From bfdc8d2d843c08bd01517256a63518d03da236f6 Mon Sep 17 00:00:00 2001
From: Marvin Borner
Date: Tue, 27 Jul 2021 21:10:34 +0200
Subject: Started treeify
---
.gitignore | 1 +
Makefile | 18 +++++----
inc/context.h | 32 +++++++++++++++
inc/lint.h | 8 ++++
inc/log.h | 9 +++++
inc/preprocess.h | 8 ++++
inc/tokenize.h | 35 +++++++++++++++++
inc/treeify.h | 71 +++++++++++++++++++++++++++++++++
src/context.c | 12 +++++-
src/inc/context.h | 23 -----------
src/inc/lib.h | 33 ----------------
src/inc/lint.h | 8 ----
src/inc/log.h | 9 -----
src/inc/preprocess.h | 8 ----
src/inc/tokenize.h | 35 -----------------
src/lint.c | 3 +-
src/log.c | 5 ++-
src/main.c | 3 +-
src/preprocess.c | 8 +++-
src/tokenize.c | 21 +++++-----
src/treeify.c | 109 +++++++++++++++++++++++++++++++++++++++++++++++++++
21 files changed, 318 insertions(+), 141 deletions(-)
create mode 100644 inc/context.h
create mode 100644 inc/lint.h
create mode 100644 inc/log.h
create mode 100644 inc/preprocess.h
create mode 100644 inc/tokenize.h
create mode 100644 inc/treeify.h
delete mode 100644 src/inc/context.h
delete mode 100644 src/inc/lib.h
delete mode 100644 src/inc/lint.h
delete mode 100644 src/inc/log.h
delete mode 100644 src/inc/preprocess.h
delete mode 100644 src/inc/tokenize.h
create mode 100644 src/treeify.c
diff --git a/.gitignore b/.gitignore
index 017fcf9..50660f9 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,3 +1,4 @@
build/
+tags
compile_commands.json
diff --git a/Makefile b/Makefile
index 7e61d70..9f74876 100644
--- a/Makefile
+++ b/Makefile
@@ -1,23 +1,25 @@
-SOURCEDIR = src
-BUILDDIR = build
+INCLUDEDIR = $(PWD)/inc
+SOURCEDIR = $(PWD)/src
+BUILDDIR = $(PWD)/build
SOURCES = $(wildcard $(SOURCEDIR)/*.c)
OBJS = $(patsubst $(SOURCEDIR)/%.c, $(BUILDDIR)/%.o, $(SOURCES))
-CC = gcc
-CFLAGS = -Ofast -Wall -Wextra -Werror -pedantic -Wshadow -Wpointer-arith -Wwrite-strings -Wredundant-decls -Wnested-externs -Wformat=1 -Wmissing-declarations -Wstrict-prototypes -Wmissing-prototypes -Wcast-qual -Wswitch-default -Wswitch-enum -Wlogical-op -Wunreachable-code -Wundef -Wold-style-definition -Wvla -std=c99 -fsanitize=address -fsanitize=undefined -I$(SOURCEDIR)/inc/
+CC = ccache gcc
+CFLAGS = -Ofast -Wall -Wextra -Werror -pedantic -Wshadow -Wpointer-arith -Wwrite-strings -Wredundant-decls -Wnested-externs -Wformat=1 -Wmissing-declarations -Wstrict-prototypes -Wmissing-prototypes -Wcast-qual -Wswitch-default -Wswitch-enum -Wlogical-op -Wunreachable-code -Wundef -Wold-style-definition -Wvla -std=c99 -fsanitize=address -fsanitize=undefined -fstack-protector-strong -I$(INCLUDEDIR)
all: $(OBJS)
- @$(CC) -o ./$(BUILDDIR)/out $(CFLAGS) $^
+ @$(CC) -o $(BUILDDIR)/out $(CFLAGS) $^
clean:
@$(RM) -rf $(BUILDDIR)
-run: clean all
- @./$(BUILDDIR)/out test.fun
+run: clean all sync
+ @$(BUILDDIR)/out test.fun
$(BUILDDIR)/%.o: $(SOURCEDIR)/%.c
@mkdir -p $(BUILDDIR)
@$(CC) -c -o $@ $(CFLAGS) $<
sync:
- @make --always-make --dry-run | grep -wE 'gcc|g\+\+' | grep -w '\-c' | jq -nR '[inputs|{directory:"$(PWD)", command:., file: match(" [^ ]+$$").string[1:]}]' > compile_commands.json
+ @ctags -R --exclude=.git --exclude=build .
