about summary refs log tree commit diff stats
path: root/apps/mulisp.subx
blob: 87479ffe224d5bd7a9c95fbe7ce3c7cbe2db0ec1 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
# Toy lisp interpreter. Incomplete.
#
# To run:
#   $ ./translate_subx init.linux 0*.subx apps/mulisp.subx
#   $ ./a.elf
#   42
#   => 42
#   ^D
#   $

== code

Entry:  # run tests if necessary, a REPL if not
    # . prologue
    89/<- %ebp 4/r32/esp
    # initialize heap
    (new-segment *Heap-size Heap)
    {
      # if (argc <= 1) break
      81 7/subop/compare *ebp 1/imm32
      7e/jump-if-lesser-or-equal break/disp8
      # if (argv[1] != "test")) break
      (kernel-string-equal? *(ebp+8) "test")  # => eax
      3d/compare-eax-and 0/imm32
      74/jump-if-equal break/disp8
      #
      (run-tests)
      # syscall(exit, *Num-test-failures)
      8b/-> *Num-test-failures 3/r32/ebx
      eb/jump $main:end/disp8
    }
    (repl Stdin Stdout)
    # syscall(exit, 0)
    bb/copy-to-ebx 0/imm32
$main:end:
    b8/copy-to-eax 1/imm32/exit
    cd/syscall 0x80/imm8

# Data structures
#
# Lisp is dynamically typed. Values always carry around knowledge of their
# type.
#
# There's several types of types in the description below, so we need a
# glossary and notational convention to disambiguate:
#   lisp type: what Lisp code can see. Looks how you type it at the prompt.
#     nil num char string symbol pair array
#   type tag: the numeric code for a lisp type. All caps.
#     NIL NUM CHAR STRING SYMBOL PAIR ARRAY
#   memory type: a type specifying memory layout at the SubX level. Starts
#   with a '$'.
#     $int $array $(addr _)
#
# Lisp values are represented in memory by the _cell_ data structure. A cell
# is 12 bytes long:
#   tag: $int (4 bytes; we're not concerned about wasting space)
#   data: 8 bytes whose contents and meaning depend on tag
#
# What values of the different Lisp types look like in memory:
#   - nil: cell{ tag: 0/NIL, data: 0 0 }
#   - num: cell{ tag: 1/NUM, data: $int 0 }
#     data contains the number
#   - char: cell{ tag: 2/CHAR, data: $int 0 }
#     data contains the utf-8 code of the character (no compound glyphs, no
#     modifiers, etc., etc.)
#   - string: cell{ tag: 3/STRING, data: $(addr stream byte)
#     data contains an (addr array byte) containing the string in utf-8
#   - symbol: cell{ tag: 4/SYMBOL, data: $(addr array byte) 0 }
#     data contains an (addr array byte) containing the name of the symbol in utf-8
#     alternatively, data could contain an index into the table of interned symbols
#   - pair: cell{ tag: 5/PAIR, data: $(addr cell) $(addr cell)  }
#     data contains pointers to car and cdr
#   - array: cell{ tag: 6/ARRAY, data: $tag $(addr stream data)
#     data contains a pointer to an array of 8-byte data fields and the common
#     tag for them all

repl:  # in : (addr buffered-file), out : (addr buffered-file)
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # . save registers
    50/push-eax
    {
      (lisp-read Stdin)  # => eax : (handle cell)
      # if (eax == 0) break
      3d/compare-eax-and 0/imm32
      74/jump-if-equal break/disp8
      #
      (lisp-eval %eax)  # => eax : (handle cell)
      (lisp-print Stdout %eax)
      eb/jump loop/disp8
    }
$repl:end:
    # . restore registers
    58/pop-to-eax
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

# numbers start with a digit and are always in hex
# characters start with a backslash
# pairs start with '('
# arrays start with '['
# symbols start with anything else but quote, backquote, unquote or splice
# only one s-expression per line
lisp-read:  # in : (addr buffered-file) -> eax : (handle cell)
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # . save registers
    51/push-ecx
    # var s/ecx : (stream byte 512)
    81 5/subop/subtract %esp 0x200/imm32
    68/push 0x200/imm32/size
    68/push 0/imm32/read
    68/push 0/imm32/write
    89/<- %ecx 4/r32/esp
    {
      # read line into s
      (clear-stream %ecx)
      (read-line-buffered *(ebp+8) %ecx)
      # if (s->write == 0) return null
      {
        81 7/subop/compare *ecx 0/imm32
        75/jump-if-not-equal break/disp8
        b8/copy-to-eax 0/imm32/eof
        eb/jump $lisp-read:end/disp8
      }
      # ...
#?       eb/jump loop/disp8
    }
    # return s
    89/<- %eax 1/r32/ecx
$lisp-read:end:
    # . reclaim locals
    81 0/subop/add %esp 0x20c/imm32
    # . restore registers
    59/pop-to-ecx
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

# lisp-read:  in : (addr buffered-file) -> (handle cell)
#   token tmp = next-mulisp-token(in)
#   if is-int(tmp) return cell(tmp)
#   if is-string(tmp) return cell(tmp)
#   if is-pair(tmp) ...
#   if is-array(tmp) ...

