LLVM 15.0.7 is now available. This release is a bug-fix release and is API and ABI compatible with 15.0.0. If there are no major issues found with this release, then it will be the last 15.0.x release.
LLVM 15.0.5 is now available. This release is a bug-fix release and is API and ABI compatible with 15.0.0. If there are no major issues found with this release, then it will be the last 15.0.x release.
LLVM 15.0.4
Changes to the LLVM IR
LLVM now uses opaque pointers. This means that different pointer types like i8*, i32* or void()** are now represented as a single ptr type. See the linked document for migration instructions.
Renamed llvm.experimental.vector.extract intrinsic to llvm.vector.extract.
Renamed llvm.experimental.vector.insert intrinsic to llvm.vector.insert.
The constant expression variants of the following instructions have been removed:
extractvalue
insertvalue
udiv
sdiv
urem
srem
fadd
fsub
fmul
fdiv
frem
Added the support for fmax and fmin in atomicrmw instruction. The comparison is expected to match the behavior of llvm.maxnum.* and llvm.minnum.* respectively.
callbr instructions no longer use blockaddress arguments for labels. Instead, label constraints starting with ! refer directly to entries in the callbr indirect destination list.
Changes to building LLVM
Omitting CMAKE_BUILD_TYPE when using a single configuration generator is now an error. You now have to pass -DCMAKE_BUILD_TYPE=<type> in order to configure LLVM. This is done to help new users of LLVM select the correct type: since building LLVM in Debug mode is very resource intensive, we want to make sure that new users make the choice that lines up with their usage. We have also improved documentation around this setting that should help new users. You can find this documentation here.
Changes to Loop Optimizations
Loop interchange legality and cost model improvements
Changes to the AMDGPU Backend
8 and 16-bit atomic loads and stores are now supported
Changes to the ARM Backend
Added support for the Armv9-A, Armv9.1-A and Armv9.2-A architectures.
Added support for the Armv8.1-M PACBTI-M extension.
Added support for the Armv9-A, Armv9.1-A and Armv9.2-A architectures.
Added support for the Armv8.1-M PACBTI-M extension.
Removed the deprecation of ARMv8-A T32 Complex IT blocks. No deprecation warnings will be generated and -mrestrict-it is now always off by default. Previously it was on by default for Armv8 and off for all other architecture versions.
Added a pass to workaround Cortex-A57 Erratum 1742098 and Cortex-A72 Erratum 1655431. This is enabled by default when targeting either CPU.
Implemented generation of Windows SEH unwind information.
Switched the MinGW target to use SEH instead of DWARF for unwind information.
Added support for the Cortex-M85 CPU.
Added support for a new -mframe-chain=(none|aapcs|aapcs+leaf) command-line option, which controls the generation of AAPCS-compliant Frame Records.
Changes to the DirectX Backend
DirectX has been added as an experimental target. Specify -DLLVM_EXPERIMENTAL_TARGETS_TO_BUILD=DirectX in your CMake configuration to enable it. The target is not packaged in pre-built binaries.
The DirectX backend supports the dxil architecture which is based on LLVM 3.6 IR encoded as bitcode and is the format used for DirectX GPU Shader programs.
Changes to the PowerPC Backend
Common PowerPC improvements: * Add a new post instruction selection pass to generate CTR loops. * Add SSE4 and BMI compatible intrinsics implementation. * Supported 16-byte lock free atomics on PowerPC8 and up. * Supported atomic load/store for pointer types. * Supported stack size larger than 2G * Add __builtin_min/__builtin_max/__abs builtins. * Code generation improvements for splat load/vector shuffle/mulli, etc. * Emit VSX instructions for vector loads and stores regardless of alignment. * The mcpu=future has its own ISA now (FutureISA). * Added the ppc-set-dscr option to set the Data Stream Control Register (DSCR). * Bug fixes.
AIX improvements: * Supported 64 bit XCOFF for integrated-as path. * Supported X86-compatible vector intrinsics. * Program code csect default alignment now is 32-byte. * Supported auxiliary header in integrated-as path. * Improved alias symbol handling.
Changes to the RISC-V Backend
A RISCVRedundantCopyElimination pass was added to remove unnecessary zero copies.
A RISC-V specific CodeGenPrepare pass was added.
The machine outliner was enabled by default for RISC-V at -Oz. Additionally, the newly introduced RISCVMakeCompressible pass will make modify instructions prior to emission at -Oz in order to increase opportunities for the compression with the RISC-V C extension.
Various bug fixes and improvements to code generation for the RISC-V vector extensions.
