Skip to content

Latest commit

 

History

History
2336 lines (1731 loc) · 81.4 KB

NEWS.md

File metadata and controls

2336 lines (1731 loc) · 81.4 KB

NEWS since Mercury 22.01

Changes that may break compatibility

  • The predicates read_term/3, read_term/4, read_term_with_op_table/4, read_term_with_op_table/5 have been removed from the term_io module of the Mercury standard library.

  • The order of two arguments in the predicates write_variable/4, write_variable/5, write_variable_with_op_table/4, and write_variable_with_op_table/5in the term_io module has been swapped.

  • The io module is being reorganized. Some predicates have been marked as obsolete, while other predicates that had been marked as obsolete before the release of Mercury 22.01 have been removed.

  • The old random number generator in the random module has been removed.

  • The system of operator priorities in the ops module has been inverted. Previously, given two operators of different priorities, the one that bound more tightly was the one with the lower numerical priority. Now, operators with higher numerical priorities bind more tightly. There are other changes as well, such as priorities no longer being interchangeable with other integers, some types and function symbols being renamed, and an updated op_table type class with a somewhat different list of methods.

    The old contents of the ops module is still available as extras/old_library_modules/old_ops.

  • The mercury_term_parser module now expects the operator table to be specified using the updated op_table type class in the new ops module.

    The old contents of the mercury_term_parser module is still available as extras/old_library_modules/old_mercury_term_parser.

  • The enum/1 typeclass has been modified by having its from_int method changed from a semidet function of arity one (plus the return value) to a semidet predicate of arity two.

  • The sparse_bitset, fat_sparse_bitset and tree_bitset modules now require the items in those sets to be instances of the uenum typeclass, not the enum typeclass.

  • The digraph_key type in the digraph module is now an instance of the uenum typeclass, not the enum typeclass.

  • Six predicates in the injection module, det_insert/4, det_set/4, det_update/4, insert/4, set/4 and update/4, have had the order of their arguments changed.

  • The character sequences <<u and >>u are now recognized as single tokens. (They are the names of new versions of the left and right shift operators that expect the shift amount to be specified by an unsigned integer; previously, shift amounts were always specified by signed integers.) This means that existing code in which a << or >> operator was followed immediately, without an intervening space, by a u character, will now be parsed differently.

  • The compiler no longer accepts : as the module name separator in backquoted operators. (Mercury switched to using . as the module name separator around 2003.)

  • We have dropped support for the x86 (32-bit) version of Cygwin.

  • The --use-subdirs and --use-grade-subdirs options now cause .mh files to be placed in a Mercury/mhs subdirectory instead of the current directory. This reduces clutter in the current directory, but may require additional options to be passed to the C compiler in order for it to find those header files.

  • We have added the option --error-files-in-subdir to place .err files under the Mercury subdirectory when using mmc --make.

  • We have changed the meaning of mmc --make name.cs.

    The mmc --make target name.cs now means "build the .cs file of the named module". To build all the .c files of a program, use the target program.all_cs.

  • Some (undocumented) mmake variables in compiler-generated .dv files have been renamed:

      program.cs          -> program.all_cs
      program.os          -> program.all_os
      program.pic_os      -> program.all_pic_os
      program.javas       -> program.all_javas
      program.css         -> program.all_css
    
      program.all_mhs     -> program.mhs_to_clean
      program.all_mihs    -> program.mihs_to_clean
      program.all_int0s   -> program.int3s_to_clean
    

Changes to the Mercury standard library

Changes to the array module

  • The following predicates have been added:

    • pred semidet_least_index/2
    • pred semidet_greatest_index/2
  • The following predicates have been removed:

    • pred random_permutation/4 (replacement: random.shuffle_array/4 or random.shuffle_array/5)

    • pred sort_fix_2014/0 (no replacement needed)

  • The following function has been marked obsolete:

    • func array_to_doc/1 (replacement: pretty_printer.array_to_doc/1)

Changes to the benchmarking module

  • The following predicates have been added:

    • pred full_memory_stats_are_available/0
    • pred report_stats/3
    • pred report_stats/4
    • pred report_standard_stats/2
    • pred report_standard_stats/3
    • pred report_full_memory_stats/2
    • pred report_full_memory_stats/3
    • pred report_tabling_statistics/2
    • pred report_tabling_statistics/3
  • The following predicates have been marked obsolete:

    • pred report_stats/0
    • pred report_full_memory_stats/0

Changes to the bitmap module

  • The following predicates have been added:

    • pred num_bytes/2
    • pred from_string/2
    • pred read_bitmap/6
    • pred read_bitmap/7
    • pred read_bitmap_range/8
    • pred read_bitmap_range/9
    • pred write_bitmap/3
    • pred write_bitmap/4
    • pred write_bitmap_range/5
    • pred write_bitmap_range/6

Changes to the builtin module

  • The following obsolete predicate and function have been removed:

    • func promise_only_solution/1
    • pred promise_only_solution_io/4

Changes to the char module

  • The following type has had its typeclass memberships changed:

    • The type character is now an instance of the new uenum typeclass.
  • The following predicate and functions have been added:

    • func to_uint/1
    • pred from_uint/2
    • func det_from_uint/1
  • The following function has been marked obsolete:

    • func char_to_doc/1 (replacement: pretty_printer.char_to_doc/1)

Changes to the construct module

  • The following predicates have been added:

    • pred num_functors/2
    • pred get_functor_ordinal/3
    • pred get_functor_lex/3
    • pred construct/4

Changes to the cord module

  • The following predicates have been added:

    • pred is_non_empty/1
    • pred foldr2/6
    • pred foldr3/8
    • pred head/2
    • pred is_singleton/2

Changes to the counter module

  • The following type has been added:

    • type ucounter/0
  • The following predicate and function have been added:

    • func uinit/1
    • pred uinit/2
    • pred uallocate/1

Changes to the diet module

  • The following obsolete predicate has been removed:

    • pred empty/1

Changes to the digraph module

  • The following type has had its typeclass memberships changed:

    • The digraph_key(T) type is now an instance of the uenum typeclass, and is no longer an instance of the enum typeclass.
  • The following functions have been added:

    • func symmetric_closure/1 (synonym for sc/1)
    • func transitive_closure/1 (synonym for tc/1)
    • func reflexive_transitive_closure/1 (synonym for rtc/1)
  • We have improved the implementations of transitive closure and reflexive transitive closure.

New edit_distance module

  • This new module defines operations that, when given a query sequence and a set of candidate sequences, finds out which candidate sequence is closest to the query sequence, in terms of the configurable distances added by operations such as inserting or replacing elements in sequences.

Changes to the enum module

  • The following typeclass has been modified:

    • The enum/1 typeclass has had its from_int method changed from a semidet function of arity one (plus the return value) to a semidet predicate of arity two.
  • The following typeclass has been added:

    • typeclass uenum/1
  • The following functions have been added:

    • func from_int/1 (replaces class method; born obsolete)
    • func det_from_uint/1

Changes to the fat_sparse_bitset module

  • The following type has had its typeclass memberships changed:

    • The fat_sparse_bitset(T) type now requires T to be an instance of the uenum typeclass, replacing the earlier requirement that it be a member of the enum typeclass.
  • The documentation of the following predicates and functions have been clarified:

    • func sorted_list_to_set/1
    • pred sorted_list_to_set/2
  • The following obsolete predicate has been removed:

    • pred empty/1

New fatter_sparse_bitset module

  • This new module defines a more compact version of the data structure defined by the fat_sparse_bitset module, and implements the same set of operations on this data structure.

