Mercury Bugs - Issues

0000583: crash in LLDS .par grades

Tue, 14 Apr 2026 12:50:30 +1000

Here are some results of trying to bootcheck the asm_fast.par.gc grade on various Linux distributions. All tests were performed in a container on the same Linux host kernel, on a x86-64 machine.

--------

Debian 12
GNU ld (GNU Binutils for Debian) 2.40
gcc version 12.2.0 (Debian 12.2.0-14+deb12u1)
- mercury-srcdist-rotd-2026-03-19 - ok

Debian 13
GNU ld (GNU Binutils for Debian) 2.44
gcc version 14.2.0 (Debian 14.2.0-19)
- mercury-srcdist-rotd-2026-03-19 - ok

--------

Ubuntu 20.04
GNU ld (GNU Binutils for Ubuntu) 2.34
gcc version 9.4.0 (Ubuntu 9.4.0-1ubuntu1~20.04.2)
- mercury-srcdist-rotd-2026-03-19 - ok

Ubuntu 22.04
GNU ld (GNU Binutils for Ubuntu) 2.38
gcc version 11.4.0 (Ubuntu 11.4.0-1ubuntu1~22.04.3)
- mercury-srcdist-rotd-2026-03-19 - ok

Ubuntu 24.04
GNU ld (GNU Binutils for Ubuntu) 2.42
gcc version 11.5.0 (Ubuntu 11.5.0-1ubuntu1~24.04.1)
- mercury-srcdist-rotd-2026-03-19 - building of stage 3 dependencies not successful
gcc version 12.4.0 (Ubuntu 12.4.0-2ubuntu1~24.04.1)
- mercury-srcdist-rotd-2026-03-19 - building of stage 3 dependencies not successful
gcc version 13.3.0 (Ubuntu 13.3.0-6ubuntu2~24.04.1) - default
- mercury-srcdist-rotd-2026-03-19 - building of stage 3 dependencies not successful
gcc version 14.2.0 (Ubuntu 14.2.0-4ubuntu2~24.04.1)
- mercury-srcdist-rotd-2026-03-19 - building of stage 3 dependencies not successful

--------

Alpine Linux 3.20
GNU ld (GNU Binutils) 2.42
gcc version 13.2.1 20240309 (Alpine 13.2.1_git20240309)
- mercury-srcdist-rotd-2026-03-19 - ok

Alpine Linux 3.22
GNU ld (GNU Binutils) 2.44
gcc version 14.2.0 (Alpine 14.2.0)
- mercury-srcdist-rotd-2026-03-19 - ok

Alpine Linux 3.23
GNU ld (GNU Binutils) 2.45.1
gcc version 15.2.0 (Alpine 15.2.0)
- mercury-srcdist-rotd-2026-03-19 - ok

--------

AlmaLinux 9
GNU ld version 2.35.2-67.el9_7.1
gcc version 11.5.0 20240719 (Red Hat 11.5.0-11) (GCC)
- mercury-srcdist-rotd-2026-03-19 - ok

AlmaLinux 10
GNU ld version 2.41-58.el10_1.2.alma.1
gcc version 14.3.1 20250617 (Red Hat 14.3.1-2) (GCC) - default
- mercury-srcdist-rotd-2026-03-19 - building of stage 3 dependencies not successful
gcc version 15.1.1 20250521 (Red Hat 15.1.1-2) (GCC)
- mercury-srcdist-rotd-2026-03-19 - building of stage 3 dependencies not successful

--------

Fedora 43
GNU ld version 2.45.1-4.fc43
gcc version 14.3.1 20250808 (Red Hat 14.3.1-1) (GCC)
- mercury-srcdist-rotd-2026-03-19 - ok
gcc version 15.2.1 20260123 (Red Hat 15.2.1-7) (GCC) - default
- mercury-srcdist-rotd-2026-03-19 - ok

--------

OpenSuSE Leap 15.6
GNU ld (GNU Binutils; SUSE Linux Enterprise 15) 2.45.0.20251103-150100757
gcc version 7.5.0 (SUSE Linux)
- mercury-srcdist-rotd-2026-03-19 - ok

OpenSuSE Tumbleweed
GNU ld (GNU Binutils; openSUSE Tumbleweed) 2.45.0.20251103-2
gcc version 15.2.1 20260202 (SUSE Linux)
- mercury-srcdist-rotd-2026-03-19 - building of stage 3 dependencies not successful

--------

Void Linux
GNU ld (GNU Binutils) 2.44
gcc version 14.2.1 20250405 (GCC)
- mercury-srcdist-rotd-2026-03-19 - ok

--------

I'm not sure what conclusions we can draw. The problem appears to be something particular to a distribution, rather than an issue with certain gcc or binutils versions. (But keeping in mind that distributions will have applied their own patches to those software packages.)