Various improvements were made to RISC-V specific optimisation passes such as RISCVSExtWRemoval and RISCVMergeBaseOffset.
llc now computes the target ABI based on the target architecture using the same logic as Clang if not explicit ABI is given.
generic is now recognized as a valid CPU name and is mapped to generic-rv32 or generic-rv64 depending on the target triple.
Support for the experimental Zvfh extension was added, enabling half-precision floating point in vectors.
Support for the Zihintpause (Pause Hint) extension.
Assembler and disassembler support for the Zfinx and Zdinx (float / double in integer register) extensions.
Assembler and disassembler support for the Zicbom, Zicboz, and Zicbop cache management operation extensions.
Support for the Zmmul extension (a subextension of the M extension, adding multiplication instructions only).
Assembler and disassembler support for the hypervisor extension and for the Sinval supervisor memory-management extension.
Changes to the X86 Backend
Support half type on SSE2 and above targets following X86 psABI.
Support rdpru instruction on Zen2 and above targets.
During this release, half type has an ABI breaking change to provide the support for the ABI of _Float16 type on SSE2 and above following X86 psABI. (D107082)
The change may affect the current use of half includes (but is not limited to):
Frontends generating half type in function passing and/or returning arguments.
Downstream runtimes providing any half conversion builtins assuming the old ABI.
Projects built with LLVM 15.0 but using early versions of compiler-rt.
When you find failures with half type, check the calling conversion of the code and switch it to the new ABI.
Changes to the LLVM tools
(Experimental) llvm-symbolizer now has --filter-markup to filter Symbolizer Markup into human-readable form.
llvm-objcopy has removed support for the legacy zlib-gnu format.
llvm-objcopy now allows --set-section-flags src=... --rename-section src=tst. --add-section=.foo1=... --rename-section=.foo1=.foo2 now adds .foo1 instead of .foo2.
New features supported on AIX for llvm-ar:
AIX big-format archive write operation (D123949)
A new object mode option, -X , to specify the type of object file llvm-ar should operate upon (D127864)
Read global symbols of AIX big archive (D124865)
New options supported for llvm-nm:
-X, to specify the type of object file that llvm-nm should examine (D118193)
--export-symbols, to create a list of symbols to export (D112735)
The LLVM gold plugin now ignores bitcode from the .llvmbc section of ELF files when doing LTO. https://github.com/llvm/llvm-project/issues/47216
llvm-objcopy now supports 32 bit XCOFF.
llvm-objdump: improved assembly printing for XCOFF.
llc now parses code-model attribute from input file.
Changes to LLDB
The “memory region” command now has a “–all” option to list all memory regions (including unmapped ranges). This is the equivalent of using address 0 then repeating the command until all regions have been listed.
Added “–show-tags” option to the “memory find” command. This is off by default. When enabled, if the target value is found in tagged memory, the tags for that memory will be shown inline with the memory contents.
Various memory related parts of LLDB have been updated to handle non-address bits (such as AArch64 pointer signatures):
“memory read”, “memory write” and “memory find” can now be used with addresses with non-address bits.
All the read and write memory methods on SBProccess and SBTarget can be used with addreses with non-address bits.
When printing a pointer expression, LLDB can now dereference the result even if it has non-address bits.
The memory cache now ignores non-address bits when looking up memory locations. This prevents us reading locations multiple times, or not writing out new values if the addresses have different non-address bits.
LLDB now supports reading memory tags from AArch64 Linux core files.
LLDB now supports the gnu debuglink section for reading debug information from a separate file on Windows
LLDB now allows selecting the C++ ABI to use on Windows (between Itanium, used for MingW, and MSVC) via the plugin.object-file.pe-coff.abi setting. In Windows builds of LLDB, this defaults to the style used for LLVM’s default target.
Other Changes
The code for the LLVM Visual Studio integration has been removed. This had been obsolete and abandoned since Visual Studio started including an integration by default in 2019.
Added the unwinder, personality, and helper functions for exception handling on AIX. (D100132) (D100504)
PGO on AIX: A new implementation that requires linker support (__start_SECTION/__stop_SECTION symbols) available on AIX 7.2 TL5 SP4 and AIX 7.3 TL0 SP2.
Tested on NetBSD-current and 9.1/amd64.
Also update packages tightly coupled to llvm, if new versions are
available.
Mark creduce, include-what-you-use, and zig broken (waiting for a
new release).
Changes:
Non-comprehensive list of changes in this release
Windows Control-flow Enforcement Technology: the -ehcontguard
option now emits valid unwind entrypoints which are validated
when the context is being set during exception handling.
