1.14.2

This is a minor bug fix release with the following changes:

• Fixed a crash when 'last_value is used with record types (#1043).
• Fixed a crash when a process sensitivity list contains an external name and the process is elaborated before the object referenced by the name (#1062).
• Fixed parser hang for integers in scientific notation (from @NikLeberg).
• Fixed a platform detection issue that caused JIT code loading to fail on Apple silicon macOS (#1069).
• The STD_INPUT and STD_OUTPUT files can now be opened and closed multiple times (#1079).
• Fixed a crash when a subtype is used for an aggregate element association range choice which associates an array slice (#1080).
• Several other minor bugs were resolved (#1038, #1055).

Special thank you to @bpadalino, @tmeissner, @Blebowski, @amb5l, @m42uko, @a-panella, @cmarqu, @albydnc, @johonkanen, and @augustofg for sponsoring me!

1.14.1

This is a minor bug fix release with the following changes:

• Fixed an error when using the work library alias and the working library has the same name as design unit being analysed (#991).
• Added a check for illegal references to signal or variable external names inside pure functions (#1010).
• Added a check that the parameter of 'stable and other implicit signal attributes is a static expression (#1018).
• Fixed an elaboration crash when an external name is passed to a procedure (#1029).
• Record subtype constraints can now be accessed over VHPI (#1035).
• Several other minor bugs were resolved (#1017, #1020, #1030, #1040).

Special thank you to @bpadalino, @tmeissner, @Blebowski, @amb5l, @m42uko, @a-panella, @cmarqu, @albydnc, @johonkanen, and @augustofg for sponsoring me!

1.14.0

This is a major new release with the following changes:

• Waiting on implicit 'stable and 'quiet signals now works correctly.
• Updated to OSVVM 2024.07 and UVVM 2024.07.03b for nvc --install.
• Added a warning when an inner declaration hides an outer declaration in a way that is likely to be unintentional (#905).
• The -c sub-command for coverage is deprecated and replaced with separate --cover-report and --cover-merge commands. See the manual for details.
• The argument to the --cover-export command is now the path to a coverage database file instead of a top-level unit name.
• The --force-init command which was deprecated in the 1.7 release has been removed.
• Coverage exclude files now support fold command to merge coverage data from independent sub-hierarchies (from @Blebowski).
• FSM state coverage bin renamed from STATE to BIN_STATE.
• A signature is now allowed in the formal part of generic map associations in VHDL-2019 mode.
• use lib.pack.all no longer makes the bare package name pack potentially visible.
• Very limited initial Verilog support has been added, including the ability to instantiate vendor-supplied UDPs from VHDL.
• Fixed a regression which caused parse errors for some concurrent assertion statements (#956).
• Arrays-of-arrays with downto direction are now dumped in the correct order (#957).
• Fixed incorrect result of 'event and 'active where the prefix is an array of records.
• Fixed a crash when the 'stable attribute is used with a record type (#960).
• Conversion functions applied to individual record elements in a port map now work correctly (#963).
• External name elaboration order checks were overly strict (#964).
• Associating a signal with an out or inout parameter in a procedure that is not within a process now produces an analysis error rather than crashing at runtime (#965).
• The --dump-arrays option now takes an optional argument to specify an upper limit on the length of nested arrays to dump (#959).
• Matching case? statements are now checked for duplicate and missing choices at analysis time (#966).
• Limited maximum instantiation depth to prevent crashes due to stack overflow when elaborating a design that has unbounded recursive entity instantiation (#969).
• Physical subtype ranges and 64-bit integer types are now handled correctly by VHPI (#978).
• Added support for generic map on subprogram call in VHDL-2019 and improved support for array type generics.
• Optimised emission of FST initial signal values which also fixes a potential crash (#979).
• Added checks for duplicate attribute specification (#977).
• Several other minor bugs were resolved (#961, #962, #971, #975, #985).

Special thank you to @bpadalino, @tmeissner, @Blebowski, @amb5l, @m42uko, @a-panella, @cmarqu, @albydnc, @johonkanen, and @augustofg for sponsoring me!

1.13.3

This release fixes a number of bugs that were discovered in the 1.13 branch:

• Type checking was not performed correctly for conversion function actuals in port maps (#945).
• Fixed a crash with signal aliases inside an instantiated generic package (#946).
• Fixed random crashes when the simulation heap size is eight gigabytes or larger (#949).
• Fixed a crash when a constant array with unconstrained element type is initialised with an aggregate (#954).
• External names in concurrent assertion statements and PSL assertions are now parsed correctly (#952).
• Several other minor bugs were resolved (#942, #943, #950, #951, #953).

Special thank you to @bpadalino, @tmeissner, @Blebowski, @amb5l, @m42uko, @a-panella, @cmarqu, @albydnc, @johonkanen, @ikwzm, and @augustofg for sponsoring me!

1.13.2

This release fixes a number of bugs that were discovered in the 1.13 branch:

• Fixed an incorrect bounds check error when a constant declaration has a partially constrained type (#932).
• Elaboration of large designs on Windows no longer fail with linker errors due to excessive command line length.
• Fixed a bug which prevented VCD files from being generated on Windows systems using the UCRT runtime library (#637).
• Fixed a crash when the type of a qualified expression is an unconstrained array-of-array (#934).
• Fixed random segmentation fault when the simulation heap size is larger than four gigabytes (#935).
• Fixed an incorrect ambiguous subprogram error when using the 1993 standard and calling a subprogram via a selected name such as ieee.numeric_std."=".

Special thank you to @bpadalino, @tmeissner, @Blebowski, @amb5l, @m42uko, @a-panella, @cmarqu, @albydnc, @johonkanen, and @ikwzm for sponsoring me!