next-mulisp-token:  # in : (addr buffered-file), line : (addr stream byte), result : (addr slice)
    # pseudocode:
    #   if (line->read >= line->write)
    #     read-line-buffered(in, line)
    #     recurse
    #   if (line->data[line->read] == ' ')
    #     skip-chars-matching-whitespace(line)
    #     recurse
    #   if (line->data[line->read] == '#')
    #     read-line-buffered(in, line)
    #     recurse
    #   eax = line->data[line->read]
    #   if (eax == '"')
    #     result->start = &line->data[line->read]
    #     skip-string(in)
    #     result->end = &line->data[line->read]
    #     return
    #   if (is-digit(eax))
    #     result->start = &line->data[line->read]
    #     skip-hex-int(in)
    #     result->end = &line->data[line->read]
    #     return
    #   if (eax in '(' ')' '[' ']')
    #     result->start = &line->data[line->read]
    #     ++line->read
    #     result->en = &line->data[line->read]
    #     return
    #   else
    #     result->start = &line->data[line->read]
    #     skip-lisp-word(line)
    #     result->en = &line->data[line->read]
    #     return
    #
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # . save registers
$next-mulisp-token:end:
    # . reclaim locals
    # . restore registers
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

new-int-cell:  # in : (addr slice) -> eax : (handle cell)

new-string-cell:  # in : (addr slice) -> eax : (handle cell)

lisp-eval:  # in : (addr cell) -> eax : (handle cell)
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # . save registers
    8b/-> *(ebp+8) 0/r32/eax
$lisp-eval:end:
    # . restore registers
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

lisp-print:  # out : (addr buffered-file), x : (addr cell)
    # . prologue
    55/push-ebp
    89/<- %ebp 4/r32/esp
    # . save registers
    # write(x)
    (write-buffered Stdout "=> ")
    (write-stream-data Stdout *(ebp+0xc))
    (flush Stdout)
$lisp-print:end:
    # . restore registers
    # . epilogue
    89/<- %esp 5/r32/ebp
    5d/pop-to-ebp
    c3/return

== data

Nil:
  0/imm32/tag
  0/imm32/data
'n605' href='#n605'>605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039
//: Mu scenarios. This will get long, but these are the tests we want to
//: support in this layer.

//: We avoid raw numeric locations in Mu -- except in scenarios, where they're
//: handy to check the values of specific variables

void test_scenario_block() {
  run_mu_scenario(
      "scenario foo [\n"
      "  run [\n"
      "    1:num <- copy 13\n"
      "  ]\n"
      "  memory-should-contain [\n"
      "    1 <- 13\n"
      "  ]\n"
      "]\n"
  );
  // checks are inside scenario
}

void test_scenario_multiple_blocks() {
  run_mu_scenario(
      "scenario foo [\n"
      "  run [\n"
      "    1:num <- copy 13\n"
      "  ]\n"
      "  memory-should-contain [\n"
      "    1 <- 13\n"
      "  ]\n"
      "  run [\n"
      "    2:num <- copy 13\n"
      "  ]\n"
      "  memory-should-contain [\n"
      "    1 <- 13\n"
      "    2 <- 13\n"
      "  ]\n"
      "]\n"
  );
  // checks are inside scenario
}

void test_scenario_check_memory_and_trace() {
  run_mu_scenario(
      "scenario foo [\n"
      "  run [\n"
      "    1:num <- copy 13\n"
      "    trace 1, [a], [a b c]\n"
      "  ]\n"
      "  memory-should-contain [\n"
      "    1 <- 13\n"
      "  ]\n"
      "  trace-should-contain [\n"
      "    a: a b c\n"
      "  ]\n"
      "  trace-should-not-contain [\n"
      "    a: x y z\n"
      "  ]\n"
      "]\n"
  );
  // checks are inside scenario
}

//:: Core data structure

:(before "End Types")
struct scenario {
  string name;
  string to_run;
  void clear() {
    name.clear();
    to_run.clear();
  }
};

:(before "End Globals")
vector<scenario> Scenarios, Scenarios_snapshot;
set<string> Scenario_names, Scenario_names_snapshot;
:(before "End save_snapshots")
Scenarios_snapshot = Scenarios;
Scenario_names_snapshot = Scenario_names;
:(before "End restore_snapshots")
Scenarios = Scenarios_snapshot;
Scenario_names = Scenario_names_snapshot;

//:: Parse the 'scenario' form.
//: Simply store the text of the scenario.

:(before "End Command Handlers")
else if (command == "scenario") {
  scenario result = parse_scenario(in);
  if (!result.name.empty())
    Scenarios.push_back(result);
}
else if (command == "pending-scenario") {
  // for temporary use only
  parse_scenario(in);  // discard
}

