles These were causing issues in 1.9 due to the changed effect ID. * Properly send a kick packet when connecting with an invalid version This means that the client no longer waits on the server screen with no indication whatsoever. However, right now the server list ping isn't implemented for unknown versions, so it'll only load "Old" on the ping. I also added a GetVarIntSize method to cByteBuffer. This helps clean up part of the code here (and I think it could clean up other parts), but it may make sense for it to be moved elsewhere (or declared in a different way). * Handle server list pings from unrecognized versions This isn't the cleanest way of writing it (it feels odd to use ProtocolRecognizer to send packets, and the addition of m_InPingForUnrecognizedVersion feels like the wrong technique), but it works and I can't think of a better way (apart from creating a full separate protocol class to handle only the ping... which would be worse). * Use cPacketizer for the disconnect packet This also should fix clang build errors. * Add 1.9.3 / 1.9.4 support * Fix incorrect indentation in APIDesc
* Bulk clearing of whitespaceLogicParrot2016-02-051-14/+14
|
* Remove a pair of unused broken functionstycho2015-10-021-34/+0
|
* Unified the doxy-comment format.Mattes D2015-07-311-2/+0
|
* Silenced and fixed many warning messages across multiple files.Samuel Barney2015-07-291-7/+7
|
* SelfTests are registered and executed after logging framework init.Mattes D2015-06-111-7/+8
| | | | Fixes #2228.
* Added VarInt64, normalized cPacketizer datatype names.Mattes D2015-03-221-8/+53
|
* Unified cPacketizer across all protocols.Mattes D2015-03-221-4/+8
|
* Unified cByteBuffer types.Mattes D2015-03-211-31/+72
| | | | cByteBuffer now reads and writes any of the [U]Int<N> types.
* ByteBuffer: Added support for reading unsigned shorts and ints.Matyas Dolak2015-01-211-1/+31
|
* cByteBuffer: Improved SingleThreadAccessChecker performance.Mattes D2015-01-031-1/+6
| | | | But it's still poor and unusable for regular testing.
* ByteBuffer: SingleThreadAccessChecker is request-only.Mattes D2014-12-241-1/+10
| | | | It slows the server down way too much, so it can't be turned on by default.
* Fixed indent.Mattes D2014-12-031-2/+2
|
* ByteBuffer: Re-added Single-thread access checker.Mattes D2014-12-031-0/+84
|
* En masse NULL -> nullptr replaceTiger Wang2014-10-201-1/+1
|
* Use std::threadTiger Wang2014-10-191-74/+0
|
* Bug fixTiger Wang2014-09-271-17/+0
|
* cByteBuffer: Simplified ReadPosition().madmaxoft2014-09-251-6/+9
| | | | Also, by popular demand, added more comments to the code.
* cByteBuffer: Fixed position reading.madmaxoft2014-09-251-3/+3
|
* Added cByteBuffer::WriteBEUShort().madmaxoft2014-09-251-0/+13
|
* 1.8: Simplified item metadata reading.Howaner2014-09-181-17/+0
|
* Use xofts ReadPosition() code.Howaner2014-09-171-3/+6
|
* 1.8: Fixed ReadItem()Howaner2014-09-121-1/+1
|
* Fixed more 1.8 packets.Howaner2014-09-081-0/+44
|
* Assume POWER is big-endian, so it compiles.Alexander Harkness2014-08-231-1/+1
| | | | | [reference](http://nadeausoftware.com/articles/2012/02/c_c_tip_how_detect_processor_type_using_compiler_predefined_macros#POWER) We may want to come back and figure out if the processor is running in little-endian mode, but for now assume they're big-endian.
* Normalized comments.madmaxoft2014-07-171-1/+1
| | | | | This was mostly done automatically and then visually inspected for obvious errors. All //-style comments should have a 2-space separation from the code, and 1 space after the comment sign.
* Nullify deleted pointers.archshift2014-06-191-0/+1
|
* Removed assert that is now informed by type systemworktycho2014-06-111-1/+0
|
* Fixed warnings in cByteBuffer.madmaxoft2014-05-011-13/+13
|
* Merged branch 'master' into SslWrappers.madmaxoft2014-04-281-2/+0
|\
| * Removed Old AssertsTycho2014-04-271-2/+0
| | | | | | | | | | Removed asserts about non-negative numbers on what are now unsigned types Fixes CID 43608
* | cByteBuffer uses void * instead of char * in data interface.madmaxoft2014-04-241-4/+5
|/ | | | This makes it compatible with any pointer type.
* Changed cByteBuffer constructor to take a size_t instead of int.madmaxoft2014-04-241-1/+1
|
* Fixed xofts issuesTycho2014-03-141-1/+1
|
* Fixed more Format issuesTycho2014-03-111-1/+1
|
* Fixed test assertsTycho2014-03-101-3/+0
|
* Fixed assertTycho2014-03-101-14/+17
|
* Fixed Alignment issue in ByteBufferTycho2014-03-101-1/+1
|
* FIxed int in testTycho2014-03-091-1/+1
|
* Added staticTycho2014-03-081-1/+1
|
* Added tons more asserts to bytebufferTycho2014-03-081-7/+17
|
* Enabled self test of bytebufferTycho2014-03-081-1/+1
|
* Fixed issues with int vs size_t and a few other warningsTycho2014-03-081-10/+10
|
* Actually Fixed ByteBufferTycho2014-03-081-3/+3
|
* WarningsTycho2014-03-071-5/+5
|
* cByteBuffer has more self-tests.madmaxoft2014-01-261-6/+29
|
* removed unneccisary castTycho Bickerstaff2013-12-311-1/+2
|
* fixed warnings in bytebuffer.cppTycho Bickerstaff2013-12-311-3/+3
|
* Fixed threading issues affecting cProtocol172.madmaxoft2013-12-131-0/+22
|
* Hopefully fixed the sparc not being detectedAlexander Harkness2013-12-041-1/+3
|
* Fixed an off-by-one error in cByteBuffer.madmaxoft2013-11-291-1/+1
| | | | Could have caused random protocol errors and maybe even crashes.
* Merge remote-tracking branch 'origin/master' into foldermove2Alexander Harkness2013-11-241-0/+54
| | | | | Conflicts: GNUmakefile
* Moved source to srcAlexander Harkness2013-11-241-0/+787