0000584: exception in high-level code generator when type referenced via import in parent module

Thu, 02 Apr 2026 17:23:44 +1100

The compiler crashes in the MLDS backend on the attached test case, where mod_a.sub1 references a type in a sibling module, relying on a import_module declaration in the parent module for visibility of that type.

% mmc -s hlc.gc -m bugmain.all_cs
Making Mercury/int3s/bugmain.int3
Making Mercury/int3s/mod_a.int3
Making Mercury/int3s/mod_a.sub2.int3
Making Mercury/int3s/mod_a.sub1.int3
Making Mercury/int0s/mod_a.int0
Making Mercury/ints/mod_a.sub2.int
Making Mercury/ints/mod_a.sub1.int
Making Mercury/ints/mod_a.int
Making Mercury/ints/bugmain.int
Making Mercury/cs/bugmain.c
Uncaught Mercury exception:
Software Error: map.lookup: key not found
Key Type: string
Key Value: "bbb"
Value Type: tree234.tree234(parse_tree.prog_data.type_ctor, hlds.hlds_data.hlds_type_defn)
Stack dump follows:
0 pred exception.throw/1-0 (erroneous) (exception.m:312)
1 pred require.error/1-0 (erroneous) (require.m:173)
2 pred require.report_lookup_error/3-0 (erroneous) (require.m:277)
3 pred map.lookup/3-0 (det) (map.m:1463)
4 pred hlds.hlds_data.lookup_type_ctor_defn/3-0 (det) (hlds_data.m:598)
5 func hlds.type_util.classify_type_ctor/2-0 (det) (type_util.m:1095)
6 func ml_backend.mlds.mercury_type_to_mlds_type/2-0 (det) (mlds.m:2870)
7 pred ml_backend.ml_code_util.ml_gen_local_var_decl/6-0 (det) (ml_code_util.m:956)
8 pred ml_backend.ml_code_gen.ml_gen_local_var_decls/6-0 (det) (ml_code_gen.m:1393)
9 pred ml_backend.ml_code_gen.ml_gen_goal/7-0 (det) (ml_code_gen.m:611)
10 pred ml_backend.ml_proc_gen.ml_gen_proc_body/10-0 (det) (ml_proc_gen.m:1701)
11 pred ml_backend.ml_proc_gen.ml_gen_proc/12-0 (det) (ml_proc_gen.m:630)
12 pred list.foldl3/8-0 (det) (list.m:4236)
13 pred ml_backend.ml_proc_gen.ml_gen_scc_code/11-0 (det) (ml_proc_gen.m:328)
14 pred ml_backend.ml_proc_gen.ml_gen_scc/13-0 (det) (ml_proc_gen.m:233)
15 pred ml_backend.ml_proc_gen.ml_gen_sccs/13-0 (det) (ml_proc_gen.m:217)
16 pred ml_backend.ml_proc_gen.ml_gen_preds/10-0 (det) (ml_proc_gen.m:123)
17 pred ml_backend.ml_top_gen.ml_code_gen/7-0 (det) (ml_top_gen.m:85)
18 pred top_level.mercury_compile_mlds_back_end.hlds_to_mlds/9-0 (det) (mercury_compile_mlds_back_end.m:133)
19 pred top_level.mercury_compile_augment.choose_and_execute_backend_passes/15-0 (det) (mercury_compile_augment.m:479)
20 pred top_level.mercury_compile_augment.after_front_end_passes/14-0 (det) (mercury_compile_augment.m:412)
21 pred top_level.mercury_compile_augment.process_augmented_module/16-0 (det) (mercury_compile_augment.m:283)
22 pred top_level.mercury_compile_augment.augment_and_process_module/14-0 (det) (mercury_compile_augment.m:209)
23 pred list.map_foldl3/9-0 (det) (list.m:4575)
24 pred top_level.mercury_compile_augment.augment_and_process_all_submodules/15-0 (det) (mercury_compile_augment.m:158)
25 pred top_level.mercury_compile_augment.augment_and_process_source_file/17-0 (det) (mercury_compile_augment.m:136)
26 pred top_level.mercury_compile_main.read_augment_and_process_module/15-0 (det) (mercury_compile_main.m:1150)
27 pred top_level.mercury_compile_main.do_process_compiler_arg/13-0 (det) (mercury_compile_main.m:788)
28 pred top_level.mercury_compile_main.do_process_compiler_cmd_line_args/15-0 (det) (mercury_compile_main.m:606)
29 pred top_level.mercury_compile_main.do_op_mode_args/13-0 (det) (mercury_compile_main.m:418)
30 pred top_level.mercury_compile_main.do_op_mode/11-0 (det) (mercury_compile_main.m:236)
31 pred top_level.mercury_compile_main.main_after_setup/6-0 (det) (mercury_compile_main.m:191)
32 pred top_level.mercury_compile_main.main_for_make/6-0 (det) (mercury_compile_main.m:147)
33 pred make.module_target.call_mercury_compile_main/7-0 (det) (make.module_target.m:613)
34 pred libs.process_util.call_child_process_io_pred/4-0 (det) (process_util.m:458)
reached unknown label
** Error making `Mercury/cs/bugmain.c'.