Flang is now included in the binary packages released by LLVM.
The debuginfo-test project has been renamed cross-project-tests
and is now intended for testing components from multiple
projects, not just debug information. The new “cross-project-tests”
name replaces “debuginfo-test” in LLVM_ENABLE_PROJECTS, and a
new check-cross-project-tests target has been added for running
all tests in the project. The pre-existing check-debuginfo-
test target remains for running just the debug information
tests. (D95339 and D96513)
Changes to the LLVM IR
The inalloca attribute now has a mandatory type field, similar
to byval and sret.
The opaque pointer type ptr has been introduced. It is still
in the process of being worked on and should not be used yet.
Using the legacy pass manager for the optimization pipeline is
deprecated and will be removed after LLVM 14. In the meantime,
only minimal effort will be made to maintain the legacy pass
manager for the optimization pipeline.
Changes to building LLVM
The build system now supports building multiple distributions,
so that you can e.g. have one distribution containing just
tools and another for libraries (to enable development). See
Multi-distribution configurations for details.
Changes to the AArch64 Backend
Introduced assembly support for Armv9-A’s Realm Management
Extension (RME) and Scalable Matrix Extension (SME).
Produce proper cross-section relative relocations on COFF
Fixed the calling convention on Windows for variadic functions
involving floats in the fixed arguments
Changes to the ARM Backend
Produce proper cross-section relative relocations on COFF
Changes to the Hexagon Target
The Hexagon target now supports V68/HVX ISA.
Changes to the C API
The C API functions LLVMGetAlignment and LLVMSetAlignment now
allow changing alignment on atomicrmw and cmpxchg instructions
A new entry LLVMDIArgListMetadataKind was added to the
LLVMMetadataKind enum, representing DIArgLists (D88175)
Type attributes have been added to LLVM-C, introducing
LLVMCreateTypeAttribute, LLVMGetTypeAttributeValue and
LLVMIsTypeAttribute. (D977763’)
The LTO_API_VERSION was bumped to 28, introducing a new function
lto_set_debug_options for parsing LTO debug options (D92611)
LLVMJITTargetSymbolFlags was renamed to LLVMJITSymbolTargetFlags
(rG8d718a0bff73af066675a6258c01307937c33cf9)
The C API received support for creating custom ORCv2
MaterializationUnits and APIs to retrieve an LLJIT instance’s
linking layers. An ABI breaking change for LLVMOrcAbsoluteSymbols
was introduced to make it consistent with
LLVMOrcCreateCustomMaterializationUnit.
(rGc8fc5e3ba942057d6c4cdcd1faeae69a28e7b671)
The C API received support for adding ORCv2 object buffers
directly to an object layer.
(rG7b73cd684a8d5fb44d34064200f10e2723085c33)
A breaking change to LLVMGetInlineAsm was introduced, adding
a ninth argument LLVMBool CanThrow (D95745)
The C API received support for calling into the new pass manager.
(D102136)
The C API function LLVMIntrinsicCopyOverloadedName has been
deprecated. Please migrate to LLVMIntrinsicCopyOverloadedName2
which takes an extra module argument and which also handles
unnamed types. (D99173)
The C API received support for dumping objects from ORCv2
(rGcec8e69f01c3374cb38c6683058381b96fab8f89)
A breaking change to LLVMOrcObjectTransformLayerTransformFunction
was introduced, changing the order of the function pointer’s
arguments. (rG8962c68ad007a525f9daa987c99eda57e0d0069a)
The C API received support for accessing utilities from the
LLJIT IRTransformLayer and ThreadSafeModule classes. (D103855)
The C API received support for creating lazy-export
MaterializationUnits (D104672)
The C API function LLVMPassBuilderOptionsSetCoroutines was
removed because couroutine passes have been enabled by default.
(D105877)
comdat noduplicates was renamed to comdat nodeduplicate and as
a result, LLVMNoDuplicatesComdatSelectionKind was renamed to
LLVMNoDeduplicateComdatSelectionKind. (D106319)
Changes to the FastISel infrastructure
FastISel no longer tracks killed registers, and instead leaves
this to the register allocator. This means that hasTrivialKill()
is removed, as well as the OpNIsKill parameters to the fastEmit_*()
family of functions.
Changes to the LLVM tools
The options --build-id-link-{dir,input,output} have been deleted.
(D96310)
Support for in-order processors has been added to llvm-mca.
