Version 0.46.0
This release significantly reworked one of the main parts of Numba, the compiler
pipeline, to make it more extensible and easier to use. The purpose of this was
to continue enhancing Numba's ability for use as a compiler toolkit. In a
similar vein, Numba now has an extension registration mechanism to allow other
Numba-using projects to automatically have their Numba JIT compilable functions
discovered. There were also a number of other related compiler toolkit
enhancement added along with some more NumPy features and a lot of bug fixes.
This release has updated the CUDA Array Interface specification to version 2,
which clarifies the `strides` attribute for C-contiguous arrays and specifies
the treatment for zero-size arrays. The implementation in Numba has been
changed and may affect downstream packages relying on the old behavior
General Enhancements:
* Add rewrite for semantic constants.
* Add np.cross support
* Make IR comparable and legalize it.
* R&D inlining, jitted and overloaded.
* Automatic JIT of called functions
* Inspection tool to check what numba supports
* Implement np.count_nonzero
* Unicode array support
* Entrypoints for numba extensions
* Literal dispatch
* Allow dtype input argument in np.sum
* New compiler.
* add support for np.append
* Refactor NRT C-API
* 0.46 scheduled deprecations
* Add env var to disable performance warnings.
* add np.array_equal support
* Implement numba.cross2d
* Add triangular indices functions
* Enable support for count() method for unicode string
Fixes:
* Fix inplace operator error for arrays
* Detect and raise unsupported on generator expressions
* Don't allow the allocation of mutable objects written into a
container to be hoisted.
* Avoid deprecated use of inspect.getargspec
* Replace GC macro with function call
* Loosen up typed container casting checks
* Fix some coding lines at the top of some files (utf8 -> utf-8)
* Replace "import \*" with explicit imports in numba/types
* Fix incorrect alg in isupper for ascii strings.
* test using jitclass in typed-list
* Add allocation hoisting info to LICM section at diagnostic L4
* Offset search box to avoid wrapping on some pages with Safari.
* Replace all "except BaseException" with "except Exception".
* Restore the "free" conda channel for NumPy 1.10 support.
* Add lowering for constant bytes.
* Add exception chaining for better error context
* Name of type should not contain user facing description for debug.
* Limit the number of return types for recursive functions
* Fixed two module teardown races in py2.
* Fix and test numpy.random.random_sample(n) for np117
* NamedTuple - Raises an error on non-iterable elements
* Add a newline in patched errors
* Fix liveness for remove dead of parfors (and other IR extensions)
* Make List.__getitem__ accept unsigned parameters
* Raise specific error at typing time for iteration on >1D array.
* Fix static_getitem with Literal type as index
* Update to inliner cost model information.
* Use specific random number seed when generating arbitrary test data
* Adjust test timeouts
* Skip unicode array tests on ppc64le that trigger an LLVM bug
* Fix packaging issue due to missing numba/cext
* Fix issue 4520 due to storage model mismatch
* Updates for llvmlite 0.30.0
CUDA Enhancements/Fixes:
* cudasim mishandling recarray
* Replace use of `np.prod` with `functools.reduce` for computing size
from shape
* Prevent taking the GIL in ForAll
* Just pass NULL for b2d_func for constant dynamic
sharedmem
* Update CUDA Array Interface & Enforce Numba compliance
* Implement math.{degrees, radians} for the CUDA target.
* Bump cuda array interface to version 2
Documentation Updates:
* Add docs for ARMv8/AArch64
* Add supported platforms to the docs.
* Add docstrings to inspect methods
* Update Python 2.7 EOL statement
* Add note about np.sum
* Minor parallel performance tips edits
* Clarify docs for typed dict with regard to arrays
* Fix example in guvectorize docstring.
* fix two typos in architecture.rst
* Document numba.extending.intrinsic and inlining.
* Fix typo in jit-compilation docs
* add dependency list to docs
* Add documentation for implementing new compiler passes.
- Removes obsolete patches.
- Updates the source code upstream to github.
Changes since 1.9.17:
1.9.18
fixes crashes when trying to run in amd64 systems
ok maya@
v0.30.0:
This release adds support for half-precision float and schedules the
deprecation of memset/memcpy accepting 5 arguments (cf. LLVM change).
