mu/archive/1.vm/041jump_target.cc

221 lines
6.2 KiB
C++
Raw Normal View History

//: Support jumps to a specific label (the 'jump target') in the same recipe.
//: Jump targets must be unique and unambiguous within any recipe.
//:
//: The 'break' and 'loop' pseudo instructions can also take jump targets.
//: Which instruction you use is just documentation about intent; use 'break'
//: to indicate you're exiting one or more loop nests, and 'loop' to indicate
//: you're skipping to the next iteration of some containing loop nest.
//: Since they have to be unique in a recipe, not all labels can be jump
//: targets.
bool is_jump_target(const string& label) {
if (label == "{" || label == "}") return false;
// End is_jump_target Special-cases
return is_label_word(label);
}
5001 - drop the :(scenario) DSL I've been saying for a while[1][2][3] that adding extra abstractions makes things harder for newcomers, and adding new notations doubly so. And then I notice this DSL in my own backyard. Makes me feel like a hypocrite. [1] https://news.ycombinator.com/item?id=13565743#13570092 [2] https://lobste.rs/s/to8wpr/configuration_files_are_canary_warning [3] https://lobste.rs/s/mdmcdi/little_languages_by_jon_bentley_1986#c_3miuf2 The implementation of the DSL was also highly hacky: a) It was happening in the tangle/ tool, but was utterly unrelated to tangling layers. b) There were several persnickety constraints on the different kinds of lines and the specific order they were expected in. I kept finding bugs where the translator would silently do the wrong thing. Or the error messages sucked, and readers may be stuck looking at the generated code to figure out what happened. Fixing error messages would require a lot more code, which is one of my arguments against DSLs in the first place: they may be easy to implement, but they're hard to design to go with the grain of the underlying platform. They require lots of iteration. Is that effort worth prioritizing in this project? On the other hand, the DSL did make at least some readers' life easier, the ones who weren't immediately put off by having to learn a strange syntax. There were fewer quotes to parse, fewer backslash escapes. Anyway, since there are also people who dislike having to put up with strange syntaxes, we'll call that consideration a wash and tear this DSL out. --- This commit was sheer drudgery. Hopefully it won't need to be redone with a new DSL because I grow sick of backslashes.
2019-03-13 01:56:55 +00:00
void test_jump_to_label() {
run(
"def main [\n"
" jump +target:label\n"
" 1:num <- copy 0\n"
" +target\n"
"]\n"
);
CHECK_TRACE_DOESNT_CONTAIN("mem: storing 0 in location 1");
}
:(before "End Mu Types Initialization")
put(Type_ordinal, "label", 0);
:(before "End Instruction Modifying Transforms")
Transform.push_back(transform_labels); // idempotent
:(code)
void transform_labels(const recipe_ordinal r) {
map<string, int> offset;
2016-10-20 05:10:35 +00:00
for (int i = 0; i < SIZE(get(Recipe, r).steps); ++i) {
const instruction& inst = get(Recipe, r).steps.at(i);
if (!inst.is_label) continue;
if (is_jump_target(inst.label)) {
if (!contains_key(offset, inst.label)) {
put(offset, inst.label, i);
}
else {
2016-02-26 21:04:55 +00:00
raise << maybe(get(Recipe, r).name) << "duplicate label '" << inst.label << "'" << end();
// have all jumps skip some random but noticeable and deterministic amount of code
put(offset, inst.label, 9999);
}
}
}
2016-10-20 05:10:35 +00:00
for (int i = 0; i < SIZE(get(Recipe, r).steps); ++i) {
instruction& inst = get(Recipe, r).steps.at(i);
if (inst.name == "jump") {
2016-02-28 18:37:37 +00:00
if (inst.ingredients.empty()) {
2017-05-26 23:43:18 +00:00
raise << maybe(get(Recipe, r).name) << "'" << to_original_string(inst) << "' expects an ingredient but got 0\n" << end();
2016-02-28 18:37:37 +00:00
return;
}
replace_offset(inst.ingredients.at(0), offset, i, r);
}
if (inst.name == "jump-if" || inst.name == "jump-unless") {
2016-02-28 18:37:37 +00:00
if (SIZE(inst.ingredients) < 2) {
2017-05-26 23:43:18 +00:00
raise << maybe(get(Recipe, r).name) << "'" << to_original_string(inst) << "' expects 2 ingredients but got " << SIZE(inst.ingredients) << '\n' << end();
2016-02-28 18:37:37 +00:00
return;
}
replace_offset(inst.ingredients.at(1), offset, i, r);
}
if ((inst.name == "loop" || inst.name == "break")
2016-02-28 18:37:37 +00:00
&& SIZE(inst.ingredients) >= 1) {
replace_offset(inst.ingredients.at(0), offset, i, r);
}
if ((inst.name == "loop-if" || inst.name == "loop-unless"
|| inst.name == "break-if" || inst.name == "break-unless")