:(code)
scenario parse_scenario(istream& in) {
  scenario result;
  result.name = next_word(in);
  if (contains_key(Scenario_names, result.name))
    raise << "duplicate scenario name: '" << result.name << "'\n" << end();
  Scenario_names.insert(result.name);
  if (result.name.empty()) {
    assert(!has_data(in));
    raise << "incomplete scenario at end of file\n" << end();
    return result;
  }
  skip_whitespace_and_comments(in);
  if (in.peek() != '[') {
    raise << "Expected '[' after scenario '" << result.name << "'\n" << end();
    exit(0);
  }
  // scenarios are take special 'code' strings so we need to ignore brackets
  // inside comments
  result.to_run = slurp_quoted(in);
  // delete [] delimiters
  if (!starts_with(result.to_run, "[")) {
    raise << "scenario " << result.name << " should start with '['\n" << end();
    result.clear();
    return result;
  }
  result.to_run.erase(0, 1);
  if (result.to_run.at(SIZE(result.to_run)-1) != ']') {
    raise << "scenario " << result.name << " has an unbalanced '['\n" << end();
    result.clear();
    return result;
  }
  result.to_run.erase(SIZE(result.to_run)-1);
  return result;
}

void test_read_scenario_with_bracket_in_comment() {
  run_mu_scenario(
      "scenario foo [\n"
      "  # ']' in comment\n"
      "  1:num <- copy 0\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "run: {1: \"number\"} <- copy {0: \"literal\"}\n"
  );
}

void test_read_scenario_with_bracket_in_comment_in_nested_string() {
  run_mu_scenario(
      "scenario foo [\n"
      "  1:text <- new [# not a comment]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "run: {1: (\"address\" \"array\" \"character\")} <- new {\"# not a comment\": \"literal-string\"}\n"
  );
}

void test_duplicate_scenarios() {
  Hide_errors = true;
  run(
      "scenario foo [\n"
      "  1:num <- copy 0\n"
      "]\n"
      "scenario foo [\n"
      "  2:num <- copy 0\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: duplicate scenario name: 'foo'\n"
  );
}

//:: Run scenarios when we run './mu test'.
//: Treat the text of the scenario as a regular series of instructions.

:(before "End Globals")
int Num_core_mu_scenarios = 0;
:(after "Check For .mu Files")
Num_core_mu_scenarios = SIZE(Scenarios);
:(before "End Tests")
Hide_missing_default_space_errors = false;
if (Num_core_mu_scenarios > 0) {
  time(&t);
  cerr << "Mu tests: " << ctime(&t);
  for (int i = 0;  i < Num_core_mu_scenarios;  ++i) {
//?     cerr << '\n' << i << ": " << Scenarios.at(i).name;
    run_mu_scenario(Scenarios.at(i));
    if (Passed) cerr << ".";
    else ++num_failures;
  }
  cerr << "\n";
}
run_app_scenarios:
if (Num_core_mu_scenarios != SIZE(Scenarios)) {
  time(&t);
  cerr << "App tests: " << ctime(&t);
  for (int i = Num_core_mu_scenarios;  i < SIZE(Scenarios);  ++i) {
//?     cerr << '\n' << i << ": " << Scenarios.at(i).name;
    run_mu_scenario(Scenarios.at(i));
    if (Passed) cerr << ".";
    else ++num_failures;
  }
  cerr << "\n";
}

//: For faster debugging, support running tests for just the Mu app(s) we are
//: loading.
:(before "End Globals")
bool Test_only_app = false;
:(before "End Commandline Options(*arg)")
else if (is_equal(*arg, "--test-only-app")) {
  Test_only_app = true;
}
:(after "End Test Run Initialization")
if (Test_only_app && Num_core_mu_scenarios < SIZE(Scenarios)) {
  goto run_app_scenarios;
}

//: Convenience: run a single named scenario.
:(after "Test Runs")
for (int i = 0;  i < SIZE(Scenarios);  ++i) {
  if (Scenarios.at(i).name == argv[argc-1]) {
    run_mu_scenario(Scenarios.at(i));
    if (Passed) cerr << ".\n";
    return 0;
  }
}

:(before "End Globals")
// this isn't a constant, just a global of type const*
const scenario* Current_scenario = NULL;
:(code)
void run_mu_scenario(const scenario& s) {
  Current_scenario = &s;
  bool not_already_inside_test = !Trace_stream;
//?   cerr << s.name << '\n';
  if (not_already_inside_test) {
    Trace_stream = new trace_stream;
    reset();
  }
  vector<recipe_ordinal> tmp = load("recipe scenario_"+s.name+" [ "+s.to_run+" ]");
  mark_autogenerated(tmp.at(0));
  bind_special_scenario_names(tmp.at(0));
  transform_all();
  if (!trace_contains_errors())
    run(tmp.front());
  // End Mu Test Teardown
  if (!Hide_errors && trace_contains_errors() && !Scenario_testing_scenario)
    Passed = false;
  if (not_already_inside_test && Trace_stream) {
    delete Trace_stream;
    Trace_stream = NULL;
  }
  Current_scenario = NULL;
}

//: Permit numeric locations to be accessed in scenarios.
:(before "End check_default_space Special-cases")
// user code should never create recipes with underscores in their names
if (starts_with(caller.name, "scenario_")) return;  // skip Mu scenarios which will use raw memory locations
if (starts_with(caller.name, "run_")) return;  // skip calls to 'run', which should be in scenarios and will also use raw memory locations

:(before "End maybe(recipe_name) Special-cases")
if (starts_with(recipe_name, "scenario_"))
  return recipe_name.substr(strlen("scenario_")) + ": ";

//: Some variables for fake resources always get special /raw addresses in scenarios.