1.13.1

This release fixes a number of bugs that were discovered in 1.13.0:

• Windows installer was missing some standard library files.
• Fixed a regression where generics and ports could not be referenced in external names (#921).
• Fixed a crash when an unconstrained port with a default value is associated with open.
• Fixed a duplicate symbol linker error with some uses of the 'delayed attribute.
• Fixed a crash when an aliased subprogram is matched by a generic subprogram <> (#924)
• Fixed a crash when a generic package containing a subprogram is passed to an entity package generic (#925).
• Fixed an intermittent "failed to suspend thread" fatal error on macOS (#922).
• Fixed a crash when a shared variable is declared inside a instantiated generic package in 2019 mode (#923).
• Fixed a crash when a port with unconstrained array-of-record type is associated with a signal that has a static subtype.
• Fixed incorrect longest static prefix calculation for array slices containing 'length, 'left, 'right and similar attributes (#930).
• Fixed wrong result when slicing a generic that has an unconstrained array type (#931).

Special thank you to @bpadalino, @tmeissner, @Blebowski, @amb5l, @m42uko, @a-panella, @cmarqu, @albydnc, and @johonkanen for sponsoring me!

1.13.0

This is a major new release with a large number of improvements and bug fixes:

• Type and package generics now work correctly with components and configurations (#883).
• Fixed a crash when a matching case? alternative contains multiple choices (#890).
• Implemented VHPI support for package declarations and vhpiPackInsts iterators (#891).
• Signatures in attribute names are now parsed correctly (#892).
• Attribute specifications for all are now parsed correctly (#889).
• Predefined "=" and "/=" operators are no longer declared for file types.
• Alias of packages now work correctly (#893).
• Stricter checking of identifiers to reject those with multiple consecutive underscores or trailing underscores (from @typingArtist).
• Stricter checking for subprograms, deferred constants, and protected types referenced before their corresponding body or full declaration.
• Decimal physical literals like 8.2 ms are now correctly rounded (to 8200 us in this case) when they are converted to the base unit (#899).
• External names can now be passed as signal actuals to subprograms (#900).
• Fixed a crash when branch coverage is enabled and an if-statement contains a return (#903).
• The NVC_CONCURRENT_JOBS environment variable can be used to scale the number of worker threads NVC creates based on the number of concurrently executing simulations.
• Fixed a crash when certain attributes such as 'val are used in the target of a signal assignment statement (#908).
• Signals in packages are now dumped to the FST file (#901).
• Function interface declarations can now be pure or impure (#917).

Special thank you to @bpadalino, @tmeissner, @Blebowski, @amb5l, @m42uko, @a-panella, @cmarqu, and @albydnc for sponsoring me!

1.12.2

This is a minor release with an important compatibility fix for the latest LLVM version.

• Fixed a crash when 'transaction is used with a record type.
• Fixed a compatibility issue with LLVM 18 that could cause an illegal instruction exception when the --jit elaboration option is used (#887).

Special thank you to @bpadalino, @tmeissner, @Blebowski, @amb5l, @m42uko, @a-panella, @cmarqu, @albydnc, and @ikwzm for sponsoring me!

1.12.1

This is a minor bug fix release fixing the following issues:

• Fixed a crash when a process contains a force or release assignment inside a process (all) statement (#877).
• Fixed a crash creating a waveform dump with --dump-arrays and the design contains an array-of-array-of-records (#878).
• Fixed a spurious error when a type declaration from a package generic is used in a port list (#881).
• Fixed an assertion failure when a procedure declared within a process (all) assigns to a signal (#882).
• Added a missing check for illegal assignment to external signal name outside a process (#884).
• Fixed a crash while printing diagnostic information when the simulation delta cycle limit is reached (#885).
• Generic actuals no longer need to be globally static, which was never required by the LRM (#886).

Special thank you to @bpadalino, @tmeissner, @Blebowski, @amb5l, @m42uko, @a-panella, @cmarqu, @albydnc, and @ikwzm for sponsoring me!

1.12.0

This is a major new release with the following features and bug fixes:

• The --jit elaboration option no longer requires --no-save.
• Fixed a crash when subtype bounds depend on a package instance generic (#815).
• Fixed various issues in the implementation of guarded blocks and disconnection specifications (#829).
• The --std option now controls which VHDL standard the --install command uses to compile third-party libraries (#836).
• The prefix of an indexed or slice name now must be another name or function call, as required by the LRM (#835).
• The implementation of conversions in port maps has been reworked and fixes a number of long-standing issues (#843).
• Added support for inertial keyword in port maps (#843).
• Fixed a bug where 'last_value could give the wrong result if the signal has multiple sources.
• Updated to OSVVM 2023.09a and UVVM 2023.09.16 for nvc --install.
• The --exit-severity= option now also controls which severity level results in a non-zero exit code (#850).
• Improvements to waveform dumping for signals with record types or types with non-locally-static bounds (#851, #852).
• The parameter for attributes with dimensions such as 'length(N) may be any integer type when --relaxed is passed (#862).
• The default standard version was changed to VHDL-2008.
• The --vhpi-trace option now implies --vhpi-debug.
• The bounds of array aggregates are now calculated correctly in several corner cases (#873).
• Added supported for VHPI foreign subprograms.
• Resolved several other minor issues (#654, #854, #855, #859, #863).

For Windows users there is a new standalone MSI installer which does not depend on any MSYS2 installation.

As always thank you to @bpadalino, @tmeissner, @Blebowski, @amb5l, @m42uko, @a-panella, and @cmarqu for sponsoring me!