2016-02-28 18:37:37 +00:00
&& SIZE(inst.ingredients) >= 2) {
replace_offset(inst.ingredients.at(1), offset, i, r);
}
}
}
void replace_offset(reagent& x, /*const*/ map<string, int>& offset, const int current_offset, const recipe_ordinal r) {
if (!is_literal(x)) {
raise << maybe(get(Recipe, r).name) << "jump target must be offset or label but is '" << x.original_string << "'\n" << end();
x.set_value(0); // no jump by default
return;
}
if (x.initialized) return;
2015-05-17 04:24:21 +00:00
if (is_integer(x.name)) return; // non-labels will be handled like other number operands
if (!is_jump_target(x.name)) {
raise << maybe(get(Recipe, r).name) << "can't jump to label '" << x.name << "'\n" << end();
x.set_value(0); // no jump by default
return;
}
if (!contains_key(offset, x.name)) {
raise << maybe(get(Recipe, r).name) << "can't find label '" << x.name << "'\n" << end();
x.set_value(0); // no jump by default
return;
}
x.set_value(get(offset, x.name) - current_offset);
}
5001 - drop the :(scenario) DSL I've been saying for a while[1][2][3] that adding extra abstractions makes things harder for newcomers, and adding new notations doubly so. And then I notice this DSL in my own backyard. Makes me feel like a hypocrite. [1] https://news.ycombinator.com/item?id=13565743#13570092 [2] https://lobste.rs/s/to8wpr/configuration_files_are_canary_warning [3] https://lobste.rs/s/mdmcdi/little_languages_by_jon_bentley_1986#c_3miuf2 The implementation of the DSL was also highly hacky: a) It was happening in the tangle/ tool, but was utterly unrelated to tangling layers. b) There were several persnickety constraints on the different kinds of lines and the specific order they were expected in. I kept finding bugs where the translator would silently do the wrong thing. Or the error messages sucked, and readers may be stuck looking at the generated code to figure out what happened. Fixing error messages would require a lot more code, which is one of my arguments against DSLs in the first place: they may be easy to implement, but they're hard to design to go with the grain of the underlying platform. They require lots of iteration. Is that effort worth prioritizing in this project? On the other hand, the DSL did make at least some readers' life easier, the ones who weren't immediately put off by having to learn a strange syntax. There were fewer quotes to parse, fewer backslash escapes. Anyway, since there are also people who dislike having to put up with strange syntaxes, we'll call that consideration a wash and tear this DSL out. --- This commit was sheer drudgery. Hopefully it won't need to be redone with a new DSL because I grow sick of backslashes.
2019-03-13 01:56:55 +00:00
void test_break_to_label() {
run(
"def main [\n"
" {\n"
" {\n"
" break +target:label\n"
" 1:num <- copy 0\n"
" }\n"
" }\n"
" +target\n"
"]\n"
);
CHECK_TRACE_DOESNT_CONTAIN("mem: storing 0 in location 1");
}
5001 - drop the :(scenario) DSL I've been saying for a while[1][2][3] that adding extra abstractions makes things harder for newcomers, and adding new notations doubly so. And then I notice this DSL in my own backyard. Makes me feel like a hypocrite. [1] https://news.ycombinator.com/item?id=13565743#13570092 [2] https://lobste.rs/s/to8wpr/configuration_files_are_canary_warning [3] https://lobste.rs/s/mdmcdi/little_languages_by_jon_bentley_1986#c_3miuf2 The implementation of the DSL was also highly hacky: a) It was happening in the tangle/ tool, but was utterly unrelated to tangling layers. b) There were several persnickety constraints on the different kinds of lines and the specific order they were expected in. I kept finding bugs where the translator would silently do the wrong thing. Or the error messages sucked, and readers may be stuck looking at the generated code to figure out what happened. Fixing error messages would require a lot more code, which is one of my arguments against DSLs in the first place: they may be easy to implement, but they're hard to design to go with the grain of the underlying platform. They require lots of iteration. Is that effort worth prioritizing in this project? On the other hand, the DSL did make at least some readers' life easier, the ones who weren't immediately put off by having to learn a strange syntax. There were fewer quotes to parse, fewer backslash escapes. Anyway, since there are also people who dislike having to put up with strange syntaxes, we'll call that consideration a wash and tear this DSL out. --- This commit was sheer drudgery. Hopefully it won't need to be redone with a new DSL because I grow sick of backslashes.