(D94928)
llvm-objdump supports -M {att,intel} now. --x86-asm-syntax is
a deprecated internal option which will be removed in LLVM
14.0.0. (D101695)
The llvm-readobj short aliases -s (previously --sections) and
-t (previously --syms) have been changed to --syms and
--section-details respectively, to match llvm-readelf. (D105055)
The llvm-nm short aliases -M (--print-armap), -U (--defined-only),
and -W (--no-weak) are now deprecated. Use the long form versions
instead. The alias --just-symbol-name is now deprecated in
favor of --format=just-symbols and -j. (D105330)
In lli the default JIT engine switched from MCJIT (-jit-kind=mcjit)
to ORC (-jit-kind=orc). (D98931)
llvm-rc got support for invoking Clang to preprocess its input.
(D100755)
llvm-rc got a GNU windres compatible frontend, llvm-windres.
(D100756)
llvm-ml has improved compatibility with MS ml.exe, managing to
assemble more asm files.
Changes to LLDB
LLDB executable is now included in pre-built LLVM binaries.
LLDB now includes full featured support for AArch64 SVE register
access.
LLDB now supports AArch64 Pointer Authentication, allowing
stack unwind with signed return address.
LLDB now supports debugging programs on AArch64 Linux that use
memory tagging (MTE).
Added memory tag read and memory tag write commands.
The memory region command will note when a region has memory
tagging enabled.
Synchronous and asynchronous tag faults are recognised.
Synchronous tag faults have memory tag annotations in addition
to the usual fault address.
Improvements to clangd
Performance
clangd’s memory usage is significantly reduced on most Linux systems. In particular, memory usage should not increase dramatically over time.
The standard allocator on most systems is glibc’s ptmalloc2, and it creates disproportionately large heaps when handling clangd’s allocation patterns. By default, clangd will now periodically call malloc_trim to release free pages on glibc systems.
Users of other allocators (such as jemalloc or tcmalloc) on glibc systems can disable this using --malloc_trim=0 or the CMake flag -DCLANGD_MALLOC_TRIM=0.
Added the $/memoryUsage request: an LSP extension. This provides a breakdown of the memory clangd thinks it is using (excluding malloc overhead etc). The clangd VSCode extension supports showing the memory usage tree.
Parsing and selection
Improved navigation of broken code in C using Recovery AST. (This has been enabled for C++ since clangd 11).
Types are understood more often in broken code. (This is the first release where Recovery AST preserves speculated types).
Heuristic resolution for dependent names in templates.
Code completion
Higher priority for symbols that were already used in this file, and symbols from namespaces mentioned in this file. (Estimated 3% accuracy improvement)
Introduced a ranking algorithm trained on snippets from a large C++ codebase. Use the flag --ranking-model=decision_forest to try this (Estimated 6% accuracy improvement). This mode is likely to become the default in future.
Note: this is a generic model, not specialized for your code. clangd does not collect any data from your code to train code completion.
Signature help works with functions with template-dependent parameter types.
Go to definition
Selecting an auto or decltype keyword will attempt to navigate to a definition of the deduced type.
Improved handling of aliases: navigate to the underlying entity more often.
Better understanding of declaration vs definition for Objective-C classes and protocols.
Selecting a pure-virtual method shows its overrides.
Find references
Indexes are smarter about not returning stale references when code is deleted.
References in implementation files are always indexed, so results should be more complete.
Find-references on a virtual method shows references to overridden methods.
New navigation features
Call hierarchy (textDocument/callHierarchy) is supported. Only incoming calls are available.
Go to implementation (textDocument/implementation) is supported on abstract classes, and on virtual methods.
Symbol search (workspace/symbol) queries may be partially qualified. That is, typing b::Foo will match the symbol a:🅱️:c::Foo.
Refactoring
New refactoring: populate switch statement with cases. (This acts as a fix for the -Wswitch-enum warning).
Renaming templates is supported, and many other complex cases were fixed.
Attempting to rename to an invalid or conflicting name can produce an error message rather than broken code. (Not all cases are detected!)
The accuracy of many code actions has been improved.
Hover
Hovers for auto and decltype show the type in the same style as other hovers. this is also now supported.
Displayed type names are more consistent and idiomatic.
Semantic highlighting
Inactive preprocessor regions (#ifdef) are highlighted as comments.
clangd 12 is the last release with support for the non-standard textDocument/semanticHighlights notification. Clients sholud migrate to the textDocument/semanticTokens request added in LSP 3.16.
Remote index (alpha)
clangd can now connect to a remote index server instead of building a project index locally. This saves resources in large codebases that are slow to index.