% mdb mmc -s hlc.gc -C bugmain --use-subdirs
Melbourne Mercury Debugger, mdb version DEV.
Copyright 1998-2012 The University of Melbourne.
Copyright 2013-2026 The Mercury team.
mdb is free software; there is absolutely no warranty for mdb.
1: 1 1 CALL pred mercury_compile.main/2-0 (det) mercury_compile.m:31
mdb> e
697392: 243564 32 EXCP pred map.lookup/3-0 (det) e;c2; map.m:1463 (hlds_data.m:598)
mdb> P
Map (arg 1) three("float_box", two(type_ctor(qualified(unqualified("private_builtin"), "float_box"), 0), hlds_type_defn/9, empty, empty), "struct", two/4, three/7, three/7, three/7)
K (arg 2) "bbb"

0000582: generate int switches for other integer types

Wed, 04 Mar 2026 20:09:32 +1100

The compiler ought to be able to generate all the different switch types on all integer types, not just `int'.

We noticed this deficiency on a fact table of the form:

:- pred action_table(int16::in, int16::in, uint16::out) is semidet.
action_table(0i16, 0i16, 1661u16).
action_table(0i16, 1i16, 1661u16).
action_table(0i16, 2i16, 1661u16).
action_table(0i16, 3i16, 1661u16).
action_table(0i16, 4i16, 1661u16).
action_table(0i16, 5i16, 1661u16).
action_table(0i16, 6i16, 1661u16).
action_table(0i16, 7i16, 1661u16).
action_table(0i16, 8i16, 1661u16).
action_table(0i16, 9i16, 1661u16).
...

In the end, we replaced the input arguments with `int's (which was fine, but shouldn't be necessary).

0000581: abort with Unexpected: SubDisjunctions0 != no

Tue, 17 Feb 2026 01:31:17 +1100

Compiling the attach program with:

mmc --allow-stubs -C zinc_rt_driver.m

results in:

Uncaught Mercury exception:
Software Error: predicate `check_hlds.scout_disjunctions.scout_disjunctions_in_goal'/10: Unexpected: SubDisjunctions0 != no
Stack dump not available in this grade.

0000580: parse foreign_proc body into comments vs non-comments

Fri, 30 Jan 2026 23:59:15 +1100

It should be relatively easy to parse C/C#/Java foreign_proc bodies into comments and non-comment sections. Then we could avoid false positive warnings about "return" statements when the word "return" appears within a comment.

There are cases where we deliberately name a variable within a comment to suppress a warning about a variable not occurring, but perhaps in those cases the variable would have an underscore prefix, e.g.

:- pragma foreign_proc("C#",
min(_Array::in, Min::out),
[will_not_call_mercury, promise_pure, thread_safe],
"
// Array not used.
Min = 0;
").

Also, in C at least, that code could have been written as (void)Array;

0000579: Unrecognised options in Mercury.options files not reported correctly

Mon, 15 Dec 2025 04:21:17 +1100

Attempting to compile a program with mmc --make (e.g. samples/hello.m) and a Mercury.options file containing the following:

MCFLAGS=--i-am-not-an-option

causes the compiler to print:

mercury_compile:
Error: the location of the directory that holds the Mercury standard library is
not specified
either by an `--mercury-stdlib-dir' option,
or by an environment variable named `MERCURY_STDLIB_DIR',
or by a make variable named `MERCURY_STDLIB_DIR' in any specified options

0000578: Switch detection regression

Thu, 20 Nov 2025 20:06:50 +1100

After commit c9e549996 ("Let switch detection handle deeper disjunctions."), the compiler no longer accepts the given test case.