Changes to the float module

  • The following function has been marked obsolete:

    • func float_to_doc/1 (replacement: pretty_printer.float_to_doc/1)

Changes to the hash_table module

  • The following predicate has been added:

    • pred lookup/3
  • The following obsolete predicates have been removed:

    • pred char_hash/2
    • pred float_hash/2
    • pred generic_hash/2
    • pred int_hash/2
    • pred string_hash/2
    • pred uint_hash/2

Changes to the injection module

  • The following predicates have had the order of their arguments changed:

    • pred det_insert/4
    • pred det_set/4
    • pred det_update/4
    • pred insert/4
    • pred set/4
    • pred update/4

Changes to the int module

  • The following type has had its typeclass memberships changed:

    • The type int is now an instance of the new uenum typeclass.
  • The following predicates and functions have been added:

    • func <<u/2
    • func >>u/2
    • func unchecked_left_ushift/2
    • func unchecked_right_ushift/2
    • func ubits_per_int/0
    • pred ubits_per_int/1
  • The following function has been marked obsolete:

    • func int_to_doc/1 (replacement: pretty_printer.int_to_doc/1)

Changes to the int8 module

  • The following functions have been added:

    • func <<u/2
    • func >>u/2
    • func unchecked_left_ushift/2
    • func unchecked_right_ushift/2
  • The following function has been marked obsolete:

    • func int8_to_doc/1 (replacement: pretty_printer.int8_to_doc/1)

Changes to the int16 module

  • The following functions have been added:

    • func <<u/2
    • func >>u/2
    • func unchecked_left_ushift/2
    • func unchecked_right_ushift/2
  • The following function has been marked obsolete:

    • func int16_to_doc/1 (replacement: pretty_printer.int16_to_doc/1)

Changes to the int32 module

  • The following functions have been added:

    • func <<u/2
    • func >>u/2
    • func unchecked_left_ushift/2
    • func unchecked_right_ushift/2
  • The following function has been marked obsolete:

    • func int32_to_doc/1 (replacement: pretty_printer.int32_to_doc/1)

Changes to the int64 module

  • The following functions have been added:

    • func <<u/2
    • func >>u/2
    • func unchecked_left_ushift/2
    • func unchecked_right_ushift/2
  • The following function has been marked obsolete:

    • func int64_to_doc/1 (replacement: pretty_printer.int64_to_doc/1)

Changes to the io module

  • The following predicates and functions have been added:

    • pred make_io_error_from_system_error/5
    • pred make_io_error_from_windows_error/5
    • pred get_errno_error/2
    • pred get_exception_object_error/2
    • pred get_system_error/2
    • pred get_system_error_name/2
    • pred get_windows_error/2
    • func init_posn/0
    • pred read_file_as_string_wf/3
    • pred read_file_as_string_wf/4
    • pred read_file_as_string_and_num_code_units_wf/3
    • pred read_file_as_string_and_num_code_units_wf/4
    • pred read_named_file_as_string_wf/4
    • pred read_named_file_as_lines_wf/4
    • pred system_error_is_success/1
    • pred write_binary_string_utf8/3
    • pred write_binary_string_utf8/4
    • pred write_prefixed_lines/4
    • pred write_prefixed_lines/5
  • The following obsolete predicates have been removed:

    • pred see/3 (replacement: prolog.see/3)
    • pred see_binary/3 (replacement: prolog.see_binary/3)
    • pred seen/2 (replacement: prolog.seen/2)
    • pred seen_binary/2 (replacement: prolog.seen_binary/2)
    • pred tell/4 (replacement: prolog.tell/4)
    • pred tell_binary/4 (replacement: prolog.tell_binary/4)
    • pred told/2 (replacement: prolog.told/2)
    • pred told_binary/2 (replacement: prolog.told_binary/2)
  • The following predicates have been marked obsolete:

    • pred read_bitmap/6 (replacement: bitmap.read_bitmap/6)

    • pred read_bitmap/7 (replacement: bitmap.read_bitmap/7)

    • pred read_bitmap/8 (replacement: bitmap.read_bitmap_range/8)

    • pred read_bitmap/9 (replacement: bitmap.read_bitmap_range/9)

    • pred write_bitmap/3 (replacement: bitmap.write_bitmap/3)

    • pred write_bitmap/4 (replacement: bitmap.write_bitmap/4)

    • pred write_bitmap/5 (replacement: bitmap.write_bitmap_range/5)

    • pred write_bitmap/6 (replacement: bitmap.write_bitmap_range/6)

    • pred report_stats/3 (replacement: benchmarking.report_stats/3)

    • pred report_stats/4 (replacement: benchmarking.report_stats/4)

    • pred report_standard_stats/2 (replacement: benchmarking.report_standard_stats/2)

    • pred report_standard_stats/3 (replacement: benchmarking.report_standard_stats/3)

    • pred report_full_memory_stats/2 (replacement: benchmarking.report_full_memory_stats/2)

    • pred report_full_memory_stats/3 (replacement: benchmarking.report_full_memory_stats/3)

    • pred report_tabling_statistics/2 (replacement: benchmarking.report_tabling_statistics/2)

    • pred report_tabling_statistics/3 (replacement: benchmarking.report_tabling_statistics/3)

    • pred remove_file/4 (replacement: io.file.remove_file/4)

    • pred remove_file_recursively/4 (replacement: io.file.remove_file_recursively/4)

    • pred rename_file/5 (replacement: io.file.rename_file/5)

    • pred have_symlinks/0 (replacement: io.file.have_symlinks/0)

    • pred make_symlink/5 (replacement: io.file.make_symlink/5)

    • pred read_symlink/4 (replacement: io.file.read_symlink/4)

    • pred check_file_accessibility/5 (replacement: io.file.check_file_accessibility/5)

    • pred file_type/5 (replacement: io.file.file_type/5)

    • pred file_modification_time/4 (replacement: io.file.file_modification_time/4)

    • pred make_temp_file/3 (replacement: io.file.make_temp_file/3)

    • pred make_temp_file/6 (replacement: io.file.make_temp_file/6)

    • pred make_temp_directory/3 (replacement: io.file.make_temp_directory/3)

    • pred make_temp_directory/6 (replacement: io.file.make_temp_directory/6)

    • pred have_make_temp_directory/0 (replacement: io.file.have_make_temp_directory/0)

    • pred get_temp_directory/3 (replacement: io.file.get_temp_directory/3)

    • pred get_environment_var/4 (replacement: io.environment.get_environment_var/4)

    • pred set_environment_var/5 (replacement: io.environment.set_environment_var/5)

    • pred set_environment_var/4 (replacement: io.environment.set_environment_var/4)

    • pred have_set_environment_var/0 (replacement: io.environment.have_set_environment_var/0)

    • pred get_environment_var_map/3 (replacement: io.environment.get_environment_var_map/3)

    • pred call_system/4 (replacement: io.call_system.call_system/4)

    • pred call_system_return_signal/4 (replacement: io.call_system.call_system_return_signal/4)

    • pred get_globals/3 (replacement: a user-defined mutable)

    • pred set_globals/3 (replacement: a user-defined mutable)

    • pred update_globals/3 (replacement: a user-defined mutable)

    • pred input_stream_foldl/5 (replacement: stream.input_stream_fold/6)

    • pred input_stream_fold/6 (replacement: stream.input_stream_fold/6)

    • pred input_stream_foldl_io/4 (replacement: stream.input_stream_fold_state/5)

    • pred input_stream_foldl_io/5 (replacement: stream.input_stream_fold_state/5)

    • pred input_stream_foldl2_io/5 (replacement: stream.input_stream_foldl_state/6)

    • pred input_stream_foldl2_io/6 (replacement: stream.input_stream_fold_state/6)

    • pred input_stream_foldl2_io_maybe_stop/5 (replacement: stream.input_stream_fold2_state_maybe_stop/6)

    • pred binary_input_stream_foldl/5) (replacement: stream.input_stream_fold/6)

    • pred binary_input_stream_foldl/6) (replacement: stream.input_stream_fold/6)

    • pred binary_input_stream_foldl_io/4 (replacement: stream.input_stream_fold_state/5)

    • pred binary_input_stream_foldl2_io/5 (replacement: stream.input_stream_fold2_state/6)

    • pred binary_input_stream_foldl2_io/6 (replacement: stream.input_stream_fold2_state/6)

    • pred binary_input_stream_foldl2_io_maybe_stop/5 (replacement: stream.input_stream_fold2_state_maybe_stop/6)

    • pred binary_input_stream_foldl2_io_maybe_stop/6 (replacement: stream.input_stream_fold2_state_maybe_stop/6)

New io.call_system module

  • This new module has these predicates and functions:

    • pred call_system/4
    • pred call_system_return_signal/4
    • func decode_system_command_exit_code/1

New io.environment module

  • This new module has these predicates:

    • pred get_environment_var/4
    • pred set_environment_var/5
    • pred set_environment_var/4
    • pred have_set_environment_var/0
    • pred get_environment_var_map/3

New io.file module

  • This new module has these predicates:

    • pred remove_file/4
    • pred remove_file_recursively/4
    • pred rename_file/5
    • pred have_symlinks/0
    • pred make_symlink/5
    • pred read_symlink/4
    • pred check_file_accessibility/5
    • pred file_type/5
    • pred file_modification_time/4
    • pred make_temp_file/3
    • pred make_temp_file/6
    • pred make_temp_directory/3
    • pred make_temp_directory/6
    • pred have_make_temp_directory/0
    • pred get_temp_directory/3

Changes to the library module

  • The following functions have been added:

    • func architecture/0
    • func mercury_version/0
    • func package_version/0

Changes to the list module

  • The following predicates and functions have been added:

    • pred chunk_foldl/5
    • pred chunk_foldl2/7
    • pred chunk_foldl3/9
    • pred chunk_foldl4/11
    • pred gap_foldl/5
    • pred head/2
    • pred tail/2
    • pred det_head/2
    • pred det_tail/2
    • func inst_preserving_condense/1
    • pred intersperse/3
    • pred intersperse_list/3
    • pred intersperse_list_last/4
    • pred is_non_empty/1
    • pred is_singleton/2
    • pred last_gap_foldl/6
    • func map_corresponding4/5
    • pred map_corresponding4/6
    • func merge_lists/1
    • func merge_lists/2
    • pred merge_lists/2
    • pred merge_lists/3
    • func merge_lists_and_remove_dups/1
    • func merge_lists_and_remove_dups/2
    • pred merge_lists_and_remove_dups/2
    • pred merge_lists_and_remove_dups/3
    • func take_while_not/2
    • pred take_while_not/3
    • pred take_while_not/4
  • The following function has been marked obsolete:

    • func list_to_doc/1 (replacement: pretty_printer.list_to_doc/1)

Changes to the one_or_more module

  • The following predicate has been added:

    • pred is_non_empty/1
  • The following function has been marked obsolete:

    • func one_or_more_to_doc/1 (replacement: pretty_printer.one_or_more_to_doc/1)

Changes to the map module

  • The following predicates have been added:

    • pred max_key/2
    • pred min_key/2
    • pred compose_maps/3
    • pred sorted_keys_match/2

Changes to the ops module

  • The representation of operator priorities has been changed. Priorities used to be represented as signed integers between 0 and 1200, with higher numbers representing operators that bind less tightly. Priorities are now represented using unsigned integers between 0u and 1500u, with higher numbers representing operators that bind more tightly. To guard against priorities being unintentionally confused with other kinds of numbers, they are now wrapped in the prio function symbol.

  • The representation of the set of operators that a string may represent has been changed. Operator tables used to map each operator string (such as "-") to a list of operator descriptions, each of which could be infix, binary prefix, prefix or postfix. This allowed a table entry to declare an operator string to be e.g. an infix operator using two separate entries, with different priorities. Operator tables now map each operator string to a structure which has

    • one slot for a description of an infix operator,
    • one slot for a description of a binary prefix operator,
    • one slot for a description of a unary prefix operator, and
    • one slot for a description of a unary postfix operator.

    Besides making some nonsensical entries that could exist in old op tables unrepresentable, the new op table also allows users of operator information to test more simply and quickly whether an operator string represents e.g. a prefix operator.

  • The assoc type has been renamed to arg_prio_gt_or_ge, which reflects its semantics. Its function symbols x and y have been renamed to arg_gt and arg_ge respectively.

  • The list of methods in the op_table type class has been changed.

  • The method max_priority/1 has been replaced by the method loosest_op_priority/1.

  • The method lookup_op/2 has been renamed to is_op/2.

  • The new method universal_priority/1 has been added. It is intended to replace uses of max_priority + 1 to denote a context that accept terms using any operator, regardless of its priority.

  • The new method tightest_op_priority/1 has been added. It is intended to replace uses of the integer constant 0 as a priority.

  • The new method comma_priority/1 has been added. It is intended to denote the priority of the comma character, which separates term arguments, when used as an operator.

  • The argument types of the lookup_op_infos/2 method have been changed.

  • The following predicates and functions have been added:

    • func decrement_priority/1
    • func increment_priority/1
    • func mercury_op_table_arg_priority/0
    • func mercury_op_table_comma_priority/0
    • func mercury_op_table_loosest_op_priority/0
    • func mercury_op_table_tightest_op_priority/0
    • func mercury_op_table_universal_priority/0
    • pred mercury_op_table_binary_prefix_op/5
    • pred mercury_op_table_infix_op/5
    • pred mercury_op_table_lookup_operator_term/3
    • pred mercury_op_table_postfix_op/4
    • pred mercury_op_table_prefix_op/4
    • pred mercury_op_table_search_binary_prefix_op/4
    • pred mercury_op_table_search_infix_op/4
    • pred mercury_op_table_search_op/1
    • pred mercury_op_table_search_op_infos/3
    • pred mercury_op_table_search_postfix_op/3
    • pred mercury_op_table_search_prefix_op/3
    • func min_priority_for_arg/2
    • pred priority_ge/2
    • pred priority_gt/2
    • pred priority_le/2
    • pred priority_lt/2
  • The following predicates have been renamed:

    • pred mercury_op_table_prefix_op/5 to op_infos_prefix_op/5
    • pred mercury_op_table_binary_prefix_op/5 to op_infos_binary_prefix_op/5
    • pred mercury_op_table_infix_op/5 to op_infos_infix_op/5
    • pred mercury_op_table_postfix_op/5 to op_infos_postfix_op/5

Changes to the pretty_printer module

  • The following predicates and functions have been added:

    • func array_to_doc/1
    • func char_to_doc/1
    • func float_to_doc/1
    • func get_formatter_map_entry_types/1
    • func int_to_doc/1
    • func int8_to_doc/1
    • func int16_to_doc/1
    • func int32_to_doc/1
    • func int64_to_doc/1
    • func list_to_doc/1
    • func one_or_more_to_doc/1
    • func string_to_doc/1
    • func tree234_to_doc/1
    • func uint_to_doc/1
    • func uint8_to_doc/1
    • func uint16_to_doc/1
    • func uint32_to_doc/1
    • func uint64_to_doc/1
    • func version_array_to_doc/1
    • pred write_doc_formatted/3
    • pred write_doc_formatted/4

Changes to the prolog module

  • The following predicate has been deleted:

    • pred is/2

New ra_list module

  • This new module contains operations on random access lists.

Changes to the random module

  • The old random number generator has been removed. It has been replaced by the new type class random number generation framework defined in this module. As a result, the following predicates have been removed:

    • pred init/2
    • pred random/3
    • pred random/5
    • pred randmax/3
    • pred randcount/3
    • pred permutation/4 (replacement: random.shuffle_list/4 or random.shuffle_list/5)

Changes to the ranges module

  • The following predicate has been added:

    • pred search_range/4

Changes to the rbtree module

  • The following predicates and functions have been added:

    • func foldr/3
    • pred foldr/4
    • pred foldr2/6
    • pred foldr_values/4

Changes to the set module

  • The following predicate has been added:

    • pred map2_fold/6
  • The following obsolete predicates and function have been removed:

    • pred empty/1
    • pred non_empty/1
    • func set/1

Changes to the set_bbbtree module

  • The following obsolete predicates have been removed:

    • pred empty/1
    • pred non_empty/1

Changes to the set_ctree234 module

  • The following obsolete predicates have been removed:

    • pred empty/1
    • pred non_empty/1

Changes to the set_ordlist module

  • The following obsolete predicates have been removed:

    • pred empty/1
    • pred non_empty/1

Changes to the set_tree234 module

  • The following obsolete predicates have been removed:

    • pred empty/1
    • pred non_empty/1

Changes to the set_unordlist module

  • The following obsolete predicates have been removed:

    • pred empty/1
    • pred non_empty/1

Changes to the sparse_bitset module

  • The following type has had its typeclass memberships changed:

    • The sparse_bitset(T) type now requires T to be an instance of the uenum typeclass, replacing the earlier requirement that it be a member of the enum typeclass.
  • The following obsolete predicate has been removed:

    • pred empty/1
  • The documentation of the following predicates and functions have been clarified:

    • func sorted_list_to_set/1
    • pred sorted_list_to_set/2

Changes to the std_util module

  • We have changed the behaviour of pow/3 so that it throws an exception if it is called with a negative integer as its second argument.

Changes to the string module

  • The following predicates and functions have been added:

    • func add_prefix/2
    • func between_code_points/3
    • pred between_code_points/4
    • pred check_well_formedness/2
    • pred code_point_offset/3
    • pred code_point_offset/4
    • pred contains_match/2
    • func count_code_points/1
    • pred count_code_points/2
    • pred find_first_char/3
    • pred find_first_char_start/4
    • pred find_last_char/3
    • func internal_string_encoding/0
    • func left_by_code_point/2
    • pred left_by_code_point/3
    • func right_by_code_point/2
    • pred right_by_code_point/3
    • pred split_by_code_point/3
    • pred split_by_code_point/4
    • pred unsafe_find_first_char_start/4
  • The following obsolete modes have been removed from the following predicates:

    • pred to_char_list(uo, in)
    • pred to_rev_char_list(in, out)
    • pred append(out, out, in)
    • pred prefix(in, out)
    • pred suffix(in, out)
  • The following predicates and functions have been marked obsolete:

    • func string_to_doc/1 (replacement: pretty_printer.string_to_doc/1)
    • func between_codepoints/3 (replacement: between_code_points/3)
    • pred between_codepoints/4 (replacement: between_code_points/4)
    • pred codepoint_offset/3 (replacement: code_point_offset/3)
    • pred codepoint_offset/4 (replacement: code_point_offset/4)
    • func count_codepoints/1 (replacement: count_code_points/1)
    • pred count_codepoints/2 (replacement: count_code_points/2)
    • func left_by_codepoint/2 (replacement: left_by_code_point/2)
    • pred left_by_codepoint/3 (replacement: left_by_code_point/3)
    • func right_by_codepoint/2 (replacement: right_by_code_point/2)
    • pred right_by_codepoint/3 (replacement: right_by_code_point/3)
    • pred split_by_codepoint/3 (replacement: split_by_code_point/3)
    • pred split_by_codepoint/4 (replacement: split_by_code_point'/4)
  • The semantics of the format_table_max function has been changed. Previously, the function's documentation was silent on how the returned string would be constructed if some columns contained strings that exceeded that column's maximum width. What it actually did was to split the line at the end of the too-long columns, continuing the row in the next line at the expected initial column. Since splitting the line is probably not what users of this function expect, we have now changed this function to never break a row into more than one line. We have also documented this behavior.