The server program is clangd-index-server, and it consumes index files produced by clangd-indexer.
This feature requires clangd to be built with the CMake flag -DCLANGD_ENABLE_REMOTE=On, which requires GRPC libraries and is not enabled by default. Unofficial releases of the remote-index-enabled client and server tools are at https://github.com/clangd/clangd/releases
Large projects can deploy a shared server, and check in a .clangd file to enable it (in the Index.External section). We hope to provide such a server for llvm-project itself in the near future.
Configuration
Static and remote indexes can be configured in the Index.External section. Different static indexes can now be used for different files. (Obsoletes the flag --index-file).
Diagnostics can be filtered or suppressed in the Diagnostics section.
Clang-tidy checks can be enabled/disabled in the Diagnostics.ClangTidy section. (Obsoletes the flag --clang-tidy-checks).
The compilation database directory can be configured in the CompileFlags section. Different compilation databases can now be specified for different files. (Obsoletes the flag --compile-commands-dir).
Errors in loaded configuration files are published as LSP diagnostics, and so should be shown in your editor.
Full reference of configuration options
System integration
Changes to compile_commands.json and compile_flags.txt will take effect the next time a file is parsed, without restarting clangd.
clangd --check=<filename> can be run on the command-line to simulate opening a file without actually using an editor. This can be useful to reproduce crashes or aother problems.
Various fixes to handle filenames correctly (and case-insensitively) on windows.
If incoming LSP messages are malformed, the logs now contain details.
Miscellaneous
“Show AST” request (textDocument/ast) added as an LSP extension. This displays a simplified view of the clang AST for selected code. The clangd VSCode extension supports this.
clangd should no longer crash while loading old or corrupt index files.
The flags --index, --recovery-ast and -suggest-missing-includes have been retired. These features are now always enabled.
Too many stability and correctness fixes to mention.
What’s New in Extra Clang Tools 10.0.0?
Some of the major new features and improvements to Extra Clang Tools are listed here. Generic improvements to Extra Clang Tools as a whole or to its underlying infrastructure are described first, followed by tool-specific sections.
Improvements to clangd
clangd documentation is now found at https://clangd.llvm.org/
Go-to-definition, hover, find-references etc use a new mechanism to identify what is under the cursor, which is (hopefully) more consistent and accurate.
clangd should be able to reliably locate the standard library/SDK on macOS.
Shutdown more cleanly on receiving a signal. In particular temporary PCH files should be cleaned up.
Find references now works on macros.
clangd can be more easily used remotely or in a docker container.
The --path-mappings flag translates between local and remote paths.
Experimental support for renaming across files (behind the --cross-file-rename flag).
Hover now exposes more information, including the type of symbols and the value of constant expressions.
Go to definition now works in dependent code in more cases, by assuming the primary template is used.
Better recovery and reporting when the compile command for a file can’t be fully parsed.
Switch header/source (an extension) now uses index information in addition to filename heuristics, and is much more robust.
Semantic selection (expand/contract selection) is supported.
Semantic highlighting is more robust, highlights more types of tokens, and as an extension provides information about inactive preprocessor regions.
Code completion results now include an extension field score.
This allows clients to incorporate clangd quality signals when re-ranking code completion after client-side fuzzy-matching.
New refactorings: define function out-of-line, define function in-line, extract function, remove using namespace directive, localize Objective-C string.
Bug fixes and performance improvements :-)
Improvements to clang-doc
clang-doc now generates documentation in HTML format.
Improvements to clang-tidy
New checks
New bugprone-bad-signal-to-kill-thread check.
Finds pthread_kill function calls when a thread is terminated by raising SIGTERM signal.
New bugprone-dynamic-static-initializers check.
Finds instances where variables with static storage are initialized dynamically in header files.
New bugprone-infinite-loop check.
Finds obvious infinite loops (loops where the condition variable is not changed at all).
New bugprone-not-null-terminated-result check
Finds function calls where it is possible to cause a not null-terminated result.
New bugprone-signed-char-misuse check.
Finds signed char to integer conversions which might indicate a programming error.
New cert-mem57-cpp check.
Checks if an object of type with extended alignment is allocated by using the default operator new.
New cert-oop58-cpp check.
Finds assignments to the copied object and its direct or indirect members in copy constructors and copy assignment operators.
New cppcoreguidelines-init-variables check.
Checks whether there are local variables that are declared without an initial value.
New darwin-dispatch-once-nonstatic check.
Finds declarations of dispatch_once_t variables without static or global storage.
New google-upgrade-googletest-case check.