:(code)
// Should contain everything passed by is_special_name but failed by is_disqualified.
void bind_special_scenario_names(const recipe_ordinal r) {
  // Special Scenario Variable Names(r)
  // End Special Scenario Variable Names(r)
}
:(before "Done Placing Ingredient(ingredient, inst, caller)")
maybe_make_raw(ingredient, caller);
:(before "Done Placing Product(product, inst, caller)")
maybe_make_raw(product, caller);
:(code)
void maybe_make_raw(reagent& r, const recipe& caller) {
  if (!is_special_name(r.name)) return;
  if (starts_with(caller.name, "scenario_"))
    r.properties.push_back(pair<string, string_tree*>("raw", NULL));
  // End maybe_make_raw
}

//: Test with some setup.
:(before "End is_special_name Special-cases")
if (s == "__maybe_make_raw_test__") return true;
:(before "End Special Scenario Variable Names(r)")
//: ugly: we only need this for this one test, but need to define it for all time
Name[r]["__maybe_make_raw_test__"] = Reserved_for_tests-1;
:(code)
void test_maybe_make_raw() {
  // check that scenarios can use local-scope and special variables together
  vector<recipe_ordinal> tmp = load(
      "def scenario_foo [\n"
      "  local-scope\n"
      "  __maybe_make_raw_test__:num <- copy 34\n"
      "]\n");
  mark_autogenerated(tmp.at(0));
  bind_special_scenario_names(tmp.at(0));
  transform_all();
  run(tmp.at(0));
  CHECK_TRACE_DOESNT_CONTAIN_ERRORS();
}

//: Watch out for redefinitions of scenario routines. We should never ever be
//: doing that, regardless of anything else.

void test_forbid_redefining_scenario_even_if_forced() {
  Hide_errors = true;
  Disable_redefine_checks = true;
  run(
      "def scenario-foo [\n"
      "  1:num <- copy 34\n"
      "]\n"
      "def scenario-foo [\n"
      "  1:num <- copy 35\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: redefining recipe scenario-foo\n"
  );
}

void test_scenario_containing_parse_error() {
  Hide_errors = true;
  run(
      "scenario foo [\n"
      "  memory-should-contain [\n"
      "    1 <- 0\n"
         // missing ']'
      "]\n"
  );
  // no crash
}

void test_scenario_containing_transform_error() {
  Hide_errors = true;
  run(
      "def main [\n"
      "  local-scope\n"
      "  add x, 1\n"
      "]\n"
  );
  // no crash
}

:(after "bool should_check_for_redefine(const string& recipe_name)")
  if (recipe_name.find("scenario-") == 0) return true;

//:: The special instructions we want to support inside scenarios.
//: These are easy to support in an interpreter, but will require more work
//: when we eventually build a compiler.

//: 'run' is a purely lexical convenience to separate the code actually being
//: tested from any setup

:(code)
void test_run() {
  run(
      "def main [\n"
      "  run [\n"
      "    1:num <- copy 13\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "mem: storing 13 in location 1\n"
  );
}

:(before "End Rewrite Instruction(curr, recipe result)")
if (curr.name == "run") {
  // Just inline all instructions inside the run block in the containing
  // recipe. 'run' is basically a comment; pretend it doesn't exist.
  istringstream in2("[\n"+curr.ingredients.at(0).name+"\n]\n");
  slurp_body(in2, result);
  curr.clear();
}

:(code)
void test_run_multiple() {
  run(
      "def main [\n"
      "  run [\n"
      "    1:num <- copy 13\n"
      "  ]\n"
      "  run [\n"
      "    2:num <- copy 13\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "mem: storing 13 in location 1\n"
      "mem: storing 13 in location 2\n"
  );
}

//: 'memory-should-contain' raises errors if specific locations aren't as expected
//: Also includes some special support for checking Mu texts.

:(before "End Globals")
bool Scenario_testing_scenario = false;
:(before "End Reset")
Scenario_testing_scenario = false;

:(code)
void test_memory_check() {
  Scenario_testing_scenario = true;
  Hide_errors = true;
  run(
      "def main [\n"
      "  memory-should-contain [\n"
      "    1 <- 13\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "run: checking location 1\n"
      "error: F - main: expected location '1' to contain 13 but saw 0\n"
  );
}

:(before "End Primitive Recipe Declarations")
MEMORY_SHOULD_CONTAIN,
:(before "End Primitive Recipe Numbers")
put(Recipe_ordinal, "memory-should-contain", MEMORY_SHOULD_CONTAIN);
:(before "End Primitive Recipe Checks")
case MEMORY_SHOULD_CONTAIN: {
  break;
}
:(before "End Primitive Recipe Implementations")
case MEMORY_SHOULD_CONTAIN: {
  if (!Passed) break;
  check_memory(current_instruction().ingredients.at(0).name);
  break;
}