Changes to the string.builder module

  • The following predicates and functions have been added:

    • pred append_char/3
    • pred append_string/3
    • pred append_strings/3
    • pred append_strings_sep/4
    • pred format/4
    • func total_num_code_points/1
    • pred total_num_code_points_is_at_most/2

Changes to the term module

  • The following type has had its typeclass memberships changed:

    • The type var/1 is now an instance of the new uenum typeclass.
  • The following predicates have been marked obsolete:

    • pred generic_term/1 (replacement: with_type annotations on expressions)

    • pred decimal_term_to_int/2 (replacement: term_int.decimal_term_to_int/2)

    • pred term_to_int/2 (replacement: term_int.term_to_int/2)

    • pred term_to_int8/2 (replacement: term_int.term_to_int8/2)

    • pred term_to_int16/2 (replacement: term_int.term_to_int16/2)

    • pred term_to_int32/2 (replacement: term_int.term_to_in32t/2)

    • pred term_to_int64/2 (replacement: term_int.term_to_int64/2)

    • pred term_to_uint/2 (replacement: term_int.term_to_uint/2)

    • pred term_to_uint8/2 (replacement: term_int.term_to_uint8/2)

    • pred term_to_uint16/2 (replacement: term_int.term_to_uint16/2)

    • pred term_to_uint32/2 (replacement: term_int.term_to_uin32t/2)

    • pred term_to_uint64/2 (replacement: term_int.term_to_uint64/2)

    • func int_to_decimal_term/2 (replacement: term_int.int_to_decimal_term/2)

    • func int8_to_decimal_term/2 (replacement: term_int.int8_to_decimal_term/2)

    • func int16_to_decimal_term/2 (replacement: term_int.int16_to_decimal_term/2)

    • func int32_to_decimal_term/2 (replacement: term_int.int32_to_decimal_term/2)

    • func int64_to_decimal_term/2 (replacement: term_int.int64_to_decimal_term/2)

    • func uint_to_decimal_term/2 (replacement: term_int.uint_to_decimal_term/2)

    • func uint8_to_decimal_term/2 (replacement: term_int.uint8_to_decimal_term/2)

    • func uint16_to_decimal_term/2 (replacement: term_int.uint16_to_decimal_term/2)

    • func uint32_to_decimal_term/2 (replacement: term_int.uint32_to_decimal_term/2)

    • func uint64_to_decimal_term/2 (replacement: term_int.uint64_to_decimal_term/2)

    • pred occurs/3 (replacement: term_subst.var_occurs_in_subst_term/2)

    • pred occurs_list/3 (replacement: term_subst.var_occurs_in_subst_terms/2)

    • pred is_ground/1 (replacement: term_subst.term_is_ground/1)

    • pred is_ground_in_bindings/2 (replacement: term_subst.term_is_ground_in_bindings/2)

    • pred rename_var_in_term/4 (replacement: term_subst.rename_var_in_term/4)

    • pred rename_var_in_terms/4 (replacement: term_subst.rename_var_in_terms/4)

    • pred apply_renaming_in_var/3 (replacement: term_subst.apply_renaming_in_var/3)

    • pred apply_renaming_in_vars/3 (replacement: term_subst.apply_renaming_in_vars/3)

    • pred apply_renaming_in_term/3 (replacement: term_subst.apply_renaming_in_term/3)

    • pred apply_renaming_in_terms/3 (replacement: term_subst.apply_renaming_in_terms/3)

    • pred substitute_var_in_term/4 (replacement: term_subst.substitute_var_in_term/4)

    • pred substitute_var_in_terms/4 (replacement: term_subst.substitute_var_in_terms/4)

    • pred substitute_corresponding_in_term/3 (replacement: term_subst.substitute_corresponding_in_term/3)

    • pred substitute_corresponding_in_terms/3 (replacement: term_subst.substitute_corresponding_in_terms/3)

    • pred apply_substitution_in_term/3 (replacement: term_subst.apply_substitution_in_term/3)

    • pred apply_substitution_in_terms/3 (replacement: term_subst.apply_substitution_in_terms/3)

    • pred apply_rec_substitution_in_term/3 (replacement: term_subst.apply_rec_substitution_in_term/3)

    • pred apply_rec_substitution_in_terms/3 (replacement: term_subst.apply_rec_substitution_in_terms/3)

    • func term_list_to_var_list/2 (replacement: term_subst.term_list_to_var_list/1)

    • pred term_list_to_var_list/2 (replacement: term_subst.term_list_to_var_list/2)

    • func var_list_to_term_list/2 (replacement: term_subst.var_list_to_term_list/1)

    • pred var_list_to_term_list/2 (replacement: term_subst.var_list_to_term_list/2)

    • pred unify_term/4 (replacement: term_unify.unify_terms/4)

    • pred unify_term_list/4 (replacement: term_unify.unify_term_lists/4)

    • pred unify_term_dont_bind/5 (replacement: term_unify.unify_terms_dont_bind/5)

    • pred unify_term_list_dont_bind/5 (replacement: term_unify.unify_term_lists_dont_bind/5)

    • pred list_subsumes/3 (replacement: term_unify.first_term_list_subsumes_second/3)

    • func vars/1 (replacement: term_vars.vars_in_term/1)

    • pred vars/2 (replacement: term_vars.vars_in_term/2)

    • func vars_2/2 (replacement: term_vars.vars_in_term_acc/3)

    • pred vars_2/3 (replacement: term_vars.vars_in_term_acc/3)

    • func vars_list/1 (replacement: term_vars.vars_in_terms/1)

    • pred vars_list/2 (replacement: term_vars.vars_in_terms/2)

    • pred contains_var/2 (replacement: term_vars.term_contains_var/2)

    • pred contains_var_list/2 (replacement: term_vars.terms_contain_var/2)

    • func context_init/2 (replacement: term_context.context_init/2)

    • pred context_init/3 (replacement: term_context.context_init/2)

    • func context_init/0 (replacement: term_context.dummy_context/0)

    • pred context_init/1 (replacement: term_context.dummy_context/0)

    • func dummy_context_init/0 (replacement: term_context.dummy_context/0)

    • pred is_dummy_context/1 (replacement: term_context.is_dummy_context/1)

    • func context_file/1 (replacement: term_context.context_file/1)

    • pred context_file/2 (replacement: term_context.context_file/1)

    • func context_line/1 (replacement: term_context.context_line/1)

    • pred context_line/2 (replacement: term_context.context_line/1)

New term_context module

  • This new module defines the term_context type, and contains predicates and functions that operate on that type.

New term_int module

  • This new module contains predicates that recognize terms that contain integers, and functions that construct such terms.

Changes to the term_io module

  • The following predicates and functions have been added:

    • func constant_to_string/2
    • func escaped_char_to_string/2
    • func quoted_char_to_string/2
    • pred format_constant/4
    • pred format_escaped_char/4
    • pred format_escaped_string/4
    • pred format_quoted_atom/4
    • pred format_quoted_char/4
    • pred format_quoted_string/4
    • pred format_term/5
    • pred format_term_nl/5
    • pred format_term_with_op_table/6
    • pred format_term_nl_with_op_table/6
    • pred format_variable/5
    • pred format_variable_with_op_table/6
    • func term_to_string/2
    • func term_nl_to_string/2
    • func term_with_op_table_to_string/3
    • func term_nl_with_op_table_to_string/3
    • func variable_to_string/2
    • func variable_with_op_table_to_string/3
    • pred write_quoted_atom/3
    • pred write_quoted_atom/4
    • pred write_quoted_char/3
    • pred write_quoted_char/4
    • pred write_quoted_string/3
    • pred write_quoted_string/4
  • The order of the arguments of the following predicates has been changed:

    • pred write_variable/4
    • pred write_variable/5
    • pred write_variable_with_op_table/4
    • pred write_variable_with_op_table/5
  • The following predicates have been removed:

    • pred read_term/3 (replacement: mercury_term_parser.read_term/3)
    • pred read_term/4 (replacement: mercury_term_parser.read_term/4)
    • pred read_term_with_op_table/4 (replacement: mercury_term_parser.read_term_with_op_table/4)
    • pred read_term_with_op_table/5 (replacement: mercury_term_parser.read_term_with_op_table/5)

New term_subst module

  • This new module contains predicates that perform substitutions of various kinds on terms.