% mmc -e switch_detection_regression.m
switch_detection_regression.m:017: Error: the determinism declaration for
switch_detection_regression.m:017: predicate `halign'/5 is not satisfied.
switch_detection_regression.m:017: Declared `det', inferred `nondet'.
switch_detection_regression.m:017: The reasons for the difference are the
switch_detection_regression.m:017: following.
switch_detection_regression.m:029:
switch_detection_regression.m:029: Unification of `Dir' and
switch_detection_regression.m:029: `switch_detection_regression.ltr' can
switch_detection_regression.m:029: fail.
switch_detection_regression.m:031:
switch_detection_regression.m:031: Unification of `Dir' and
switch_detection_regression.m:031: `switch_detection_regression.rtl' can
switch_detection_regression.m:031: fail.
switch_detection_regression.m:032:
switch_detection_regression.m:032: Inside the case end/0 of the switch on
switch_detection_regression.m:032: TextAlign:
switch_detection_regression.m:032: disjunction has more than one disjunct
switch_detection_regression.m:032: with solutions.
switch_detection_regression.m:041:
switch_detection_regression.m:041: This later disjunct may have a solution.
switch_detection_regression.m:032:
switch_detection_regression.m:032: Inside the case start/0 of the switch on
switch_detection_regression.m:032: TextAlign:
switch_detection_regression.m:032: disjunction has more than one disjunct
switch_detection_regression.m:032: with solutions.
switch_detection_regression.m:041:
switch_detection_regression.m:041: This later disjunct may have a solution.
switch_detection_regression.m:038:
switch_detection_regression.m:038: Unification of `Dir' and
switch_detection_regression.m:038: `switch_detection_regression.ltr' can
switch_detection_regression.m:038: fail.
switch_detection_regression.m:040:
switch_detection_regression.m:040: Unification of `Dir' and
switch_detection_regression.m:040: `switch_detection_regression.rtl' can
switch_detection_regression.m:040: fail.

0000575: type qualification scope bug

Wed, 13 Aug 2025 07:35:47 +1000

Explicit type qualification is processed before quantification, leading to variables being given explicit types without taking into account that variables with the same name might be in different scopes.

:- module type_qual_bug.
:- interface.

:- pred bug1 is det.
:- pred bug2 is det.

:- implementation.

:- type foo1
---> foo(int).

:- type foo2
---> foo(int).

bug1 :-
% X is assigned the type foo1 due to an explicit type qualification.
% Later, when the Xs are renamed apart, the type is copied over to both
% new variables. Obviously, the two Xs are not supposed to be related.
some [X] (
X = foo(1) : foo1
),
some [X] (
X = foo(2) : foo2
).

bug2 :-
% Both X variables end up with type foo2, so this compiles.
% But the type of the first X should be ambiguous.
some [X] (
X = foo(1)
),
some [X] (
X = foo(2) : foo2
).

0000577: composite enums

Wed, 13 Aug 2025 03:27:29 +1000

Types such as the ext type in the compiler, which have

- several function symbols,
- all of which one argument,
- and all those arguments have types that are effectively enums,

should be themselves represented as one large enum.

Given e.g.

:- type t
---> f(t1)
; g(t2).

where t1 is an enum with values 0-12 and t2 is an enum
with values 0-42, then t should be such a "composite enum"
with values 0-55, with values 0-12 corresponding to function
symbol f, and values 13-55 corresponding to g (with the value
of g's argument being t's value minus 13.

The natural way to do switches on such composite enums
would be by binary search, not on the raw values, but on
the dividing lines between the integer ranges assigned to
the function symbols. Each binary step would rule out half
(rounded down) of the remaining possible function symbols.

0000576: Abort in pre_typecheck

Sat, 05 Jul 2025 19:54:52 +1000

Compiling the attached module causes the following abort:

Uncaught Mercury exception:
Software Error: function `check_hlds.pre_typecheck.arg_num_pieces'/2: Unexpected: ArgNum is not numbered correctly for return value

0000574: Higher-order inst gets lost when using combined higher-order type and inst

Mon, 06 May 2024 00:46:04 +1000

The attached program is a very cut-down version of a problem reported by a user on twitter.
Compiling the attached results in:

bug574.m:017: In clause for `process_job(in, di, uo)':
bug574.m:017: in argument 1 of call to predicate `bug574.do_process_job1'/3:
bug574.m:017: mode error: variable `Pred' has instantiatedness `ground',
bug574.m:017: expected instantiatedness was `(pred(in, di, uo) is det)'.

I can see no reason why Pred has that instantiatedness, unless it is some limitation of the combined higher-order types and insts implementation or an actual bug.

0000573: mmc --make regression remaking .opt files unnecessarily

Fri, 23 Feb 2024 16:44:43 +1100

Commit 8b5cc5125 ("Use string builders for .*opt files.")
introduces a regression where .opt files are remade unnecessarily.
Using the samples/diff program for example:

% mmc --intermod-opt -j32 -m diff
[... trimmed output ...]
Making diff

% touch match.m

% mmc --intermod-opt -m diff
Making Mercury/int3s/match.int3
Making Mercury/ints/match.int
Making Mercury/opts/match.opt
** Nothing to be done for `diff'.