:(code)
void check_memory(const string& s) {
  istringstream in(s);
  in >> std::noskipws;
  set<int> locations_checked;
  while (true) {
    skip_whitespace_and_comments(in);
    if (!has_data(in)) break;
    string lhs = next_word(in);
    if (lhs.empty()) {
      assert(!has_data(in));
      raise << maybe(current_recipe_name()) << "incomplete 'memory-should-contain' block at end of file (0)\n" << end();
      return;
    }
    if (!is_integer(lhs)) {
      check_type(lhs, in);
      continue;
    }
    int address = to_integer(lhs);
    skip_whitespace_and_comments(in);
    string _assign;  in >> _assign;  assert(_assign == "<-");
    skip_whitespace_and_comments(in);
    string rhs = next_word(in);
    if (rhs.empty()) {
      assert(!has_data(in));
      raise << maybe(current_recipe_name()) << "incomplete 'memory-should-contain' block at end of file (1)\n" << end();
      return;
    }
    if (!is_integer(rhs) && !is_noninteger(rhs)) {
      if (!Hide_errors) cerr << '\n';
      raise << "F - " << maybe(current_recipe_name()) << "location '" << address << "' can't contain non-number " << rhs << '\n' << end();
      if (!Scenario_testing_scenario) Passed = false;
      return;
    }
    double value = to_double(rhs);
    if (contains_key(locations_checked, address))
      raise << maybe(current_recipe_name()) << "duplicate expectation for location '" << address << "'\n" << end();
    trace(Callstack_depth+1, "run") << "checking location " << address << end();
    if (get_or_insert(Memory, address) != value) {
      if (!Hide_errors) cerr << '\n';
      raise << "F - " << maybe(current_recipe_name()) << "expected location '" << address << "' to contain " << no_scientific(value) << " but saw " << no_scientific(get_or_insert(Memory, address)) << '\n' << end();
      if (!Scenario_testing_scenario) Passed = false;
      return;
    }
    locations_checked.insert(address);
  }
}

void check_type(const string& lhs, istream& in) {
  reagent x(lhs);
  if (is_mu_array(x.type) && is_mu_character(array_element(x.type))) {
    x.set_value(to_integer(x.name));
    skip_whitespace_and_comments(in);
    string _assign = next_word(in);
    if (_assign.empty()) {
      assert(!has_data(in));
      raise << maybe(current_recipe_name()) << "incomplete 'memory-should-contain' block at end of file (2)\n" << end();
      return;
    }
    assert(_assign == "<-");
    skip_whitespace_and_comments(in);
    string literal = next_word(in);
    if (literal.empty()) {
      assert(!has_data(in));
      raise << maybe(current_recipe_name()) << "incomplete 'memory-should-contain' block at end of file (3)\n" << end();
      return;
    }
    int address = x.value;
    // exclude quoting brackets
    if (*literal.begin() != '[') {
      raise << maybe(current_recipe_name()) << "array:character types inside 'memory-should-contain' can only be compared with text literals surrounded by [], not '" << literal << "'\n" << end();
      return;
    }
    literal.erase(literal.begin());
    assert(*--literal.end() == ']');  literal.erase(--literal.end());
    check_mu_text(address, literal);
    return;
  }
  // End Scenario Type Special-cases
  raise << "don't know how to check memory for '" << lhs << "'\n" << end();
}

void check_mu_text(int start, const string& literal) {
  trace(Callstack_depth+1, "run") << "checking text length at " << start << end();
  int array_length = static_cast<int>(get_or_insert(Memory, start));
  if (array_length != SIZE(literal)) {
    if (!Hide_errors) cerr << '\n';
    raise << "F - " << maybe(current_recipe_name()) << "expected location '" << start << "' to contain length " << SIZE(literal) << " of text [" << literal << "] but saw " << array_length << " (for text [" << read_mu_characters(start+/*skip length*/1, array_length) << "])\n" << end();
    if (!Scenario_testing_scenario) Passed = false;
    return;
  }
  int curr = start+1;  // now skip length
  for (int i = 0;  i < SIZE(literal);  ++i) {
    trace(Callstack_depth+1, "run") << "checking location " << curr+i << end();
    if (get_or_insert(Memory, curr+i) != literal.at(i)) {
      if (!Hide_errors) cerr << '\n';
      raise << "F - " << maybe(current_recipe_name()) << "expected location " << (curr+i) << " to contain " << literal.at(i) << " but saw " << no_scientific(get_or_insert(Memory, curr+i)) << '\n' << end();
      if (!Scenario_testing_scenario) Passed = false;
      return;
    }
  }
}

void test_memory_check_multiple() {
  Scenario_testing_scenario = true;
  Hide_errors = true;
  run(
      "def main [\n"
      "  memory-should-contain [\n"
      "    1 <- 0\n"
      "    1 <- 0\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: main: duplicate expectation for location '1'\n"
  );
}

void test_memory_check_mu_text_length() {
  Scenario_testing_scenario = true;
  Hide_errors = true;
  run(
      "def main [\n"
      "  1:num <- copy 3\n"
      "  2:num <- copy 97  # 'a'\n"
      "  3:num <- copy 98  # 'b'\n"
      "  4:num <- copy 99  # 'c'\n"
      "  memory-should-contain [\n"
      "    1:array:character <- [ab]\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: F - main: expected location '1' to contain length 2 of text [ab] but saw 3 (for text [abc])\n"
  );
}

void test_memory_check_mu_text() {
  run(
      "def main [\n"
      "  1:num <- copy 3\n"
      "  2:num <- copy 97\n"  // 'a'
      "  3:num <- copy 98\n"  // 'b'
      "  4:num <- copy 99\n"  // 'c'
      "  memory-should-contain [\n"
      "    1:array:character <- [abc]\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "run: checking text length at 1\n"
      "run: checking location 2\n"
      "run: checking location 3\n"
      "run: checking location 4\n"
  );
}