* Fix use of -fPIC flag in wheels
* Remove restriction on sphinx version from Anaconda distro
* fix for block labels which contain "interesting" characters
* Deprecate the use of memset/memcpy with alias
* Add fp16 Intrinsics
* Add Half-Precision Type
* Add -fPIC flag for manylinux1 wheel building
* Fix incorrect hierarchy in the documentation for ir.Constant.
* Update docs
* Fix leak on string returning APIs.
9.0.0:
Non-comprehensive list of changes in this release
ELF Improvements
ld.lld now has typo suggestions for flags: $ ld.lld --call-shared now prints unknown argument '--call-shared', did you mean '--call_shared'.
--allow-shlib-undefined and --no-allow-shlib-undefined options are added. --no-allow-shlib-undefined is the default for executables.
-nmagic and -omagic options are fully supported.
Segment layout has changed. PT_GNU_RELRO, which was previously placed in the middle of readable/writable PT_LOAD segments, is now placed at the beginning of them. This change permits lld-produced ELF files to be read correctly by GNU strip older than 2.31, which has a bug to discard a PT_GNU_RELRO in the former layout.
-z common-page-size is supported.
Diagnostics messages have improved. A new flag --vs-diagnostics alters the format of diagnostic output to enable source hyperlinks in Microsoft Visual Studio IDE.
Linker script compatibility with GNU BFD linker has generally improved.
The clang --dependent-library form of autolinking is supported.
This feature is added to implement the Windows-style autolinking for Unix. On Unix, in order to use a library, you usually have to include a header file provided by the library and then explicitly link the library with the linker -l option. On Windows, header files usually contain pragmas that list needed libraries. Compilers copy that information to object files, so that linkers can automatically link needed libraries. --dependent-library is added for implementing that Windows semantics on Unix.
AArch64 BTI and PAC are supported.
lld now supports replacing JAL with JALX instructions in case of MIPS-microMIPS cross-mode jumps.
lld now creates LA25 thunks for MIPS R6 code.
Put MIPS-specific .reginfo, .MIPS.options, and .MIPS.abiflags sections into corresponding PT_MIPS_REGINFO, PT_MIPS_OPTIONS, and PT_MIPS_ABIFLAGS segments.
The quality of RISC-V and PowerPC ports have greatly improved. Many applications can now be linked by lld. PowerPC64 is now almost production ready.
The Linux kernel for arm32_7, arm64, ppc64le and x86_64 can now be linked by lld.
x86-64 TLSDESC is supported.
DF_STATIC_TLS flag is set for i386 and x86-64 when needed.
The experimental partitioning feature is added to allow a program to be split into multiple pieces.
The feature allows you to semi-automatically split a single program into multiple ELF files called “partitions”. Since all partitions share the same memory address space and don’t use PLT/GOT, split programs run as fast as regular programs.
With the mechanism, you can start a program only with a “main” partition and load remaining partitions on-demand. For example, you can split a web browser into a main partition and a PDF reader sub-partition and load the PDF reader partition only when a user tries to open a PDF file.
See the documentation for more information.
If “-” is given as an output filename, lld writes the final result to the standard output. Previously, it created a file “-” in the current directory.
-z ifunc-noplt option is added to reduce IFunc function call overhead in a freestanding environment such as the OS kernel.
Functions resolved by the IFunc mechanism are usually dispatched via PLT and thus slower than regular functions because of the cost of indirection. With -z ifunc-noplt, you can eliminate it by doing text relocations at load-time. You need a special loader to utilize this feature. This feature is added for the FreeBSD kernel but can be used by any operating systems.
--undefined-glob option is added. The new option is an extension to --undefined to take a glob pattern instead of a single symbol name.
COFF Improvements
Like the ELF driver, lld-link now has typo suggestions for flags.
lld-link now correctly reports duplicate symbol errors for object files that were compiled with /Gy.
lld-link now correctly reports duplicate symbol errors when several resource (.res) input files define resources with the same type, name and language. This can be demoted to a warning using /force:multipleres.
lld-link now rejects more than one resource object input files, matching link.exe. Previously, lld-link would silently ignore all but one. If you hit this: Don’t pass resource object files to the linker, instead pass res files to the linker directly. Don’t put resource files in static libraries, pass them on the command line.
Having more than two /natvis: now works correctly; it used to not work for larger binaries before.
Undefined symbols are now printed only in demangled form. Pass /demangle:no to see raw symbol names instead.