2019-03-13 01:56:55 +00:00
void test_jump_if_to_label() {
run(
"def main [\n"
" {\n"
" {\n"
" jump-if 1, +target:label\n"
" 1:num <- copy 0\n"
" }\n"
" }\n"
" +target\n"
"]\n"
);
CHECK_TRACE_DOESNT_CONTAIN("mem: storing 0 in location 1");
}
5001 - drop the :(scenario) DSL I've been saying for a while[1][2][3] that adding extra abstractions makes things harder for newcomers, and adding new notations doubly so. And then I notice this DSL in my own backyard. Makes me feel like a hypocrite. [1] https://news.ycombinator.com/item?id=13565743#13570092 [2] https://lobste.rs/s/to8wpr/configuration_files_are_canary_warning [3] https://lobste.rs/s/mdmcdi/little_languages_by_jon_bentley_1986#c_3miuf2 The implementation of the DSL was also highly hacky: a) It was happening in the tangle/ tool, but was utterly unrelated to tangling layers. b) There were several persnickety constraints on the different kinds of lines and the specific order they were expected in. I kept finding bugs where the translator would silently do the wrong thing. Or the error messages sucked, and readers may be stuck looking at the generated code to figure out what happened. Fixing error messages would require a lot more code, which is one of my arguments against DSLs in the first place: they may be easy to implement, but they're hard to design to go with the grain of the underlying platform. They require lots of iteration. Is that effort worth prioritizing in this project? On the other hand, the DSL did make at least some readers' life easier, the ones who weren't immediately put off by having to learn a strange syntax. There were fewer quotes to parse, fewer backslash escapes. Anyway, since there are also people who dislike having to put up with strange syntaxes, we'll call that consideration a wash and tear this DSL out. --- This commit was sheer drudgery. Hopefully it won't need to be redone with a new DSL because I grow sick of backslashes.
2019-03-13 01:56:55 +00:00
void test_loop_unless_to_label() {
run(
"def main [\n"
" {\n"
" {\n"
" loop-unless 0, +target:label\n" // loop/break with a label don't care about braces
" 1:num <- copy 0\n"
" }\n"
" }\n"
" +target\n"
"]\n"
);
CHECK_TRACE_DOESNT_CONTAIN("mem: storing 0 in location 1");
}
5001 - drop the :(scenario) DSL I've been saying for a while[1][2][3] that adding extra abstractions makes things harder for newcomers, and adding new notations doubly so. And then I notice this DSL in my own backyard. Makes me feel like a hypocrite. [1] https://news.ycombinator.com/item?id=13565743#13570092 [2] https://lobste.rs/s/to8wpr/configuration_files_are_canary_warning [3] https://lobste.rs/s/mdmcdi/little_languages_by_jon_bentley_1986#c_3miuf2 The implementation of the DSL was also highly hacky: a) It was happening in the tangle/ tool, but was utterly unrelated to tangling layers. b) There were several persnickety constraints on the different kinds of lines and the specific order they were expected in. I kept finding bugs where the translator would silently do the wrong thing. Or the error messages sucked, and readers may be stuck looking at the generated code to figure out what happened. Fixing error messages would require a lot more code, which is one of my arguments against DSLs in the first place: they may be easy to implement, but they're hard to design to go with the grain of the underlying platform. They require lots of iteration. Is that effort worth prioritizing in this project? On the other hand, the DSL did make at least some readers' life easier, the ones who weren't immediately put off by having to learn a strange syntax. There were fewer quotes to parse, fewer backslash escapes. Anyway, since there are also people who dislike having to put up with strange syntaxes, we'll call that consideration a wash and tear this DSL out. --- This commit was sheer drudgery. Hopefully it won't need to be redone with a new DSL because I grow sick of backslashes.