Finds uses of deprecated Googletest APIs with names containing case and replaces them with equivalent APIs with suite.
New linuxkernel-must-use-errs check.
Checks Linux kernel code to see if it uses the results from the functions in linux/err.h.
New llvm-prefer-register-over-unsigned check.
Finds historical use of unsigned to hold vregs and physregs and rewrites them to use Register
New objc-missing-hash check.
Finds Objective-C implementations that implement -isEqual: without also appropriately implementing -hash.
New performance-no-automatic-move check.
Finds local variables that cannot be automatically moved due to constness.
New performance-trivially-destructible check.
Finds types that could be made trivially-destructible by removing out-of-line defaulted destructor declarations.
New readability-make-member-function-const check.
Finds non-static member functions that can be made const because the functions don’t use this in a non-const way.
New readability-qualified-auto check.
Adds pointer and const qualifications to auto-typed variables that are deduced to pointers and const pointers.
New readability-redundant-access-specifiers check.
Finds classes, structs, and unions that contain redundant member access specifiers.
New aliases
New alias cert-pos44-c to bugprone-bad-signal-to-kill-thread was added.
New alias llvm-qualified-auto to readability-qualified-auto was added.
Changes in existing checks
Improved bugprone-posix-return check.
Now also checks if any calls to pthread_* functions expect negative return values.
Improved hicpp-signed-bitwise check.
The check now supports the IgnorePositiveIntegerLiterals option.
Improved modernize-avoid-bind check.
The check now supports supports diagnosing and fixing arbitrary callables instead of only simple free functions. The PermissiveParameterList option has also been added to address situations where the existing fix-it logic would sometimes generate code that no longer compiles.
The modernize-use-equals-default fix no longer adds semicolons where they would be redundant.
Improved modernize-use-override check.
The check now supports the AllowOverrideAndFinal option to eliminate conflicts with gcc -Wsuggest-override or gcc -Werror=suggest-override.
The modernize-use-using check now converts typedefs containing struct definitions and multiple comma-separated types.
Improved readability-magic-numbers check.
The check now supports the IgnoreBitFieldsWidths option to suppress the warning for numbers used to specify bit field widths.
The check was updated to eliminate some false positives (such as using class enumeration as non-type template parameters, or the synthetically computed length of a static user string literal.)
Improved readability-redundant-member-init check.
The check now supports the IgnoreBaseInCopyConstructors option to avoid “base class ‘Foo’ should be explicitly initialized in the copy constructor” warnings or errors with gcc -Wextra or gcc -Werror=extra.
The readability-redundant-string-init check now supports a StringNames option enabling its application to custom string classes.
Renamed checks
The ‘objc-avoid-spinlock’ check was renamed to darwin-avoid-spinlock
9.0.0:
Improvements to clangd
* Background indexing is on by default
When using clangd, it will build an index of your code base (all files listed in your compile database). This index enables go-to-definition, find-references, and even code completion to find symbols across your project.
This feature can consume a lot of CPU. It can be disabled using the --background-index=false flag, and respects -j to use fewer threads. The index is written to .clangd/index in the project root.
* Contextual code actions
Extract variable, expand auto, expand macro, convert string to raw string. More to come in the future!
* Clang-tidy warnings are available
These will be produced for projects that have a .clang-tidy file in their source tree, as described in the clang-tidy documentation.
* Improved diagnostics
Errors from headers are now shown (on the #including line). The message now indicates if fixes are available. Navigation between errors and associated notes is improved (for editors that support Diagnostic.relatedInformation).
* Suggested includes
When a class or other name is not found, clangd may suggest to fix this by adding the corresponding #include directive.
* Semantic highlighting
clangd can push syntax information to the editor, allowing it to highlight e.g. member variables differently from locals. (requires editor support)
This implements the proposed protocol from https://github.com/microsoft/vscode-languageserver-node/pull/367
* Type hierachy
Navigation to base/derived types is possible in editors that support the proposed protocol from https://github.com/microsoft/vscode-languageserver-node/pull/426
* Improvements to include insertion
Only headers with #include-guards will be inserted, and the feature can be disabled with the --header-insertion=never flag.
Standard library headers should now be inserted more accurately, particularly for C++ other than libstdc++, and for the C standard library.
* Code completion
Overloads are bundled into a single completion item by default. (for editors that support signature-help).
Redundant const/non-const overloads are no longer shown.
Before clangd is warmed up (during preamble build), limited identifier- and index-based code completion is available.