% mmc --intermod-opt -m diff
Making Mercury/int3s/match.int3
Making Mercury/opts/match.opt
** Nothing to be done for `diff'.

where every subsequent invocation of mmc --make will keep remaking match.opt.

0000572: Code generator abort in debugging grades

Tue, 20 Feb 2024 17:38:33 +1100

The attached is a cut-down version of a program that causes an abort in debugging grades originally reported by Volker Wysk on the reviews list on 3 January 2024. Compiling it in a debugging grade results in:

Uncaught Mercury exception:
Software Error: predicate `ll_backend.liveness.require_equal'/4: Unexpected: branches of if-then-else disagree on liveness
First: ArgDesc_3, Values_4, MDef_5, Default_6, TypeClassInfo_for_argument_21, TypeInfo_22_22
Rest: ArgDesc_3, Values_4, MDef_5, Default_6, TypeInfo_22_22

0000571: Improve unsatisfiable typeclass constraint errors

Fri, 09 Feb 2024 15:29:30 +1100

The attached program generates the following error messages:

unsat_error.m:015: In clause for predicate `create_schedule'/3:
unsat_error.m:015: unsatisfiable typeclass constraint:
unsat_error.m:015: `unsat_error.logger(L)'.

For cases like this it would better if we could provide the user some indication of the cause of this constraint being unsatisfiable, namely that the there is no constraint on the type variable Logger in the predicate declaration.

0000570: merge switch after switch bug

Wed, 13 Dec 2023 19:06:42 +1100

The compiler aborts during the deforestation, when simplify tries to merge a switch after switch:

% mmc -C --deforestation bug.m
Uncaught Mercury exception:
Software Error: map.lookup: key not found
Key Type: parse_tree.prog_data.cons_id
Key Value: cons(qualified(unqualified("bug"), "prepare_edit"), 0, type_ctor(qualified(unqualified("bug"), "prepare_temp"), 0))
Value Type: uint
Stack dump follows:
0 pred exception.throw/1-0 (erroneous) (exception.m:312)
1 pred require.error/1-0 (erroneous) (require.m:173)
2 pred require.report_lookup_error/3-0 (erroneous) (require.m:277)
3 pred map.lookup/3-0 (det) (map.m:1333)
4 pred check_hlds.simplify.simplify_goal_conj.build_maps_second_switch_cons_id/5-0 (det) (simplify_goal_conj.m:663)
5 pred check_hlds.simplify.simplify_goal_conj.build_maps_second_switch/7-0 (det) (simplify_goal_conj.m:648)
6 pred check_hlds.simplify.simplify_goal_conj.try_to_merge_switch_after_switch/6-0 (det) (simplify_goal_conj.m:586)
7 6* pred check_hlds.simplify.simplify_goal_conj.simplify_conj/10-0 (det) (simplify_goal_conj.m:285 and others)
13 pred check_hlds.simplify.simplify_goal_conj.simplify_goal_plain_conj/10-0 (det) (simplify_goal_conj.m:82)
14 pred check_hlds.simplify.simplify_goal.simplify_goal_expr/10-0 (det) (simplify_goal.m:331)
15 pred check_hlds.simplify.simplify_goal.simplify_goal/8-0 (det) (simplify_goal.m:285)
16 2* pred check_hlds.simplify.simplify_goal_switch.simplify_switch_cases/15-0 (det) (simplify_goal_switch.m:278 and others)
18 pred check_hlds.simplify.simplify_goal_switch.simplify_goal_switch/10-0 (det) (simplify_goal_switch.m:85)
19 pred check_hlds.simplify.simplify_goal.simplify_goal_expr/10-0 (det) (simplify_goal.m:344)
20 pred check_hlds.simplify.simplify_goal.simplify_goal/8-0 (det) (simplify_goal.m:285)
21 pred check_hlds.simplify.simplify_proc.do_simplify_top_level_goal/6-0 (det) (simplify_proc.m:708)
22 pred check_hlds.simplify.simplify_proc.simplify_top_level_goal/7-0 (det) (simplify_proc.m:611)
23 pred check_hlds.simplify.simplify_proc.simplify_goal_update_vars_in_proc/11-0 (det) (simplify_proc.m:201)
24 pred transform_hlds.pd_util.pd_simplify_goal/5-0 (det) (pd_util.m:264)
25 pred transform_hlds.deforest.push_goal_into_goal/8-0 (det) (deforest.m:1844)
26 pred transform_hlds.deforest.handle_deforestation/9-0 (det) (deforest.m:761)
27 3* pred transform_hlds.deforest.deforest_conj/6-0 (det) (deforest.m:566 and others)
30 pred transform_hlds.deforest.deforest_goal_expr/6-0 (det) (deforest.m:319)
31 pred transform_hlds.deforest.deforest_goal/4-0 (det) (deforest.m:294)
32 pred transform_hlds.deforest.deforest_proc_deltas/5-0 (det) (deforest.m:230)
33 pred transform_hlds.deforest.deforest_proc/3-0 (det) (deforest.m:204)
34 5* pred list.foldl/4-0 (det) (list.m:3731 and others)
39 pred transform_hlds.deforest.deforest_module/3-0 (det) (deforest.m:132)
40 pred top_level.mercury_compile_middle_passes.maybe_deforestation/7-0 (det) (mercury_compile_middle_passes.m:1130)
41 pred top_level.mercury_compile_middle_passes.middle_pass/10-0 (det) (mercury_compile_middle_passes.m:232)
42 pred top_level.mercury_compile_main.after_front_end_passes/15-0 (det) (mercury_compile_main.m:2177)
43 pred top_level.mercury_compile_main.process_augmented_module/17-0 (det) (mercury_compile_main.m:2052)
44 pred top_level.mercury_compile_main.augment_and_process_module/15-0 (det) (mercury_compile_main.m:1926)
45 pred list.map_foldl3/9-0 (det) (list.m:4080)
46 pred top_level.mercury_compile_main.augment_and_process_all_submodules/16-0 (det) (mercury_compile_main.m:1872)
47 pred top_level.mercury_compile_main.read_augment_and_process_module_ok/17-0 (det) (mercury_compile_main.m:1730)
48 pred top_level.mercury_compile_main.read_augment_and_process_module/15-0 (det) (mercury_compile_main.m:1672)
49 pred top_level.mercury_compile_main.do_process_compiler_arg/13-0 (det) (mercury_compile_main.m:1238)
50 pred top_level.mercury_compile_main.setup_and_process_compiler_arg/17-0 (det) (mercury_compile_main.m:1102)
51 pred top_level.mercury_compile_main.setup_and_process_compiler_cmd_line_args/19-0 (det) (mercury_compile_main.m:1013)
52 pred top_level.mercury_compile_main.do_op_mode_args/16-0 (det) (mercury_compile_main.m:859)
53 pred top_level.mercury_compile_main.do_op_mode/14-0 (det) (mercury_compile_main.m:662)
54 pred top_level.mercury_compile_main.main_after_setup/9-0 (det) (mercury_compile_main.m:614)
55 pred top_level.mercury_compile_main.real_main_after_expansion/5-0 (det) (mercury_compile_main.m:315)
56 pred top_level.mercury_compile_main.real_main/2-0 (det) (mercury_compile_main.m:153)
57 pred mercury_compile.main/2-0 (det) (mercury_compile.m:32)