New term_unify module

  • This new module contains predicates that perform unifications.

New term_vars module

  • This new module contains predicates and functions that find variables in terms.

Changes to the thread module

  • The following predicates have been added:

    • pred spawn_native_joinable/5
    • pred join_thread/4

Changes to the tree_bitset module

  • The following predicates have been added:

    • pred remove_leq/3
    • pred remove_gt/3
  • The documentation of the following predicates and functions have been clarified:

    • func sorted_list_to_set/1
    • pred sorted_list_to_set/2

Changes to the tree234 module

  • The following predicates have been added:

    • pred max_key/2
    • pred min_key/2
    • pred sorted_keys_match/2
  • The following function has been marked obsolete:

    • func tree234_to_doc/1 (replacement: pretty_printer.tree234_to_doc/1)

Changes to the tree_bitset module

  • The following obsolete predicate has been removed:

    • pred empty/1

Changes to the type_desc module

  • The following predicates have been added:

    • pred make_type/3
    • pred pseudo_type_ctor/2
    • pred pseudo_type_args/2

Changes to the uint module

  • The following type has had its typeclass memberships changed:

    • The type uint is now an instance of the new uenum typeclass.
  • The following functions have been added:

    • func <<u/2
    • func >>u/2
    • func unchecked_left_ushift/2
    • func unchecked_right_ushift/2
    • func ubits_per_uint/0
  • The following function has been marked obsolete:

    • func uint_to_doc/1 (replacement: pretty_printer.uint_to_doc/1)

Changes to the uint8 module

  • The following functions have been added:

    • func <<u/2
    • func >>u/2
    • func unchecked_left_ushift/2
    • func unchecked_right_ushift/2
  • The following function has been marked obsolete:

    • func uint8_to_doc/1 (replacement: pretty_printer.uint8_to_doc/1)

Changes to the uint16 module

  • The following functions have been added:

    • func <<u/2
    • func >>u/2
    • func unchecked_left_ushift/2
    • func unchecked_right_ushift/2
  • The following function has been marked obsolete:

    • func uint16_to_doc/1 (replacement: pretty_printer.uint16_to_doc/1)

Changes to the uint32 module

  • The following functions have been added:

    • func <<u/2
    • func >>u/2
    • func unchecked_left_ushift/2
    • func unchecked_right_ushift/2
  • The following function has been marked obsolete:

    • func uint32_to_doc/1 (replacement: pretty_printer.uint32_to_doc/1)

Changes to the uint64 module

  • The following functions have been added:

    • func <<u/2
    • func >>u/2
    • func unchecked_left_ushift/2
    • func unchecked_right_ushift/2
  • The following function has been marked obsolete:

    • func uint64_to_doc/1 (replacement: pretty_printer.uint64_to_doc/1)

Changes to the version_array module

  • The following predicate has been added:

    • pred lookup/3
  • The following function has been marked obsolete:

    • func version_array_to_doc/1 (replacement: pretty_printer.version_array_to_doc/1)

Changes to the version_hash_table module

  • The following predicate has been added:

    • pred lookup/3
  • The following obsolete predicates have been removed:

    • pred char_hash/2
    • pred float_hash/2
    • pred generic_hash/2
    • pred int_hash/2
    • pred string_hash/2
    • pred uint_hash/2

Changes to the Mercury language

  • The new pragma format_call can now be used to ask the compiler to perform the same checks on calls to a named predicate or function as the compiler performs on calls to string.format, io.format and stream.string_writer.format. These checks test whether the format string matches the provided list of values.

  • You can now disable another two kinds of warning with a disable_warning scope.

    First, code such as

    disable_warning [unknown_format_calls] (
        ...
    )
    

    will disable warnings about any calls in the scope to string.format, io.format and/or stream.string_writer.format for which the compiler cannot tell whether there are any mismatches between the format string and the supplied values, even if the module is compiled with --warn-unknown-format-calls.

    Second, code such as

    disable_warning [warn_repeated_singleton_vars] (
        ...
    )
    

    will disable warnings in the scope about variables whose names start with an underscore, indicating that they should be singleton variables, but which nevertheless occur more than once.

  • Floating point number literals may now contain an underscore between the digits forming the fraction part and the e or E that introduces the exponent part.

  • You can now use the escape sequence '\e' to refer to the ESC (escape) character, just as you can use '\n' to refer to the newline character.

Changes to the Mercury compiler

  • We now enforce a (previously unstated) requirement that existentially quantified type variables and existential class constraints in a subtype definition are listed in the same order as in the supertype definition.

  • We have fixed a bug where the compiler generated incorrect code in C# and Java grades when a subtype type definition did not repeat the field names of its base type.

  • We have fixed a bug where --warn-unused-imports did not warn about unused modules that are also imported by an ancestor of the current module.

  • The compiler now uses colors in error messages to draw programmers' attention to the most relevant parts of diagnostics. Users can change the color scheme using the new option --color-scheme or using the new environment variable MERCURY_COLOR_SCHEME. They can also disable the use of color entirely using the new option --no-color-diagnostics or by setting the NO_COLOR environment variable. For the details, please see the new Diagnostic output chapter of the Mercury Users' guide.

  • We have added a new option --warn-ambiguous-pragmas that tells the compiler to generate warnings for pragmas that specify a name/arity pair, when there is both a predicate and a function with that name and arity, and the pragma does not specify which one it is for.

  • We have added a new option --warn-redundant-coerce, that tells the compiler to generate warnings for type conversions from one type to the same type. This option is enabled by default.

  • We have added a new option --warn-repeated-singleton-variables, which may be shortened to --warn-repeated-singleton-vars, that tells the compiler to generate warnings for variables whose name starts with an underscore, indicating that they should be singletons, that nevertheless occur more than once. Until now, warnings about such variables were controlled by the --warn-singleton-variables option. This option is enabled by default.

  • We have raised the default value of --output-compile-error-lines to 100.

  • We have added a new option --reverse-error-order that tells the compiler to output error messages for higher line numbers before error messages for lower line numbers.

  • We have added a new option --show-local-call-tree that tells the compiler to output a representation of the call tree of the predicates and functions of the module as ordered by a depth-first left-to-right traversal of the bodies of the first (i.e. main) procedures of those predicates and functions. This can suggest a good order for the positions of those predicates and functions in the source code.

  • We have added a new option --show-pred-movability. If a user specifies this option one or more times, each time with the name of a predicate or function as the option's argument, then the compiler will generate a report about the effect of moving the named predicates and/or functions to a new module. This report will list all the other predicates and/or functions that should be moved with them, and will warn about any unwanted coupling the move would create between the old module and the new.

  • We have added a new option `--warn-can-fail-function'. If the user specifies this option, the compiler will generate a warning for every function whose primary mode (which means the mode in which all its arguments are input) can fail without generating a return value.

    This option is for people who prefer that all functions be total, and that functions that compute a value only in some circumstances (i.e. partial functions) be expressed as predicates instead.

  • The deprecated option --trail-segments has been deleted and the grade component trseg is no longer accepted as a synonym for tr.

  • --javac-flags is now accepted as a synonym for --java-flags. Similarly --javac-flag is now accepted as a synonym for --java-flags.

  • The new options --java-runtime-flags and --java-runtime-flags can be used to set flags for the Java interpreter in the launcher shell scripts or batch files generated by the compiler when creating executables in the java grade.

  • On MacOS systems, the configure script now uses the value of the MACOSX_DEPLOYMENT_TARGET environment variable as the default value for the deployment target, if set.

  • The --no-ansi-c option has been deprecated and will be removed in a future release. The option has not had any effect for a long time.

  • We have deleted the --java-object-file-extension option, which has only ever had one useful value, which is the default value, ".class".

  • We have deleted the --generate-bytecode option, which has never been useful, because the bytecode interpreter it generated code for was never complete enough to be usable.

  • The --install-command-dir-option option has been deprecated and will be removed in a future release. This option has not had any effect for a long time.

Changes to the Mercury debugger

  • The dump command has a new option: dump -p Var will dump the value of the given variable in a prettyprinted form.

  • The stack command now identifies cliques by drawing boxes next to them on the left hand side of the screen.

Changes to the Mercury deep profiler

  • Programs that have been compiled in a deep profiling grade will now, by default, put profiling data in files whose names contain both the name of the program being profiled, and the date and time of the profiling run. The file names will have the form <prog>_on_<date>_at_<time>.data and <prog>_on_<date>_at_<time>.procrep.

    To get the old behaviour, pass the runtime option --deep-std-name via the MERCURY_OPTIONS environment variable. This will put the profiling data in files named Deep.data and Deep.procrep.