* Format-on-type
A new implementation of format-on-type is triggered by hitting enter: it attempts to reformat the previous line and reindent the new line. (Requires editor support).
* Toolchain header detection
Projects that use an embedded gcc toolchain may only work when used with the corresponding standard library. clangd can now query the toolchain to find these headers. The compilation database must correctly specify this toolchain, and the --query-driver=/path/to/toolchain/bin/* flag must be passed to clangd.
* Miscellaneous improvements
Hover now produces richer Markdown-formatted text (for supported editors).
Rename is safer and more helpful, though is still within one file only.
Files without extensions (e.g. C++ standard library) are handled better.
clangd can understand offsets in UTF-8 or UTF-32 through command-line flags or protocol extensions. (Useful with editors/platforms that don’t speak UTF-16).
Editors that support edits near the cursor in code-completion can set the textDocument.completion.editsNearCursor capability to true, and clangd will provide completions that correct . to ->, and vice-versa.
Improvements to clang-tidy
New OpenMP module.
New abseil-duration-addition check.
New abseil-duration-conversion-cast check.
New abseil-duration-unnecessary-conversion check.
New abseil-time-comparison check.
New abseil-time-subtraction check.
New android-cloexec-pipe check.
New android-cloexec-pipe2 check.
New bugprone-branch-clone check.
New bugprone-posix-return check.
New bugprone-unhandled-self-assignment check.
New fuchsia-default-arguments-calls check.
New fuchsia-default-arguments-declarations check.
New google-objc-avoid-nsobject-new check.
New google-readability-avoid-underscore-in-googletest-name check.
New llvm-prefer-isa-or-dyn-cast-in-conditionals check.
New modernize-use-trailing-return-type check.
New objc-super-self check.
New openmp-exception-escape check.
New openmp-use-default-none check.
New readability-convert-member-functions-to-static check.
New alias cert-oop54-cpp to bugprone-unhandled-self-assignment was added.
New alias cppcoreguidelines-explicit-virtual-functions to modernize-use-override was added.
Added UseAssignment option to cppcoreguidelines-pro-type-member-init
The fuchsia-default-arguments check has been removed.
The google-runtime-int check has been disabled in Objective-C++.
The modernize-use-override now supports OverrideSpelling and FinalSpelling options.
The misc-throw-by-value-catch-by-reference now supports WarnOnLargeObject and MaxSize options to warn on any large trivial object caught by value.
The Acronyms and IncludeDefaultAcronyms options for the objc-property-declaration check have been removed.
Improvements to pp-trace
* Added a new option -callbacks to filter preprocessor callbacks. It replaces the -ignore option.
LLVM 8.0.1 is now available! Download it now, or read the release notes.
This release contains bug-fixes for the LLVM 8.0.0 release. This
release is API and ABI compatible with 8.0.0.
Clang Tools 8.0.0:
Improvements to clangd
clangd now adds namespace qualifiers in code completion, for example, if you type “vec”, the list of completions will include “std::vector”.
When a global index is available, clangd will use it to augment the results of “go to definition” and “find references” queries. Global index also enables global code completion, which suggests symbols that are not imported in the current file and automatically inserts the missing #include directives.
clangd stores the symbol index on disk in a new compact binary serialization format. It is 10x more compact than YAML and 40% more compact than gzipped YAML.
clangd has a new efficient symbol index suitable for complex and fuzzy queries and large code bases (e.g., LLVM, Chromium). This index is used for code completion, go to definition, and cross-references. The architecture of the index allows for complex and fuzzy retrieval criteria and sophisticated scoring.
clangd has a new LSP extension that communicates information about activity on clangd’s per-file worker thread. This information can be displayed to users to let them know that the language server is busy with something. For example, in clangd, building the AST blocks many other operations.
clangd has a new LSP extension that allows the client to supply the compilation commands over LSP, instead of finding compile_commands.json on disk.
clangd has a new LSP extension that allows the client to request fixes to be sent together with diagnostics, instead of asynchronously.
clangd has a new LSP extension that allows the client to resolve a symbol in a light-weight manner, without retrieving further information (like definition location, which may require consulting an index).
Improvements to clang-query
A new command line parameter --preload was added to run commands from a file and then start the interactive interpreter.
The command q can was added as an alias for quit to exit the clang-query interpreter.
It is now possible to bind to named values (the result of let expressions).
It is now possible to write comments in clang-query code. This is primarily useful when using script-mode. Comments are all content following the # character on a line.
The new set print-matcher true command now causes clang-query to print the evaluated matcher together with the resulting bindings.