0000569: timestamp regression with nested submodules in mmc --make

Wed, 06 Dec 2023 12:51:54 +1100

Here is a problem with nested submodules attributed to commit 4d4ef3e3a "Cache timestamps by target file."

This sequence of commands:
```
#!/bin/sh
touch nested_module_test.m
mmc -m nested_module_test
sleep 1
echo --------
mmc -m nested_module_test
```

exhibits this incorrect behaviour:

Making Mercury/int3s/nested_module_test.int3
Making Mercury/int0s/nested_module_test.int0
Making Mercury/ints/nested_module_test.int
Making Mercury/cs/nested_module_test.c
Making Mercury/os/nested_module_test.o
Making nested_module_test
--------
Making Mercury/int3s/nested_module_test.int3
Making Mercury/os/nested_module_test.sub.o
Making nested_module_test

where nested_module_test.sub.o and nested_module_test are re-made by the second call to mmc --make, when they should already be up-to-date.

The previous (correct) behaviour is:

Making Mercury/int3s/nested_module_test.int3
Making Mercury/int0s/nested_module_test.int0
Making Mercury/ints/nested_module_test.int
Making Mercury/cs/nested_module_test.c
Making Mercury/os/nested_module_test.o
Making Mercury/os/nested_module_test.sub.o
Making nested_module_test
--------
Making Mercury/int3s/nested_module_test.int3
** Nothing to be done for `nested_module_test'.

[Actually, nested_module_test.int3 should not need to be remade either - that is a long-standing bug, but a minor one.]

0000568: nested sub-modules broken in mmc --make

Tue, 05 Dec 2023 13:42:32 +1100

mmc --make fails to make programs making using of nested sub-modules.

The problem first appears with commit 764eab72a "Stop using module_dep_infos in computing filenames."

