2018-10-22 04:04:38 +00:00
|
|
|
//: Automatically aggregate functions starting with 'test-' into a test suite
|
|
|
|
//: called 'run-tests'. Running this function will run all tests.
|
|
|
|
//:
|
|
|
|
//: This is actually SubX's first (trivial) compiler. We generate all the code
|
|
|
|
//: needed for the 'run-tests' function.
|
|
|
|
//:
|
|
|
|
//: By convention, temporary functions needed by tests will start with
|
|
|
|
//: '_test-'.
|
2018-09-21 20:44:16 +00:00
|
|
|
|
2018-09-23 04:56:00 +00:00
|
|
|
//: We don't rely on any transforms running in previous layers, but this layer
|
|
|
|
//: knows about labels and will emit labels for previous layers to transform.
|
2018-09-21 20:44:16 +00:00
|
|
|
:(after "Begin Transforms")
|
2018-09-23 04:56:00 +00:00
|
|
|
// Begin Level-4 Transforms
|
2018-09-21 20:44:16 +00:00
|
|
|
Transform.push_back(create_test_function);
|
2018-09-23 04:56:00 +00:00
|
|
|
// End Level-4 Transforms
|
2018-09-21 20:44:16 +00:00
|
|
|
|
|
|
|
:(code)
|
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_run_test() {
|
2019-05-14 00:27:45 +00:00
|
|
|
Mem.push_back(vma(0xbd000000)); // manually allocate memory
|
|
|
|
Reg[ESP].u = 0xbd000100;
|
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
|
|
|
run(
|
2019-05-18 07:00:18 +00:00
|
|
|
"== code 0x1\n" // code segment
|
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
|
|
|
"main:\n"
|
|
|
|
" e8/call run-tests/disp32\n" // 5 bytes
|
|
|
|
" f4/halt\n" // 1 byte
|
|
|
|
"test-foo:\n" // offset 7
|
|
|
|
" 01 d8\n" // just some unique instruction: add EBX to EAX
|
|
|
|
" c3/return\n"
|
|
|
|
);
|
|
|
|
// check that code in test-foo ran (implicitly called by run-tests)
|
|
|
|
CHECK_TRACE_CONTENTS(
|
|
|
|
"run: 0x00000007 opcode: 01\n"
|
|
|
|
);
|
|
|
|
}
|
|
|
|
|
2018-09-21 20:44:16 +00:00
|
|
|
void create_test_function(program& p) {
|
|
|
|
if (p.segments.empty()) return;
|
2019-05-18 07:00:18 +00:00
|
|
|
segment& code = *find(p, "code");
|
2019-02-25 08:17:46 +00:00
|
|
|
trace(3, "transform") << "-- create 'run-tests'" << end();
|
2018-09-21 20:44:16 +00:00
|
|
|
vector<line> new_insts;
|
|
|
|
for (int i = 0; i < SIZE(code.lines); ++i) {
|
|
|
|
line& inst = code.lines.at(i);
|
|
|
|
for (int j = 0; j < SIZE(inst.words); ++j) {
|
|
|
|
const word& curr = inst.words.at(j);
|
|
|
|
if (*curr.data.rbegin() != ':') continue; // not a label
|
2018-10-06 02:49:47 +00:00
|
|
|
if (!starts_with(curr.data, "test-")) continue;
|
2018-09-21 20:44:16 +00:00
|
|
|
string fn = drop_last(curr.data);
|
|
|
|
new_insts.push_back(call(fn));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (new_insts.empty()) return; // no tests found
|
2018-10-06 02:49:47 +00:00
|
|
|
code.lines.push_back(label("run-tests"));
|
2018-09-21 20:44:16 +00:00
|
|
|
code.lines.insert(code.lines.end(), new_insts.begin(), new_insts.end());
|
|
|
|
code.lines.push_back(ret());
|
|
|
|
}
|
|
|
|
|
|
|
|
string to_string(const segment& s) {
|
|
|
|
ostringstream out;
|
|
|
|
for (int i = 0; i < SIZE(s.lines); ++i) {
|
|
|
|
const line& l = s.lines.at(i);
|
|
|
|
for (int j = 0; j < SIZE(l.words); ++j) {
|
|
|
|
if (j > 0) out << ' ';
|
|
|
|
out << to_string(l.words.at(j));
|
|
|
|
}
|
|
|
|
out << '\n';
|
|
|
|
}
|
|
|
|
return out.str();
|
|
|
|
}
|
|
|
|
|
|
|
|
line call(string s) {
|
|
|
|
line result;
|
|
|
|
result.words.push_back(call());
|
|
|
|
result.words.push_back(disp32(s));
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
|
|
|
word call() {
|
|
|
|
word result;
|
|
|
|
result.data = "e8";
|
|
|
|
result.metadata.push_back("call");
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
|
|
|
word disp32(string s) {
|
|
|
|
word result;
|
|
|
|
result.data = s;
|
|
|
|
result.metadata.push_back("disp32");
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
|
|
|
line ret() {
|
|
|
|
line result;
|
|
|
|
result.words.push_back(word());
|
|
|
|
result.words.back().data = "c3";
|
|
|
|
result.words.back().metadata.push_back("return");
|
|
|
|
return result;
|
|
|
|
}
|