A new output mode detailed-ast was added to clang-query. The existing dump output mode is now a deprecated alias for detailed-ast
Output modes can now be enabled or disabled non-exclusively.
Improvements to clang-tidy
New abseil-duration-comparison check.
Checks for comparisons which should be done in the absl::Duration domain instead of the float of integer domains.
New abseil-duration-division check.
Checks for uses of absl::Duration division that is done in a floating-point context, and recommends the use of a function that returns a floating-point value.
New abseil-duration-factory-float check.
Checks for cases where the floating-point overloads of various absl::Duration factory functions are called when the more-efficient integer versions could be used instead.
New abseil-duration-factory-scale check.
Checks for cases where arguments to absl::Duration factory functions are scaled internally and could be changed to a different factory function.
New abseil-duration-subtraction check.
Checks for cases where subtraction should be performed in the absl::Duration domain.
New abseil-faster-strsplit-delimiter check.
Finds instances of absl::StrSplit() or absl::MaxSplits() where the delimiter is a single character string literal and replaces with a character.
New abseil-no-internal-dependencies check.
Gives a warning if code using Abseil depends on internal details.
New abseil-no-namespace check.
Ensures code does not open namespace absl as that violates Abseil’s compatibility guidelines.
New abseil-redundant-strcat-calls check.
Suggests removal of unnecessary calls to absl::StrCat when the result is being passed to another absl::StrCat or absl::StrAppend.
New abseil-str-cat-append check.
Flags uses of absl::StrCat() to append to a std::string. Suggests absl::StrAppend() should be used instead.
New abseil-upgrade-duration-conversions check.
Finds calls to absl::Duration arithmetic operators and factories whose argument needs an explicit cast to continue compiling after upcoming API changes.
New bugprone-too-small-loop-variable check.
Detects those for loops that have a loop variable with a “too small” type which means this type can’t represent all values which are part of the iteration range.
New cppcoreguidelines-macro-usage check.
Finds macro usage that is considered problematic because better language constructs exist for the task.
New google-objc-function-naming check.
Checks that function names in function declarations comply with the naming conventions described in the Google Objective-C Style Guide.
New misc-non-private-member-variables-in-classes check.
Finds classes that not only contain the data (non-static member variables), but also have logic (non-static member functions), and diagnoses all member variables that have any other scope other than private.
New modernize-avoid-c-arrays check.
Finds C-style array types and recommend to use std::array<> / std::vector<>.
New modernize-concat-nested-namespaces check.
Checks for uses of nested namespaces in the form of namespace a { namespace b { ... }} and offers change to syntax introduced in C++17 standard: namespace a::b { ... }.
New modernize-deprecated-ios-base-aliases check.
Detects usage of the deprecated member types of std::ios_base and replaces those that have a non-deprecated equivalent.
New modernize-use-nodiscard check.
Adds [[nodiscard]] attributes (introduced in C++17) to member functions to highlight at compile time which return values should not be ignored.
New readability-const-return-type check.
Checks for functions with a const-qualified return type and recommends removal of the const keyword.
New readability-isolate-decl check.
Detects local variable declarations declaring more than one variable and tries to refactor the code to one statement per declaration.
New readability-magic-numbers check.
Detects usage of magic numbers, numbers that are used as literals instead of introduced via constants or symbols.
New readability-redundant-preprocessor check.
Finds potentially redundant preprocessor directives.
New readability-uppercase-literal-suffix check.
Detects when the integral literal or floating point literal has non-uppercase suffix, and suggests to make the suffix uppercase. The list of destination suffixes can be optionally provided.
New alias cert-dcl16-c to readability-uppercase-literal-suffix added.
New alias cppcoreguidelines-avoid-c-arrays to modernize-avoid-c-arrays added.
New alias cppcoreguidelines-non-private-member-variables-in-classes to misc-non-private-member-variables-in-classes added.
New alias hicpp-avoid-c-arrays to modernize-avoid-c-arrays added.
New alias hicpp-uppercase-literal-suffix to readability-uppercase-literal-suffix added.
The cppcoreguidelines-narrowing-conversions check now detects more narrowing conversions: - integer to narrower signed integer (this is compiler implementation defined), - integer - floating point narrowing conversions, - floating point - integer narrowing conversions, - constants with narrowing conversions (even in ternary operator).
The objc-property-declaration check now ignores the Acronyms and IncludeDefaultAcronyms options.
The readability-redundant-smartptr-get check does not warn about calls inside macros anymore by default.
The readability-uppercase-literal-suffix check does not warn about literal suffixes inside macros anymore by default.