void test_memory_invalid_string_check() {
  Scenario_testing_scenario = true;
  Hide_errors = true;
  run(
      "def main [\n"
      "  memory-should-contain [\n"
      "    1 <- [abc]\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: F - main: location '1' can't contain non-number [abc]\n"
  );
}

void test_memory_invalid_string_check2() {
  Hide_errors = true;
  run(
      "def main [\n"
      "  1:num <- copy 3\n"
      "  2:num <- copy 97\n"  // 'a'
      "  3:num <- copy 98\n"  // 'b'
      "  4:num <- copy 99\n"  // 'c'
      "  memory-should-contain [\n"
      "    1:array:character <- 0\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: main: array:character types inside 'memory-should-contain' can only be compared with text literals surrounded by [], not '0'\n"
  );
}

void test_memory_check_with_comment() {
  Scenario_testing_scenario = true;
  Hide_errors = true;
  run(
      "def main [\n"
      "  memory-should-contain [\n"
      "    1 <- 34  # comment\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_DOESNT_CONTAIN("error: location 1 can't contain non-number 34  # comment");
  // but there'll be an error signalled by memory-should-contain
}

//: 'trace-should-contain' is like the '+' lines in our scenarios so far
//: Like runs of contiguous '+' lines, order is important. The trace checks
//: that the lines are present *and* in the specified sequence. (There can be
//: other lines in between.)

void test_trace_check_fails() {
  Scenario_testing_scenario = true;
  Hide_errors = true;
  run(
      "def main [\n"
         // nothing added to the trace
      "  trace-should-contain [\n"
      "    a: b\n"
      "    a: d\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: F - main: missing [b] in trace with label 'a'\n"
  );
}

:(before "End Primitive Recipe Declarations")
TRACE_SHOULD_CONTAIN,
:(before "End Primitive Recipe Numbers")
put(Recipe_ordinal, "trace-should-contain", TRACE_SHOULD_CONTAIN);
:(before "End Primitive Recipe Checks")
case TRACE_SHOULD_CONTAIN: {
  break;
}
:(before "End Primitive Recipe Implementations")
case TRACE_SHOULD_CONTAIN: {
  if (!Passed) break;
  check_trace(current_instruction().ingredients.at(0).name);
  break;
}

:(code)
// simplified version of check_trace_contents() that emits errors rather
// than just printing to stderr
void check_trace(const string& expected) {
  Trace_stream->newline();
  vector<trace_line> expected_lines = parse_trace(expected);
  if (expected_lines.empty()) return;
  int curr_expected_line = 0;
  for (vector<trace_line>::iterator p = Trace_stream->past_lines.begin();  p != Trace_stream->past_lines.end();  ++p) {
    if (expected_lines.at(curr_expected_line).label != p->label) continue;
    if (expected_lines.at(curr_expected_line).contents != trim(p->contents)) continue;
    // match
    ++curr_expected_line;
    if (curr_expected_line == SIZE(expected_lines)) return;
  }
  if (!Hide_errors) cerr << '\n';
  raise << "F - " << maybe(current_recipe_name()) << "missing [" << expected_lines.at(curr_expected_line).contents << "] "
        << "in trace with label '" << expected_lines.at(curr_expected_line).label << "'\n" << end();
  if (!Hide_errors)
    DUMP(expected_lines.at(curr_expected_line).label);
  if (!Scenario_testing_scenario) Passed = false;
}

vector<trace_line> parse_trace(const string& expected) {
  vector<string> buf = split(expected, "\n");
  vector<trace_line> result;
  const string separator = ": ";
  for (int i = 0;  i < SIZE(buf);  ++i) {
    buf.at(i) = trim(buf.at(i));
    if (buf.at(i).empty()) continue;
    int delim = buf.at(i).find(separator);
    if (delim == -1) {
      raise << maybe(current_recipe_name()) << "lines in 'trace-should-contain' should be of the form <label>: <contents>. Both parts are required.\n" << end();
      result.clear();
      return result;
    }
    result.push_back(trace_line(/*contents*/  trim(buf.at(i).substr(delim+SIZE(separator)        )),
                                /*label*/     trim(buf.at(i).substr(0,                      delim))));
  }
  return result;
}

void test_trace_check_fails_in_nonfirst_line() {
  Scenario_testing_scenario = true;
  Hide_errors = true;
  run(
      "def main [\n"
      "  run [\n"
      "    trace 1, [a], [b]\n"
      "  ]\n"
      "  trace-should-contain [\n"
      "    a: b\n"
      "    a: d\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: F - main: missing [d] in trace with label 'a'\n"
  );
}

void test_trace_check_passes_silently() {
  Scenario_testing_scenario = true;
  run(
      "def main [\n"
      "  run [\n"
      "    trace 1, [a], [b]\n"
      "  ]\n"
      "  trace-should-contain [\n"
      "    a: b\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_DOESNT_CONTAIN("error: missing [b] in trace with label 'a'");
  CHECK_TRACE_COUNT("error", 0);
}

//: 'trace-should-not-contain' checks each line in its argument for absence.
//: Order is *not* important, so you can't say things like "B should not exist
//: after A."