2019-03-13 01:56:55 +00:00
void test_jump_runs_code_after_label() {
run(
"def main [\n"
// first a few lines of padding to exercise the offset computation
" 1:num <- copy 0\n"
" 2:num <- copy 0\n"
" 3:num <- copy 0\n"
" jump +target:label\n"
" 4:num <- copy 0\n"
" +target\n"
" 5:num <- copy 0\n"
"]\n"
);
CHECK_TRACE_CONTENTS(
"mem: storing 0 in location 5\n"
);
CHECK_TRACE_DOESNT_CONTAIN("mem: storing 0 in location 4");
}
5001 - drop the :(scenario) DSL I've been saying for a while[1][2][3] that adding extra abstractions makes things harder for newcomers, and adding new notations doubly so. And then I notice this DSL in my own backyard. Makes me feel like a hypocrite. [1] https://news.ycombinator.com/item?id=13565743#13570092 [2] https://lobste.rs/s/to8wpr/configuration_files_are_canary_warning [3] https://lobste.rs/s/mdmcdi/little_languages_by_jon_bentley_1986#c_3miuf2 The implementation of the DSL was also highly hacky: a) It was happening in the tangle/ tool, but was utterly unrelated to tangling layers. b) There were several persnickety constraints on the different kinds of lines and the specific order they were expected in. I kept finding bugs where the translator would silently do the wrong thing. Or the error messages sucked, and readers may be stuck looking at the generated code to figure out what happened. Fixing error messages would require a lot more code, which is one of my arguments against DSLs in the first place: they may be easy to implement, but they're hard to design to go with the grain of the underlying platform. They require lots of iteration. Is that effort worth prioritizing in this project? On the other hand, the DSL did make at least some readers' life easier, the ones who weren't immediately put off by having to learn a strange syntax. There were fewer quotes to parse, fewer backslash escapes. Anyway, since there are also people who dislike having to put up with strange syntaxes, we'll call that consideration a wash and tear this DSL out. --- This commit was sheer drudgery. Hopefully it won't need to be redone with a new DSL because I grow sick of backslashes.
2019-03-13 01:56:55 +00:00
void test_jump_fails_without_target() {
Hide_errors = true;
run(
"def main [\n"
" jump\n"
"]\n"
);
CHECK_TRACE_CONTENTS(
"error: main: 'jump' expects an ingredient but got 0\n"
);
}
2016-02-28 18:37:37 +00:00
5001 - drop the :(scenario) DSL I've been saying for a while[1][2][3] that adding extra abstractions makes things harder for newcomers, and adding new notations doubly so. And then I notice this DSL in my own backyard. Makes me feel like a hypocrite. [1] https://news.ycombinator.com/item?id=13565743#13570092 [2] https://lobste.rs/s/to8wpr/configuration_files_are_canary_warning [3] https://lobste.rs/s/mdmcdi/little_languages_by_jon_bentley_1986#c_3miuf2 The implementation of the DSL was also highly hacky: a) It was happening in the tangle/ tool, but was utterly unrelated to tangling layers. b) There were several persnickety constraints on the different kinds of lines and the specific order they were expected in. I kept finding bugs where the translator would silently do the wrong thing. Or the error messages sucked, and readers may be stuck looking at the generated code to figure out what happened. Fixing error messages would require a lot more code, which is one of my arguments against DSLs in the first place: they may be easy to implement, but they're hard to design to go with the grain of the underlying platform. They require lots of iteration. Is that effort worth prioritizing in this project? On the other hand, the DSL did make at least some readers' life easier, the ones who weren't immediately put off by having to learn a strange syntax. There were fewer quotes to parse, fewer backslash escapes. Anyway, since there are also people who dislike having to put up with strange syntaxes, we'll call that consideration a wash and tear this DSL out. --- This commit was sheer drudgery. Hopefully it won't need to be redone with a new DSL because I grow sick of backslashes.