% mmc --make nested_module_test
** dependencies for `Mercury/int3s/nested_module_test.int3' do not exist: nested_module_test.sub.m
** This indicates a bug in `mmc --make'.

0000567: regression due to merging consecutive switches

Wed, 08 Nov 2023 17:59:38 +1100

I have bisected a regression to commit 43dd12bd5b9608dcd150ea0c7bdc1e3f256f15d2
"Merge consecutive switches on the same variable."

See the attached test case.

This is incorrect:

% mmc -O2 regression_merge_switches && ./regression_merge_switches
header(field_name("Message-ID"), header_value("Tue, 7 Nov 2023 12:29:43 +1100"))
header(field_name("Message-ID"), header_value("<20231107122943.GB219@example>"))

The correct output is this:

% mmc -O1 regression_merge_switches && ./regression_merge_switches
header(field_name("Date"), header_value("Tue, 7 Nov 2023 12:29:43 +1100"))
header(field_name("Message-ID"), header_value("<20231107122943.GB219@example>"))

0000566: undefined variable warnings with GNU Make 4.4.0.90

Mon, 16 Oct 2023 17:01:55 +1100

Since GNU Make 4.4.0.90, we get warnings about undefined variables '*' and '@':

/tmp/mmake.RTEA5X:64: warning: undefined variable '*'
/tmp/mmake.RTEA5X:64: warning: undefined variable '@'
/tmp/mmake.RTEA5X:64: warning: undefined variable '@'

The problem first appears with GNU Make commit 15dfad96d77c9445d11be939a5042675e4ca8c65 "[SV 63439, SV 63452] Don't warn on undefined internal variables"
https://git.savannah.gnu.org/cgit/make.git/commit/?id=15dfad96d77c9445d11be939a5042675e4ca8c65

The code that ultimately triggers these warnings is in Mmake.vars.in, from lines 374 onwards, that says:

# `TARGET_FLAGS' is used to pass different flags to for
# different targets. By setting MCFLAGS-foo, for example, you can add
# extra values to TARGET_MCFLAGS that will used only for compiling foo.m.
#
# For each program , `TARGET_FLAGS' expands to
# `$(FLAGS-$@)' and/or `$(FLAGS-$*)'.
# $@ and $* are the builtin variables that expand to the
# current target and (for pattern rules) the base name of the current target,
# respectively. So $* is used for flags variables used by pattern rules,
# whereas $@ is used for flags variables used by hard-coded rules,
# including the ones in the generated .dep files.
#
# The code below uses some tricky GNU Make hacks to avoid expanding
# FLAGS-$* and FLAGS-$@ if these vars are not defined,
# so that we can avoid spurious warnings if the warning about
# undefined variables is enabled.

0000208: add help and version command line options to mslice and mdice

Sun, 08 Oct 2023 12:24:18 +1100

-h and -v are pretty standard command line options, but mslice and mdice don't support them. For example:

ajrross@goofy:~$ mslice -h
unrecognized option `-h'

Since the usage information of both programs is already displayed when the user gives an incorrect number of arguments, -h could show the same information.

0000563: type constructor for equivalence type not found

Mon, 04 Sep 2023 15:05:07 +1000

The compiler throws an exception when compiling the attached test case in hlc (and probably other MLDS grades).

quirks.m defines an equivalence type in the implementation which uses a type constructor that it imports:
:- import_module set.
:- type quirks == set(string).

But when generating code for a module that imports quirks.m, that type constructor is not found.