Changes to the Mercury implementation

  • We have upgraded the bundled Boehm GC to v8.2.8 and libatomic_ops to v7.8.2.

Portability improvements

  • We now support using the 64-bit (i.e. x64) version of Microsoft Visual C++ as a C compiler with Mercury. See README.MS-VisualC.md for further details.

  • We have updated the script tools/configure_cross to support cross-compiling using clang.

Changes to the extras distribution

  • We have deleted the old_term_parser library.

Other changes

  • We have added a new option, --with-pkgversion, to the configure script. This option lets you specify an identifying string that will be included in the version information displayed by programs in the Mercury system.

NEWS for Mercury 22.01.8

This is a bug-fix release.

  • [Mantis bug #565]. We have reverted a change that broke mmake in Mercury versions 22.01.6 and 22.01.7.

  • We have fixed a problem that was causing the generation of the debugger documentation to fail on Cygwin.

  • We have increased the C stack size for the compiler on 64-bit Cygwin systems, due to stack overflows occurring with the default size.

  • We have fixed a bug in the debugger, which, in some rare circumstances, could cause it to print terms to the wrong stream.

NEWS for Mercury 22.01.7

This is a bug-fix release.

  • [Github issue #123]. We have fixed a problem that was causing compilation with GCC 13 to fail.

  • We have renamed a macro used in the compiler to avoid a clash with the system headers in the MSYS2 UCRT64 environment.

  • We have fixed a problem that was causing the generation of the debugger documentation to fail on Windows.

  • We have fixed a problem where MSVC could not compile the C code generated for some tabled Mercury predicates. This was caused by the C code inadvertently making use of some GNU C extensions that MSVC does not support.

NEWS for Mercury 22.01.6

This is a bug-fix release.

  • We have fixed the implementations of transitive closure and reflexive transitive closure in the digraph module.

  • We have a fixed a bug in string.format and related predicates where int8, int16 and int32 values were being incorrectly promoted to int and sign extended when one of the unsigned conversion specifiers was applied.

  • [Github issue #116]. We have fixed a problem building programs with --intermodule-optimization where the compiler was searching for library files that were not installed unless the library was also built with --intermodule-optimization enabled.

  • [Github issue #117]. We have fixed a bug where the compiler would silently fail when compiling a program that consists of a single source file and some of the required libraries are not available in the selected grade.

  • The documentation of the following predicates in the term_io module has been fixed:

    • pred quote_char/3
    • pred quote_atom/3
    • pred write_escaped_char/3
    • pred write_escaped_string/3

    Their documentation said that they write to stdout, when in fact they write to the current text output stream. By default, this is stdout, but calls to io.set_output_stream may change this.

NEWS for Mercury 22.01.5

This is a bug-fix release.

  • We have fixed a bug that was causing invalid fact table file names to be written to dependency files.

  • We have fixed a bug with the MLDS backend that was causing crashes with certain switches on strings.

  • We have fixed a bug that was causing some transitive intermodule optimization interfaces to be silently ignored.

  • We have fixed a bug where termination analysis could cause a compiler abort when invoked on some rarely-occurring predicate structures.

NEWS for Mercury 22.01.4

This is a bug-fix release.

  • We have fixed and clarified the behaviour of the following string module predicates when called on a string containing ill-formed code unit sequences:

    • pred all_match/2
    • pred index_next_repl/5
    • pred unsafe_index_next_repl/5
    • pred prev_index_repl/5
    • pred unsafe_prev_index_repl/5
  • We have fixed a bug in the getopt and getopt_io modules where negating a maybe_string option would set its option table entry to an incorrect value.

  • We have fixed an issue where flags specific to GCC were being passed to MSVC when building Mercury using MSVC.

NEWS for Mercury 22.01.3

This is a bug-fix release.

  • [Github issue #103]. We have disabled the use of some GCC optimizations that were causing segmentation faults in code compiled in the asm_fast grades with GCC versions 11 and 12 on x86_64 systems.

NEWS for Mercury 22.01.2

This is a bug-fix release.

  • [Mantis bug #558]. We have fixed a bug that was causing some of the manual pages to sometimes build incorrectly.

  • [Mantis bug #560]. We have disabled the use of the asm_fast* grades on AArch64 when using GCC 9+ due to incompatibilities with position-independent code.

  • We have fixed a bug that was causing assertion failures or segmentation violations in the Mercury runtime on AArch64 systems with versions of GCC before 10.

  • [Github issue #102]. We have disabled the use of the GCC option -Warray-bounds for GCC version 12 and onwards. This is due to GCC emitting spurious warnings when that option is enabled. (The use of that option was already disabled for GCC versions 9-11.)

NEWS for Mercury 22.01.1

This is a bug-fix release.

  • [Mantis bug #557]. We have fixed a bug in module qualification that was causing the compiler to fail without printing an error message.

NEWS for Mercury 22.01

Changes that may break compatibility

  • We have removed is as a synonym for unification.

  • We have reserved =</2 as a type name.

  • A term with a top-level functor coerce/1 is now treated as a type conversion expression. To call a function named coerce/1, you can module qualify the name at the call site, or wrap parentheses around the name, e.g. (coerce)(Arg).

  • We have renamed the lexer and parser modules of the Mercury standard library to mercury_term_lexer and mercury_term_parser respectively.

  • We have made slight changes to the names and/or the functionality of several predicates in the getopt and getopt_io modules.

  • We have removed the legacy support for the Alpha architecture.

  • We have dropped support for macOS 10.8 and earlier.

  • We have removed the Erlang backend as it was unmaintained.

Changes to the Mercury standard library

New module: random.system_rng

  • This module provides an interface to a platform specific cryptographically secure random number generator that is seeded from the OS entropy pool.

Changes to the array module

  • The following obsolete predicates and functions have been removed:

    • func bsearch/3 (replacement: binary_search/3)
    • pred bsearch/4 (replacement: binary_search/4)
    • func least_index/1 (replacement: min/1)
    • func greatest_index/1 (replacement: max/1)
  • The following predicate has been added:

    • pred generate_foldl2/7

Changes to the array2d module

  • The following predicates and functions have been added:

    • func lookup/3 (synonym for Array ^ elem(R, C))
    • pred lookup/4 (synonym for Array ^ elem(R, C))
    • func unsafe_lookup/3 (synonym for Array ^ unsafe_elem(R, C))
    • pred unsafe_lookup/4 (synonym for Array ^ unsafe_elem(R, C))
  • The lists/1 function now returns an empty list for a 0x0 array.

Changes to the assoc_list module

  • The following predicates and functions have been added:

    • func common_subset/2
    • pred maybe_from_corresponding_lists/3

Changes to the bag module

  • The following obsolete predicates and functions have been removed:

    • func to_set_without_duplicates/1 (replacement: func to_set/1)
    • pred to_set_without_duplicates/2 (replacement: func to_set/1)
  • The following predicates and functions have been added:

    • func from_sorted_list_without_duplicates/1
    • pred from_sorted_list_without_duplicates/2

Changes to the bitmap module

  • The following predicates and functions have been added:

    • func get_bit/2
    • func get_bits/2
    • func get_byte/2
    • pred set_bit/4
    • pred set_bits/4
    • pred set_byte/4
    • func unsafe_get_bit/2
    • func unsafe_get_bits/2
    • func unsafe_get_byte/2
    • pred unsafe_set_bit/4
    • pred unsafe_set_bits/4
    • pred unsafe_set_byte/4

Changes to the char module

  • The following predicates have been added:

    • pred unsafe_base_digit_to_int/3
    • pred to_utf8_uint8/2
    • pred to_utf16_uint16/2
  • The following obsolete predicates and functions have been removed:

    • func det_int_to_digit/1 (replacement: det_int_to_decimal_digit/1)
    • pred det_int_to_digit/2 (replacement: det_int_to_decimal_digit/2)
    • pred digit_to_int/2 (replacement: decimal_digit_to_int/2)
    • pred int_to_digit/2 (replacement: int_to_decimal_digit/2)

Changes to the cord module

  • The following predicates have been added:

    • pred foldl2/6
    • pred foldl3/8

Changes to the dir module

  • The following predicate has been added:

    • pred general_foldl2/8

Changes to the getopt module

  • The following new predicates have been added:

    • pred record_arguments/8
    • pred expand_file_specials/8
  • The following variants of the existing process_options predicate have been added:

    • pred process_options_io/6
    • pred process_options_io/7
    • pred process_options_track_io/9
    • pred process_options_userdata/8
    • pred process_options_userdata_io/10

    The ones whose names have an _io suffix do the same jobs as the corresponding predicates without the _io suffix, with the exception that they also have a pair of I/O state arguments that allow them to implement file_special options. This functionality used to be available only from the getopt_io module.