void test_trace_negative_check_fails() {
  Scenario_testing_scenario = true;
  Hide_errors = true;
  run(
      "def main [\n"
      "  run [\n"
      "    trace 1, [a], [b]\n"
      "  ]\n"
      "  trace-should-not-contain [\n"
      "    a: b\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: F - main: unexpected [b] in trace with label 'a'\n"
  );
}

:(before "End Primitive Recipe Declarations")
TRACE_SHOULD_NOT_CONTAIN,
:(before "End Primitive Recipe Numbers")
put(Recipe_ordinal, "trace-should-not-contain", TRACE_SHOULD_NOT_CONTAIN);
:(before "End Primitive Recipe Checks")
case TRACE_SHOULD_NOT_CONTAIN: {
  break;
}
:(before "End Primitive Recipe Implementations")
case TRACE_SHOULD_NOT_CONTAIN: {
  if (!Passed) break;
  check_trace_missing(current_instruction().ingredients.at(0).name);
  break;
}

:(code)
// simplified version of check_trace_contents() that emits errors rather
// than just printing to stderr
bool check_trace_missing(const string& in) {
  Trace_stream->newline();
  vector<trace_line> lines = parse_trace(in);
  for (int i = 0;  i < SIZE(lines);  ++i) {
    if (trace_count(lines.at(i).label, lines.at(i).contents) != 0) {
      raise << "F - " << maybe(current_recipe_name()) << "unexpected [" << lines.at(i).contents << "] in trace with label '" << lines.at(i).label << "'\n" << end();
      if (!Scenario_testing_scenario) Passed = false;
      return false;
    }
  }
  return true;
}

void test_trace_negative_check_passes_silently() {
  Scenario_testing_scenario = true;
  run(
      "def main [\n"
      "  trace-should-not-contain [\n"
      "    a: b\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_DOESNT_CONTAIN("error: unexpected [b] in trace with label 'a'");
  CHECK_TRACE_COUNT("error", 0);
}

void test_trace_negative_check_fails_on_any_unexpected_line() {
  Scenario_testing_scenario = true;
  Hide_errors = true;
  run(
      "def main [\n"
      "  run [\n"
      "    trace 1, [a], [d]\n"
      "  ]\n"
      "  trace-should-not-contain [\n"
      "    a: b\n"
      "    a: d\n"
      "  ]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: F - main: unexpected [d] in trace with label 'a'\n"
  );
}

void test_trace_count_check() {
  run(
      "def main [\n"
      "  run [\n"
      "    trace 1, [a], [foo]\n"
      "  ]\n"
      "  check-trace-count-for-label 1, [a]\n"
      "]\n"
  );
  // checks are inside scenario
}

:(before "End Primitive Recipe Declarations")
CHECK_TRACE_COUNT_FOR_LABEL,
:(before "End Primitive Recipe Numbers")
put(Recipe_ordinal, "check-trace-count-for-label", CHECK_TRACE_COUNT_FOR_LABEL);
:(before "End Primitive Recipe Checks")
case CHECK_TRACE_COUNT_FOR_LABEL: {
  if (SIZE(inst.ingredients) != 2) {
    raise << maybe(get(Recipe, r).name) << "'check-trace-count-for-label' requires exactly two ingredients, but got '" << to_original_string(inst) << "'\n" << end();
    break;
  }
  if (!is_mu_number(inst.ingredients.at(0))) {
    raise << maybe(get(Recipe, r).name) << "first ingredient of 'check-trace-count-for-label' should be a number (count), but got '" << inst.ingredients.at(0).original_string << "'\n" << end();
    break;
  }
  if (!is_literal_text(inst.ingredients.at(1))) {
    raise << maybe(get(Recipe, r).name) << "second ingredient of 'check-trace-count-for-label' should be a literal string (label), but got '" << inst.ingredients.at(1).original_string << "'\n" << end();
    break;
  }
  break;
}
:(before "End Primitive Recipe Implementations")
case CHECK_TRACE_COUNT_FOR_LABEL: {
  if (!Passed) break;
  int expected_count = ingredients.at(0).at(0);
  string label = current_instruction().ingredients.at(1).name;
  int count = trace_count(label);
  if (count != expected_count) {
    if (!Hide_errors) cerr << '\n';
    raise << "F - " << maybe(current_recipe_name()) << "expected " << expected_count << " lines in trace with label '" << label << "' in trace\n" << end();
    if (!Hide_errors) DUMP(label);
    if (!Scenario_testing_scenario) Passed = false;
  }
  break;
}