% mmake GRADE=hlc.gc getprop.depend
% mmake GRADE=hlc.gc getprop.c
% cat getprop.err
Uncaught Mercury exception:
Software Error: map.lookup: key not found
Key Type: string
Key Value: "set"
Value Type: tree234.tree234(parse_tree.prog_data.type_ctor, hlds.hlds_data.hlds_type_defn)
Stack dump follows:
0 pred exception.throw/1-0 (erroneous) (exception.m:313)
1 pred require.error/1-0 (erroneous) (require.m:172)
2 pred require.report_lookup_error/3-0 (erroneous) (require.m:276)
3 pred map.lookup/3-0 (det) (map.m:1272)
4 pred hlds.hlds_data.lookup_type_ctor_defn/3-0 (det) (hlds_data.m:598)
5 func check_hlds.type_util.classify_type_ctor/2-0 (det) (type_util.m:1038)
6 func ml_backend.mlds.mercury_type_to_mlds_type/2-0 (det) (mlds.m:2661)
7 pred ml_backend.ml_code_util.ml_gen_local_var_decl/6-0 (det) (ml_code_util.m:998)
8 pred ml_backend.ml_code_gen.ml_gen_local_var_decls/6-0 (det) (ml_code_gen.m:1392)
9 pred ml_backend.ml_code_gen.ml_gen_goal/7-0 (det) (ml_code_gen.m:611)
10 pred ml_backend.ml_proc_gen.ml_gen_proc_body/10-0 (det) (ml_proc_gen.m:1688)
11 pred ml_backend.ml_proc_gen.ml_gen_proc/11-0 (det) (ml_proc_gen.m:620)
12 pred list.foldl3/8-0 (det) (list.m:3556)
13 pred ml_backend.ml_proc_gen.ml_gen_scc_code/10-0 (det) (ml_proc_gen.m:322)
14 pred ml_backend.ml_proc_gen.ml_gen_scc/12-0 (det) (ml_proc_gen.m:227)
15 pred ml_backend.ml_proc_gen.ml_gen_sccs/12-0 (det) (ml_proc_gen.m:211)
16 pred ml_backend.ml_proc_gen.ml_gen_preds/9-0 (det) (ml_proc_gen.m:121)
17 pred ml_backend.ml_top_gen.ml_code_gen/6-0 (det) (ml_top_gen.m:84)
18 pred top_level.mercury_compile_mlds_back_end.mlds_backend/8-0 (det) (mercury_compile_mlds_back_end.m:128)
19 pred top_level.mercury_compile_main.after_front_end_passes/13-0 (det) (mercury_compile_main.m:1930)
20 pred top_level.mercury_compile_main.process_augmented_module/14-0 (det) (mercury_compile_main.m:1756)
21 pred top_level.mercury_compile_main.augment_and_process_module/15-0 (det) (mercury_compile_main.m:1646)
22 pred list.map_foldl3/9-0 (det) (list.m:3892)
23 pred top_level.mercury_compile_main.augment_and_process_all_submodules/15-0 (det) (mercury_compile_main.m:1588)
24 pred top_level.mercury_compile_main.read_augment_and_process_module_ok/13-0 (det) (mercury_compile_main.m:1420)
25 pred top_level.mercury_compile_main.read_augment_and_process_module/11-0 (det) (mercury_compile_main.m:1362)
26 pred top_level.mercury_compile_main.do_process_compiler_arg/10-0 (det) (mercury_compile_main.m:1138)
27 pred top_level.mercury_compile_main.setup_and_process_compiler_arg/14-0 (det) (mercury_compile_main.m:1010)
28 pred top_level.mercury_compile_main.setup_and_process_compiler_cmd_line_args/16-0 (det) (mercury_compile_main.m:929)
29 pred top_level.mercury_compile_main.do_op_mode_args/13-0 (det) (mercury_compile_main.m:742)
30 pred top_level.mercury_compile_main.do_op_mode/12-0 (det) (mercury_compile_main.m:577)
31 pred top_level.mercury_compile_main.main_after_setup/7-0 (det) (mercury_compile_main.m:532)
32 pred top_level.mercury_compile_main.real_main_after_expansion/3-0 (det) (mercury_compile_main.m:276)
33 pred top_level.mercury_compile_main.real_main/2-0 (det) (mercury_compile_main.m:138)
34 pred mercury_compile.main/2-0 (det) (mercury_compile.m:32)

0000565: Installed mmake cannot build programs

Mon, 04 Sep 2023 14:07:04 +1000

Attempting to build hello word, using the mmake from an installed compiler results in:

gmake: *** No rule to make target `builtin.int', needed by `hello.c_date'. Stop.

This appears to be occurring as far back as rotd-2023-07-01 (which is the oldest compiler I have
installed on my system).

0000494: hard_coded/nonascii fails on Windows

Wed, 02 Aug 2023 11:11:03 +1000

hard_coded/nonascii fails on Windows (mingw64-{32,64},MSVC); it outputs "premature EOF" for most of the output.

0000162: No compiler error for a mode involving an instantation state that doesn't match the argument type.

Tue, 11 Oct 2022 16:43:53 +1100

In the following code I expect to receive a compiler error telling me that the instantiation state that I'm trying to use is incorrect for the type that I'm using.

If the body of the predicate is more detailed and for example uses the empty list for the head variable then the compiler complains that the empty list doesn't match the instantiation state for the some_type type.

This is reproducible on taura (Linux x86-64, Mercury version rotd-2010-08-22, gcc (Ubuntu 4.4.1-4ubuntu9) 4.4.1 with default compilation options (asm_fast.gc)

0000151: Incorrect documentation for string.unsafe_set_char

Tue, 11 Oct 2022 16:35:47 +1100

James Cussens found that the documentation for this standard library predicate and function doesn't match the implementation. The documentation says that it is a constant time operation but it involves a strcpy.

The unsafe part of the name says that there is no bounds check and that a strlen call isn't performed.