Several speed and memory usage improvements.
lld-link now supports resource object files created by GNU windres and MS cvtres, not only llvm-cvtres.
The generated thunks for delayimports now share the majority of code among thunks, significantly reducing the overhead of using delayimport.
IMAGE_REL_ARM{,64}_REL32 relocations are supported.
Range extension thunks for AArch64 are now supported, so lld can create large executables for Windows/ARM64.
The following flags have been added: /functionpadmin
WebAssembly Improvements
Imports from custom module names are supported.
Symbols that are in llvm.used are now exported by default.
Initial support for PIC and dynamic linking has landed.
wasm-ld now add __start_/__stop_ symbols for data sections.
wasm-ld now doesn’t report an error on archives without a symbol index.
The following flags have been added: --emit-relocs
MinGW Improvements
lld now correctly links crtend.o as the last object file, handling terminators for the sections such as .eh_frame properly, fixing DWARF exception handling with libgcc and gcc’s crtend.o.
lld now also handles DWARF unwind info generated by GCC, when linking with libgcc.
PDB output can be requested without manually specifying the PDB file name, with the new option -pdb= with an empty value to the option. (The old existing syntax -pdb <filename> was more cumbersome to use with an empty parameter value.)
--no-insert-timestamp option is added as an alias to /timestamp:0.
Many more GNU ld options are now supported, which e.g. allows the lld MinGW frontend to be called by GCC.
The following options are added: --exclude-all-symbols, --appcontainer, --undefined
Libc++ 9.0.0
Fixes
Minor fixes to std::chrono operators.
libc++ now correctly handles Objective-C++ ARC qualifiers in std::is_pointer.
std::span general updates and fixes.
Updates to the std::abs implementation.
std::to_chars now adds leading zeros.
Ensure std::tuple is trivially constructible.
std::aligned_union now works in C++03.
Output of nullptr to std::basic_ostream is formatted properly.
Features
Implemented P0608: sane variant converting constructor.
Added ssize function.
Added front and back methods in std::span.
std::is_unbounded_array and std::is_bounded_array added to type traits.
std::atomic now includes many new features and specialization including improved Freestanding support.
Added std::midpoint and std::lerp math functions.
Added the function std::is_constant_evaluated.
Erase-like algorithms now return size type.
Added contains method to container types.
std::swap is now a constant expression.
Updates
libc++ dropped support for GCC 4.9; we now support GCC 5.1 and above.
libc++ added explicit support for WebAssembly System Interface (WASI).
Progress towards full support of rvalues and variadics in C++03 mode. std::move and std::forward now both work in C++03 mode.
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.
1.01 Tue Oct 08 2019
Add less.pm to MANIFEST
Fix interactive test 16
1.00 Mon Oct 07 2019
Fork Term::Pager to IO::Pager::Perl, with many new features.
NOTE: This is a breaking change from 0.43 & 0.44. What was
IO::Pager::Perl in those versions is now IO::Pager::less i.e;
a shim for pager-less systems to use the pure perl pager
implementation in IO::Pager::Perl.
Update bloody version number transclusions. RT#130643
0.44 Mon Sep 30 2019
Remove debugging calls mistakenly left in code. RT#130595
Skip testing on dumb "terminal" setups. RT#130596
Add (unexposed) code for less -J search line highlighting,
and more form feed break to tp (IO::Pager::Perl).
0.43 Sun Sep 29 2019
Roll out a suboptimal but functional IO::Pager::Perl,
fixing RT#130461 and RT#130565
0.42 Thu Sep 05 2019
Disambiguate tee open. RT#127551
Stringify version. RT#127342
Add bignum dependency RT#130319
Add support for eventual Term::Pager v1.5
Update META files.
9.0.0:
Major New Features
* Experimental support for C++ for OpenCL has been added.
Non-comprehensive list of changes in this release
* The __VERSION__ macro has been updated. Previously this macro contained the string ‘4.2.1 Compatible’ to achieve compatibility with GCC 4.2.1, but that should no longer be necessary. However, to retrieve Clang’s version, please favor the one of the macro defined in clang namespaced version macros.
New Compiler Flags
* -ftime-trace and ftime-trace-granularity=N Emits flame chart style compilation time report in chrome://tracing and speedscope.app compatible format. A trace .json file is written next to the compiled object file, containing hierarchical time information about frontend activities (file parsing, template instantiation) and backend activities (modules and functions being optimized, optimization passes).