  • The following predicates have had their argument types changed:

    • pred process_options/6
    • pred process_options/7
    • pred process_options_track/7

    All these predicates used to return error indications in the form of a simple string. They now return error indications using the existing structured type option_error, which can be converted into a string on demand.

  • The following predicates have been deleted:

    • pred process_options_se/6
    • pred process_options_se/7
    • pred process_options_track_se/7

    Their functionality is now available from the predicates with the same name minus the _se suffix.

Changes to the getopt_io module

  • This module has been deprecated. For now, it exports the same functionality as the updated getopt module, but it is scheduled to be deleted after the next release.

Changes to the int module

  • The following functions have been added:

    • func uint_to_lc_hex_string/2 (synonym for uint_to_hex_string/2)
    • func uint64_to_lc_hex_string/2 (synonym for uint64_to_hex_string/2)
  • The following obsolete predicates and functions have been removed:

    • pred is/2 (replacement: =, i.e. unification)
    • func legacy_left_shift/2 (replacement: <<)
    • func legacy_right_shift/2 (replacement: >>)

Changes to the int32 module

  • The following functions have been added:

    • func cast_to_int8
    • func cast_from_int8
    • func cast_to_int16
    • func cast_from_int16
    • func cast_to_int64
    • func cast_from_int64

Changes to the integer module

  • The following obsolete functions have been removed:

    • func from_base_string/2 (replacement: pred from_base_string/3)
    • func from_string/2 (replacement: pred from_string/2)
    • func int/2 (replacement: det_to_int/1)

Changes to the io module

  • The following predicates have been added to this module:

    • pred get_environment_var_map/3
    • pred read_binary_int8_unboxed/5
    • pred read_binary_uint8_unboxed/5
    • pred read_named_file_as_string/4
    • pred read_named_file_as_lines/4
    • pred write_line_cc/4
  • The following obsolete predicates have been removed:

    • pred make_temp/3 (replacement: make_temp_file/3)
    • pred make_temp/5 (replacement: make_temp_file/5)
  • The following predicates have been marked as obsolete:

    • pred see/4 (replacement: prolog.see/4)
    • pred see_binary/4 (replacement: prolog.see_binary/4)
    • pred seen/2 (replacement: prolog.seen/2)
    • pred seen_binary/2 (replacement: prolog.seen_binary/2)
    • pred tell/4 (replacement: prolog.tell/4)
    • pred tell_binary/4 (replacement: prolog.tell_binary/4)
    • pred told/2 (replacement: prolog.told/2)
    • pred told_binary/2 (replacement: prolog.told_binary/2)
  • The following predicate has been renamed:

    • pred report_stats/2 to report_standard_stats/2.

Changes to the lexer module

  • This module has been renamed to mercury_term_lexer, to make name clashes between it and user-written modules less likely.

Changes to the list module

  • The following predicates have been added:

    • pred delete_nth/3
    • pred foldl7/16
    • pred foldl8/18
    • pred foldl4_corresponding/11
  • The following obsolete predicate has been removed:

    • pred takewhile/4 (replacement: take_while/4)

Changes to the map module

  • The following predicates have been added to this module:

    • pred foldl6/14
    • pred foldl6_values/14
    • pred foldr6/14

Changes to the maybe module

  • The following predicates have been added to this module:

    • pred foldl3_maybe/8
    • pred foldl4_maybe/10
    • pred foldl5_maybe/12
    • pred map_foldl4_maybe/11
    • pred map_foldl5_maybe/13

Changes to the parser module

  • This module has been renamed to mercury_term_parser, to make name clashes between it and user-written modules less likely.

Changes to the prolog module

  • The following predicate has been added to this module:

    • pred is/2 (moved here from the int module)
    • pred see/4 (moved here from the io module)
    • pred see_binary/4 (moved here from the io module)
    • pred seen/2 (moved here from the io module)
    • pred seen_binary/2 (moved here from the io module)
    • pred tell/4 (moved here from the io module)
    • pred tell_binary/4 (moved here from the io module)
    • pred told/2 (moved here from the io module)
    • pred told_binary/2 (moved here from the io module)

Changes to the random module

  • The following obsolete predicate has been removed:

    • pred test/4 (replacement: none)

Changes to the std_util module

  • The following obsolete predicates and functions have been removed:

    • func maybe_func/2 (replacement: func maybe.func_to_maybe/1)
    • pred maybe_pred/3 (replacement: func maybe.pred_to_maybe/1)

Changes to the string module

  • The following functions have been added:

    • func add_suffix/2
    • func split_into_lines/1
    • pred to_uint/2
    • func det_to_uint/1
    • pred base_string_to_uint/3
    • func det_base_string_to_uint/2
    • func uint_to_hex_string/1
    • func uint_to_uc_hex_string/1
    • func uint_to_octal_string/1
    • func uint64_to_hex_string/1
    • func uint64_to_uc_hex_string/1
    • func uint64_to_octal_string/1
  • The following function symbols have been added to the type poly_type:

    • i8(int8)
    • i16(int16)
    • i32(int32)
    • i64(int64)
    • u8(uint8)
    • u16(uint16)
    • u32(uint32)
    • u64(uint64)

    This allows predicates such as string.format and io.format to operate on values of not just the word sized integer types int and uint, but on sized versions of them as well.

Changes to the term module

  • The following obsolete predicates and functions have been removed:

    • func var_id/1 (replacement: var_to_int/1)

    • func relabel_variable/3 (replacement: rename_var_in_term/4)

    • pred relabel_variable/4 (replacement: rename_var_in_term/4)

    • func relabel_variables/3 (replacement: rename_var_in_terms/4)

    • pred relabel_variables/4 (replacement: rename_var_in_terms/4)

    • func rename/3 (replacement: rename_var_in_term/4)

    • pred rename/4 (replacement: rename_var_in_term/4)

    • func rename_list/3 (replacement: rename_var_in_terms/4)

    • pred rename_list/4 (replacement: rename_var_in_terms/4)

    • func apply_renaming/3 (replacement: apply_renaming_in_term/3)

    • pred apply_renaming/3 (replacement: apply_renaming_in_term/3)

    • func apply_renaming_to_list/3 (replacement: apply_renaming_in_terms/3)

    • pred apply_renaming_to_list/3 (replacement: apply_renaming_in_terms/3)

    • func apply_variable_renaming/2 (replacement: apply_renaming_in_term/3)

    • pred apply_variable_renaming/3 (replacement: apply_renaming_in_term/3)

    • func apply_variable_renaming_to_list/2 (replacement: apply_renaming_in_terms/3)

    • pred apply_variable_renaming_to_list/3 (replacement: apply_renaming_in_terms/3)

    • func apply_variable_renaming_to_var/2 (replacement: apply_renaming_in_var/3)

    • pred apply_variable_renaming_to_var/3 (replacement: apply_renaming_in_var/3)

    • func apply_variable_renaming_to_vars/2 (replacement: apply_renaming_in_vars/3)

    • pred apply_variable_renaming_to_vars/3 (replacement: apply_renaming_in_vars/3)

    • func substitute/3 (replacement: substitute_var_in_term/4)

    • pred substitute/4 (replacement: substitute_var_in_term/4)

    • func substitute_list/3 (replacement: substitute_var_in_terms/4)

    • pred substitute_list/4 (replacement: substitute_var_in_terms/4)

    • func substitute_corresponding/3 (replacement: substitute_corresponding_in_term/4)

    • pred substitute_corresponding/4 (replacement: substitute_corresponding_in_term/4)

    • func substitute_corresponding_list/3 (replacement: substitute_corresponding_in_terms/4)

    • pred substitute_corresponding_list/4 (replacement: substitute_corresponding_in_terms/4)

    • func apply_substitution/2 (replacement: apply_substitution_in_term/3)

    • pred apply_substitution/3 (replacement: apply_substitution_in_term/3)

    • func apply_substitution_to_list/2 (replacement: apply_substitution_in_terms/3)

    • pred apply_substitution_to_list/3 (replacement: apply_substitution_in_terms/3)

    • func apply_rec_substitution/2 (replacement: apply_rec_substitution_in_term/3)

    • pred apply_rec_substitution/3 (replacement: apply_rec_substitution_in_term/3)

    • func apply_rec_substitution_to_list/2 (replacement: apply_rec_substitution_in_terms/3)

    • pred apply_rec_substitution_to_list/3 (replacement: apply_rec_substitution_in_terms/3)

Changes to the thread module

  • The following predicate and functions have been added:

    • func init_thread_options/0
    • pred set_min_stack_size/3
    • pred spawn_native/5

Changes to the thread.mvar module

  • The following obsolete function has been removed:

    • func init/1 (replacement: impure_init/1)

Changes to the thread.semaphore module

  • The following obsolete function has been removed:

    • func init/1 (replacement: impure_init/1)

Changes to the time module

  • The following obsolete functions have been removed:

    • func ctime/1 (replacement: localtime/4 and asctime/1)
    • func localtime/1 (replacement: localtime/4)
    • func mktime/1 (replacement: mktime/4)

Changes to the tree234 module

  • The following predicates have been added to this module:

    • pred foldl6/14
    • pred foldl6_values/14
    • pred foldr6/14

Changes to the uint16 module

  • The following functions have been added:

    • pred from_uint/2
    • func det_from_uint/1
    • func cast_from_uint/1
    • func rotate_left/2
    • func rotate_right/2
    • func unchecked_rotate_left/2
    • func unchecked_rotate_right/2
    • func set_bit/2
    • func unchecked_set_bit/2
    • func clear_bit/2
    • func unchecked_clear_bit/2
    • func flip_bit/2
    • func unchecked_flip_bit/2
    • func bit_is_set/2
    • func unchecked_bit_is_set/2
    • func bit_is_clear/2
    • func unchecked_bit_is_clear/2
    • func cast_from_uint8/1
    • func cast_to_uint8/1

Changes to the uint32 module

  • The following functions have been added:

    • pred from_uint/2
    • func det_from_uint/1
    • func rotate_left/2
    • func rotate_right/2
    • func unchecked_rotate_left/2
    • func unchecked_rotate_right/2
    • func set_bit/2
    • func unchecked_set_bit/2
    • func clear_bit/2
    • func unchecked_clear_bit/2
    • func flip_bit/2
    • func unchecked_flip_bit/2
    • func bit_is_set/2
    • func unchecked_bit_is_set/2
    • func bit_is_clear/2
    • func unchecked_bit_is_clear/2
    • func cast_from_uint8/1
    • func cast_to_uint8/1
    • func cast_from_uint16/1
    • func cast_to_uint16/1

Changes to the uint64 module

  • The following functions have been added:

    • func cast_from_uint/1
    • func rotate_left/2
    • func rotate_right/2
    • func unchecked_rotate_left/2
    • func unchecked_rotate_right/2
    • func set_bit/2
    • func unchecked_set_bit/2
    • func clear_bit/2
    • func unchecked_clear_bit/2
    • func flip_bit/2
    • func unchecked_flip_bit/2
    • func bit_is_set/2
    • func unchecked_bit_is_set/2
    • func bit_is_clear/2
    • func unchecked_bit_is_clear/2
    • func cast_from_uint8/1
    • func cast_to_uint8/1

Changes to the uint8 module

  • The following functions have been added:

    • pred from_uint/2
    • func det_from_uint/1
    • func cast_from_uint/1
    • func rotate_left/2
    • func rotate_right/2
    • func unchecked_rotate_left/2
    • func unchecked_rotate_right/2
    • func set_bit/2
    • func unchecked_set_bit/2
    • func clear_bit/2
    • func unchecked_clear_bit/2
    • func flip_bit/2
    • func unchecked_flip_bit/2
    • func bit_is_set/2
    • func unchecked_bit_is_set/2
    • func bit_is_clear/2
    • func unchecked_bit_is_clear/2

Changes to the varset module

  • The following obsolete predicates have been removed:

    • pred merge_subst/4 (replacement: merge_renaming/4)
    • pred merge_subst_without_names/4 (replacement: merge_renaming_without_names/4)
  • The following functions and predicates have been added:

    • func lookup_name_default_prefix/3
    • pred lookup_name_default_prefix/4
    • func unname_var/2
    • pred unname_var/3
    • pred undo_default_names/2
  • The following function and predicate have been marked obsolete:

    • func lookup_name/3 (replacement: lookup_name_default_prefix/3)
    • pred lookup_name/4 (replacement: lookup_name_default_prefix/4)
  • The following predicate has had its argument order changed:

    • pred set_bindings/3

Changes to the Mercury language

  • The type system now supports subtypes, which work in tandem with type conversion expressions ("coerce"). For example, the following defines a subtype real_color of a discriminated union type color:

      :- type color
          --->    rgb(float, float, float)
          ;       cmyk(float, float, float, float)
          ;       named(string).
    
      :- type real_color =< color
          --->    rgb(float, float, float)
          ;       cmyk(float, float, float, float).
    

    A term of type real_color can be converted to a term of type color with coerce(Term). A term of type color can be converted to a term of type real_color if it has an appropriate inst. Subtypes share a common data representation with their base types, so the type conversions do not cost anything at runtime.

  • Field names no longer need to be unique within a module.

  • The compiler can implement tabling only when generating C code. When compiling a predicate that has a pragma memo specified for it in a non-C grade, it necessarily ignores the pragma, but normally it prints a warning about this fact. The compiler now supports a new attribute, disable_warning_if_ignored, that suppresses such warnings for a pragma memo if included in the pragma's attribute list, like this:

      :- pragma memo(predname/arity, [disable_warning_if_ignored])]).
    
  • A pragma foreign_proc declaration can now include an attribute may_not_export_body that prevents its body (i.e. the foreign code) from being duplicated outside of the target file for that module by intermodule optimization. This is useful when the foreign code refers to types, functions, etc. that should be kept local to the target file of that module. Unlike the may_not_duplicate attribute, may_not_export_body does not prevent inlining of the foreign procedure into other procedures in the same module.

  • Many pragmas contain a name/arity pair for specifying the predicate or function they apply to; an example is

      :- pragma inline(init/1).
    

    If the module in which this pragma occurs contains both a function init/1 and a predicate init/1, then this pragma is ambiguous. Traditionally, the Mercury compiler applied such ambiguous pragmas to both the function and the predicate. The programmers who wished the pragma to apply to only one of them had to rename the other.

    Now, pragmas that take a name/arity pair can specify whether they are intended to apply to a function or to a predicate by putting a func() or pred() wrapper around the name/arity pair, like this:

      :- pragma inline(func(init/1)).     % applies ONLY to func init/1.
      :- pragma inline(pred(init/1)).     % applies ONLY to pred init/1.
    

    This new syntax applies to all of the following kinds of pragmas:

      check_termination
      consider_used
      does_not_terminate
      fact_table
      inline
      loop_check
      memo
      minimal_model
      mode_check_clauses
      no_determinism_warning
      no_inline
      obsolete
      promise_equivalent_clauses
      promise_pure
      promise_semipure
      require_tail_recursion
      terminates
      type_spec
    

Changes to the Mercury compiler

  • If a command line first enables an optimization (say opt1), and then sets the optimization level to N with -O<N>, then the compiler will now keep opt1 enabled even if opt1 is not normally enabled at optimization level N.

  • Due to a bug fix, the --warn-unused-imports option is now stricter in warning about modules that are imported in the interface section but are not used in the interface section.

  • We have fixed parsing of reverse implication goals (A <= B).

  • By default, the compiler now checks some aspects of a module semantics when generating the .int and .int2 interface files for that module. (For example, it generates error messages for references to undefined types, insts and modes.) This new behavior can be switched off for now with the new option --no-halt-at-invalid-interface. This new option replaces the old --no-print-errors-warnings-when-generating-interface option, but once any problems caused by the new approach have been ironed out, we intend to delete the --halt-at-invalid-interface option as well.

  • In an earlier release, we extended the syntax of :- inst declarations to allow programmers to specify which type constructor's values the inst is meant for. At that time, these functioned only as documentation, but the compiler now reports error messages for situations in which an inst that was declared to be intended for values of one type constructor is applied to values of another type constructor.

  • The new option --output-stdlib-grades outputs the grades in which the Mercury standard library is available with this compiler.

  • By default, the compiler now warns if the module being compiled has a name that shadows that of Mercury standard library module. The warning can be disabled using the new option --no-warn-stdlib-shadowing.

  • The new options --output-java-class-dir and --output-java-class-directory are now supported as synonyms for --output-class-dir.

  • The new option --halt-at-warn-make-interface causes the compiler to treat all warnings as if they were errors when generating interface files.

  • The new option --halt-at-warn-make-opt causes the compiler to treat all warnings as if they were errors when generating optimization files.

  • The new option --warn-potentially-ambiguous-pragma causes the compiler to generate a warning for all pragmas that include a name/arity pair, but lack a func() or pred() wrapper around it to indicate whether they are intended to apply to a function or to a predicate.

Portability improvements

  • We have ported Mercury to Linux AArch64 (ARM64) systems.

  • The tools/configure_mingw_cross script has been replaced by tools/configure_cross. It now supports aarch64-linux-gnu and aarch64-linux-musl as targets (i.e. Linux on aarch64 with GNU or musl C libraries).

Changes to the Mercury debugger

  • The list command may now call an external command to print source listings; the command is set using list_cmd. For example, the command could produce syntax highlighted source listings.

  • We have removed support for browsing terms as XML (browse --xml) as it was unmaintained and did not work any more. The browse --web command provides an alternative method for interactively exploring a term.

For news about earlier versions, see the HISTORY file.