99a718c781
Accepting request 1238626 from home:StefanBruens:branches:science
Stefan Brüns2025-01-19 04:28:38 +00:00
275b49b3d9
- update to 2.8.7: * Bug fixes: + Fix bug that caused where clauses to be removed from comprehensions. + Fix optimisation of Boolean constraints that could sometimes trigger an assertion failure in debug builds. - update to 2.8.6: * Changes: + Variable in operators on arrays of (nested) records and tuples containing only enumerable (enum, int, or bool) will now be rewritten into table global constraints. + Values that are par in the left hand side will be checked and filtered from the resulting table constraint. + Variable <, <=, > and >= operators on (nested) records and tuples containing only enumerable (enum, int, or bool) will now be rewritten into lex_less or lex_lesseq global constraints. + Automatically detect SCIP 9.0.1 and potential future versions on Windows. + The interface to the HiGHS solver now requires version 1.7.2 or later. + Enforce strict enum type correctness for set operations. + Add par opt overloads of min / max and return enum values instead of coercing to integers. + Use half-reification only when there isn’t a reification that is a more specific match. + Add multidimensional overloads for the ++ array concatenation operator. * Bug fixes: + Fix incorrect type error for array literals where the first element is <> and subsequent elements are sets. + Fix missing code block in visualisation feature documentation. + Emit type error when using variable if-then-else/array access containing tuples or records containing non-varifiable types. + Emit type error when using a variable comprehension with a tuple/record type. + Emit type error when using optional array access for arrays of tuples/records. + Emit type error for array literals with <> and var set elements. + Fix missing cross-references for IDE preference dialog documentation figures. + Fix crash when using nested tuples/record types inside var if-then-else expressions. + Fix evaluation error for comprehensions with an initial assignment generator and a where clause which doesn’t involve any of the generator identifiers. + Fix type checking of if then else expressions with _ in the else branch. + Fix parsing of the most negative integer literal. + Fix typo in warning messages for undefined results in function calls. + Fix bug in flattening identifier causing debug assertion to fail. + Fix missing type when flattening indexed var comprehension leading to error or incorrect value type. + Fix crash when generating error message for mismatched index set where an enum index set is empty. + Fix crash due to incorrect handling of negated constraints matching entry in CSE map. + Fix type error due to missing overload of max. + Fix handling of contexts and maybe_partial annotations when flattening par expressions. + Fix par evaluation of boolean operators where an operand is an array access into an empty array literal. + Fix crash when concatenating arrays of annotations. + Ensure DZN output of enum index sets of multidimensional arrays is properly quoted. + Fix parsing of enumerated values in JSON for integer parameters. + Fix crash during flattening of clauses with opt bool arguments. + Fix the coercions of index set for empty array literals to allow any of the index sets to be empty. + Fix behaviour of overloading for the ++ operator. + Avoid outputting objective statistic when value is invalid for MIP solvers. + Fix bug in unification of aliased output variables when using -O2 and above.
Stefan Brüns2025-01-19 04:28:38 +00:00
563b5a08ad
Accepting request 1200255 from science
Ana Guerrero2024-09-13 12:34:12 +00:00
7ff4a88f33
Accepting request 1200255 from science
Ana Guerrero2024-09-13 12:34:12 +00:00
242872d3b9
Accepting request 1199473 from home:dirkmueller:Factory
Dmitry Roshchin2024-09-12 06:48:05 +00:00
3d1d9f345a
- update to 2.8.5: * Fix specification for constraint items and annotations in FlatZinc JSON. * Fix flattening of negated let expressions with constraints. * Fix crash when equating identifiers to tuples/records in if- then-else expressions. * Fix string_length function to return the number of Unicode code points rather than the number of bytes (i.e., to interpret the string as UTF-8). * Emit type error for opt tuples and records created through type aliases. * Fix evaluation of par partial functions returning arrays. * Fix type inference for if-then-else expressions with different tuple or record types in the branches. * Fix evaluation error caused by removal of fixed FlatZinc variables without fixing them in the output model. * Fix computed type when flattening cv comprehensions containing tuple or record types. * Fix unsatisfiability caused by reification in negated boolean context (:bugref:813). * Fix bug in computation of common type of incompatible record types. * Fix crash when type checking nested arrays of tuples or records. * Fix incorrect unification of flattened tuple/record fields with paths enabled. * Add string_split function that splits a string into an array of strings based on a separator. * Add json_section annotation to mark an output item as a JSON section (:bugref:816).
Dmitry Roshchin2024-09-12 06:48:05 +00:00
Dominique Leuenberger
Stefan Brüns
Ana Guerrero
Dmitry Roshchin
Dirk Stoecker
Richard Brown
Matthias Mailänder