:(before "End Primitive Recipe Declarations")
CHECK_TRACE_COUNT_FOR_LABEL_GREATER_THAN,
:(before "End Primitive Recipe Numbers")
put(Recipe_ordinal, "check-trace-count-for-label-greater-than", CHECK_TRACE_COUNT_FOR_LABEL_GREATER_THAN);
:(before "End Primitive Recipe Checks")
case CHECK_TRACE_COUNT_FOR_LABEL_GREATER_THAN: {
  if (SIZE(inst.ingredients) != 2) {
    raise << maybe(get(Recipe, r).name) << "'check-trace-count-for-label' requires exactly two ingredients, but got '" << to_original_string(inst) << "'\n" << end();
    break;
  }
  if (!is_mu_number(inst.ingredients.at(0))) {
    raise << maybe(get(Recipe, r).name) << "first ingredient of 'check-trace-count-for-label' should be a number (count), but got '" << inst.ingredients.at(0).original_string << "'\n" << end();
    break;
  }
  if (!is_literal_text(inst.ingredients.at(1))) {
    raise << maybe(get(Recipe, r).name) << "second ingredient of 'check-trace-count-for-label' should be a literal string (label), but got '" << inst.ingredients.at(1).original_string << "'\n" << end();
    break;
  }
  break;
}
:(before "End Primitive Recipe Implementations")
case CHECK_TRACE_COUNT_FOR_LABEL_GREATER_THAN: {
  if (!Passed) break;
  int expected_count = ingredients.at(0).at(0);
  string label = current_instruction().ingredients.at(1).name;
  int count = trace_count(label);
  if (count <= expected_count) {
    if (!Hide_errors) cerr << '\n';
    raise << maybe(current_recipe_name()) << "expected more than " << expected_count << " lines in trace with label '" << label << "' in trace\n" << end();
    if (!Hide_errors) {
      cerr << "trace contents:\n";
      DUMP(label);
    }
    if (!Scenario_testing_scenario) Passed = false;
  }
  break;
}

:(before "End Primitive Recipe Declarations")
CHECK_TRACE_COUNT_FOR_LABEL_LESSER_THAN,
:(before "End Primitive Recipe Numbers")
put(Recipe_ordinal, "check-trace-count-for-label-lesser-than", CHECK_TRACE_COUNT_FOR_LABEL_LESSER_THAN);
:(before "End Primitive Recipe Checks")
case CHECK_TRACE_COUNT_FOR_LABEL_LESSER_THAN: {
  if (SIZE(inst.ingredients) != 2) {
    raise << maybe(get(Recipe, r).name) << "'check-trace-count-for-label' requires exactly two ingredients, but got '" << to_original_string(inst) << "'\n" << end();
    break;
  }
  if (!is_mu_number(inst.ingredients.at(0))) {
    raise << maybe(get(Recipe, r).name) << "first ingredient of 'check-trace-count-for-label' should be a number (count), but got '" << inst.ingredients.at(0).original_string << "'\n" << end();
    break;
  }
  if (!is_literal_text(inst.ingredients.at(1))) {
    raise << maybe(get(Recipe, r).name) << "second ingredient of 'check-trace-count-for-label' should be a literal string (label), but got '" << inst.ingredients.at(1).original_string << "'\n" << end();
    break;
  }
  break;
}
:(before "End Primitive Recipe Implementations")
case CHECK_TRACE_COUNT_FOR_LABEL_LESSER_THAN: {
  if (!Passed) break;
  int expected_count = ingredients.at(0).at(0);
  string label = current_instruction().ingredients.at(1).name;
  int count = trace_count(label);
  if (count >= expected_count) {
    if (!Hide_errors) cerr << '\n';
    raise << "F - " << maybe(current_recipe_name()) << "expected less than " << expected_count << " lines in trace with label '" << label << "' in trace\n" << end();
    if (!Hide_errors) {
      cerr << "trace contents:\n";
      DUMP(label);
    }
    if (!Scenario_testing_scenario) Passed = false;
  }
  break;
}

:(code)
void test_trace_count_check_2() {
  Scenario_testing_scenario = true;
  Hide_errors = true;
  run(
      "def main [\n"
      "  run [\n"
      "    trace 1, [a], [foo]\n"
      "  ]\n"
      "  check-trace-count-for-label 2, [a]\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: F - main: expected 2 lines in trace with label 'a' in trace\n"
  );
}

//: Minor detail: ignore 'system' calls in scenarios, since anything we do
//: with them is by definition impossible to test through Mu.
:(after "case _SYSTEM:")
  if (Current_scenario) break;

//:: Warn if people use '_' manually in recipe names. They're reserved for internal use.

:(code)
void test_recipe_name_with_underscore() {
  Hide_errors = true;
  run(
      "def foo_bar [\n"
      "]\n"
  );
  CHECK_TRACE_CONTENTS(
      "error: foo_bar: don't create recipes with '_' in the name\n"
  );
}

:(before "End recipe Fields")
bool is_autogenerated;
:(before "End recipe Constructor")
is_autogenerated = false;
:(code)
void mark_autogenerated(recipe_ordinal r) {
  get(Recipe, r).is_autogenerated = true;
}

:(after "void transform_all()")
  for (map<recipe_ordinal, recipe>::iterator p = Recipe.begin();  p != Recipe.end();  ++p) {
    const recipe& r = p->second;
    if (r.name.find('_') == string::npos) continue;
    if (r.is_autogenerated) continue;  // created by previous call to transform_all()
    raise << r.name << ": don't create recipes with '_' in the name\n" << end();
  }

//:: Helpers

:(code)
// just for the scenarios running scenarios in C++ layers
void run_mu_scenario(const string& form) {
  Scenario_names.clear();
  istringstream in(form);
  in >> std::noskipws;
  skip_whitespace_and_comments(in);
  string _scenario = next_word(in);
  if (_scenario.empty()) {
    assert(!has_data(in));
    raise << "no scenario in string passed into run_mu_scenario()\n" << end();
    return;
  }
  assert(_scenario == "scenario");
  scenario s = parse_scenario(in);
  run_mu_scenario(s);
}