Modified Compiler Flags
* clang -dumpversion now returns the version of Clang itself.
Windows Support
* clang-cl now treats non-existent files as possible typos for flags, clang-cl /diagnostic:caret /c test.cc for example now produces clang: error: no such file or directory: '/diagnostic:caret'; did you mean '/diagnostics:caret'?
* clang now parses the __declspec(allocator) specifier and generates debug information, so that memory usage can be tracked in Visual Studio.
* The -print-search-dirs option now separates elements with semicolons, as is the norm for path lists on Windows
* Improved handling of dllexport in conjunction with explicit template instantiations for MinGW, to allow building a shared libc++ for MinGW without --export-all-symbols to override the dllexport attributes
9.0.0:
Non-comprehensive list of changes in this release:
* Two new extension points, namely EP_FullLinkTimeOptimizationEarly and EP_FullLinkTimeOptimizationLast are available for plugins to specialize the legacy pass manager full LTO pipeline.
* llvm-objcopy/llvm-strip got support for COFF object files/executables, supporting the most common copying/stripping options.
* The CMake parameter CLANG_ANALYZER_ENABLE_Z3_SOLVER has been replaced by LLVM_ENABLE_Z3_SOLVER.
* The RISCV target is no longer “experimental” (see Changes to the RISCV Target below for more details).
* The ORCv1 JIT API has been deprecated. Please see Transitioning from ORCv1 to ORCv2.
* Support for target-independent hardware loops in IR has been added, with PowerPC and Arm implementations.
Overview of changes in Glib::Object::Introspection 0.047
========================================================
* Add missing version checks to t/param-specs.t
* Fix a version check in t/arrays.t
* README: updated project Git URIs
0.22 2019-09-13T15:02:28Z
- Add the statistics method.
- The delete_user method can now take parameters, allowing for use of
the hard_delete parameter.
0.21 2019-08-24T18:56:13Z
- The release endpoint returns a 403 on an unknown tag rather than a
404 like it should. This condition is now detected and treated
like a 404.
0.20 2019-07-23T21:42:37Z
- The preexisting create_release and edit_release methods have been
renamed to create_tag_release and update_tag_release to make room
for the new releases API.
- Added methods for the new (added in GitLab 11.7) releases and
release links APIs.
Updated devel/p5-File-Find-Object to 0.3.4
Updated devel/p5-File-Next to 1.1
Updated devel/p5-File-Slurp to 9999.28
Updated devel/p5-Getopt-Long to 2.51
1.18 Tue Aug 27 17:08:38 CDT 2019
[SPEEDUP]
Updated the internals to minimize the number of times that stat()
gets called on each file or directory.
[FIXES]
t/from_file.t would fail under Windows because File::Temp would
put a backslash in front of the name of the tempfile. Thanks,
Varadinsky. (GH#32)
Had to specify a minimum version of File::Temp. Thanks, Justin Mohr.
(GH#29)
pkgsrc changes:
to avoid adding a new dependency on boost, poppler now uses cairo as
the default graphics backend. the qt5 package seems to require splash (?)
so that one gets a new boost dependency.
we could also avoid it by forcing cmake detection to fail, but in this
case it prints a warning about performance losses.
Release 0.81.0:
core:
* Splash: Always enable support for CMYK rasterization
* CairoOutputDev: Check scaled dimensions for 0. Issue #737
* BaseCryptStream: Fix potential uninitialized memory read
* SplashBitmap: Fix wrong width condition for splashModeDeviceN8
* Fix crashes in malformed files
Release 0.80.0:
core:
* Annotations: Implement support for setting a different text in the appearance stream than the real text
* Splash: Optionally use small_vector from boost
* Fix memory leaks on broken files
* Fix abort on broken files
* Small code simplifications
* Remove USE_FIXEDPOINT support. Issue #821
qt5:
* Fix MSVC build
* Add subsitute-font information
* Fix since marker of some functions
* Fix leak when aborting text extraction
* Small code simplifications
glib:
* Make print scaling getter visible
* Make Duplex/NumCopies/PrintPageRange preference available in API
* Implement Movie API
utils:
* pdftotext: Add -nodiag flag to remove diagonal text on output
build system:
* Mark external lib include dirs as SYSTEM