2019-03-13 01:56:55 +00:00
void test_jump_fails_without_target_2() {
Hide_errors = true;
run(
"def main [\n"
" jump-if true\n"
"]\n"
);
CHECK_TRACE_CONTENTS(
"error: main: 'jump-if true' expects 2 ingredients but got 1\n"
);
}
2016-02-28 18:37:37 +00:00
5001 - drop the :(scenario) DSL I've been saying for a while[1][2][3] that adding extra abstractions makes things harder for newcomers, and adding new notations doubly so. And then I notice this DSL in my own backyard. Makes me feel like a hypocrite. [1] https://news.ycombinator.com/item?id=13565743#13570092 [2] https://lobste.rs/s/to8wpr/configuration_files_are_canary_warning [3] https://lobste.rs/s/mdmcdi/little_languages_by_jon_bentley_1986#c_3miuf2 The implementation of the DSL was also highly hacky: a) It was happening in the tangle/ tool, but was utterly unrelated to tangling layers. b) There were several persnickety constraints on the different kinds of lines and the specific order they were expected in. I kept finding bugs where the translator would silently do the wrong thing. Or the error messages sucked, and readers may be stuck looking at the generated code to figure out what happened. Fixing error messages would require a lot more code, which is one of my arguments against DSLs in the first place: they may be easy to implement, but they're hard to design to go with the grain of the underlying platform. They require lots of iteration. Is that effort worth prioritizing in this project? On the other hand, the DSL did make at least some readers' life easier, the ones who weren't immediately put off by having to learn a strange syntax. There were fewer quotes to parse, fewer backslash escapes. Anyway, since there are also people who dislike having to put up with strange syntaxes, we'll call that consideration a wash and tear this DSL out. --- This commit was sheer drudgery. Hopefully it won't need to be redone with a new DSL because I grow sick of backslashes.
2019-03-13 01:56:55 +00:00
void test_recipe_fails_on_duplicate_jump_target() {
Hide_errors = true;
run(
"def main [\n"
" +label\n"
" 1:num <- copy 0\n"
" +label\n"
" 2:num <- copy 0\n"
"]\n"
);
CHECK_TRACE_CONTENTS(
"error: main: duplicate label '+label'\n"
);
}
5001 - drop the :(scenario) DSL I've been saying for a while[1][2][3] that adding extra abstractions makes things harder for newcomers, and adding new notations doubly so. And then I notice this DSL in my own backyard. Makes me feel like a hypocrite. [1] https://news.ycombinator.com/item?id=13565743#13570092 [2] https://lobste.rs/s/to8wpr/configuration_files_are_canary_warning [3] https://lobste.rs/s/mdmcdi/little_languages_by_jon_bentley_1986#c_3miuf2 The implementation of the DSL was also highly hacky: a) It was happening in the tangle/ tool, but was utterly unrelated to tangling layers. b) There were several persnickety constraints on the different kinds of lines and the specific order they were expected in. I kept finding bugs where the translator would silently do the wrong thing. Or the error messages sucked, and readers may be stuck looking at the generated code to figure out what happened. Fixing error messages would require a lot more code, which is one of my arguments against DSLs in the first place: they may be easy to implement, but they're hard to design to go with the grain of the underlying platform. They require lots of iteration. Is that effort worth prioritizing in this project? On the other hand, the DSL did make at least some readers' life easier, the ones who weren't immediately put off by having to learn a strange syntax. There were fewer quotes to parse, fewer backslash escapes. Anyway, since there are also people who dislike having to put up with strange syntaxes, we'll call that consideration a wash and tear this DSL out. --- This commit was sheer drudgery. Hopefully it won't need to be redone with a new DSL because I grow sick of backslashes.
2019-03-13 01:56:55 +00:00
void test_jump_ignores_nontarget_label() {
Hide_errors = true;
run(
"def main [\n"
// first a few lines of padding to exercise the offset computation
" 1:num <- copy 0\n"
" 2:num <- copy 0\n"
" 3:num <- copy 0\n"
" jump $target:label\n"
" 4:num <- copy 0\n"
" $target\n"
" 5:num <- copy 0\n"
"]\n"
);
CHECK_TRACE_CONTENTS(
"error: main: can't jump to label '$target'\n"
);
}