+ @make --always-make --dry-run | grep -wE 'gcc|g\+\+' | grep -w '\-c' | jq -nR '[inputs|{directory:".", command:., file: match(" [^ ]+$$").string[1:]}]' > compile_commands.json
diff --git a/inc/context.h b/inc/context.h
new file mode 100644
index 0000000..3f7b686
--- /dev/null
+++ b/inc/context.h
@@ -0,0 +1,32 @@
+#ifndef CONTEXT_H
+#define CONTEXT_H
+
+#include <stddef.h>
+
+typedef struct {
+ size_t start, end;
+} ctx_string;
+
+struct ctx {
+ size_t line;
+ size_t column;
+ const char *path;
+
+ char *raw;
+ char *data;
+ size_t size;
+
+ size_t token_count;
+ struct token *tokens;
+
+ struct {
+ struct tree *head;
+ struct node *current;
+ } tree;
+};
+
+struct ctx *context_create(const char *path);
+void context_destroy(struct ctx *ctx);
+void context_rewind(struct ctx *ctx);
+
+#endif
diff --git a/inc/lint.h b/inc/lint.h
new file mode 100644
index 0000000..9c3f808
--- /dev/null
+++ b/inc/lint.h
@@ -0,0 +1,8 @@
+#ifndef LINT_H
+#define LINT_H
+
+#include <context.h>
+
+void lint(struct ctx *ctx);
+
+#endif
diff --git a/inc/log.h b/inc/log.h
new file mode 100644
index 0000000..4698f9c
--- /dev/null
+++ b/inc/log.h
@@ -0,0 +1,9 @@
+#ifndef LOG_H
+#define LOG_H
+
+#include <context.h>
+
+__attribute__((noreturn)) void errln(struct ctx *ctx, const char *fmt, ...);
+__attribute__((noreturn)) void err(const char *fmt, ...);
+
+#endif
diff --git a/inc/preprocess.h b/inc/preprocess.h
new file mode 100644
index 0000000..e57af10
--- /dev/null
+++ b/inc/preprocess.h
@@ -0,0 +1,8 @@
+#ifndef PREPROCESS_H
+#define PREPROCESS_H
+
+#include <context.h>
+
+void preprocess(struct ctx *ctx);
+
+#endif
diff --git a/inc/tokenize.h b/inc/tokenize.h
new file mode 100644
index 0000000..0142bb9
--- /dev/null
+++ b/inc/tokenize.h
@@ -0,0 +1,35 @@
+#ifndef TOKENIZE_H
+#define TOKENIZE_H
+
+#include <context.h>
+
+#define TOKENS_MAX 4096
+
+enum token_type {
+ UNKNOWN,
+
+ TYPE,
+ TYPEDELIM,
+ PARAM,
+
+ IDENT,
+ OPERATOR,
+
+ LPAREN,
+ RPAREN,
+ EQUAL,
+
+ NEWLINE,
+ EOL,
+ END,
+};
+
+struct token {
+ enum token_type type;
+ ctx_string string;
+};
+
+void token_print(struct ctx *ctx, struct token *tok);
+void tokenize(struct ctx *ctx);
+
+#endif
diff --git a/inc/treeify.h b/inc/treeify.h
new file mode 100644
index 0000000..3af67f8
--- /dev/null
+++ b/inc/treeify.h
@@ -0,0 +1,71 @@
+#ifndef TREEIFY_H
+#define TREEIFY_H
+
+#include <context.h>
+
+enum node_type {
+ EXPRESSION,
+ DECLARATION,
+};
+
+/**
+ * Expressions
+ */
+
+// (*f* x y)
+struct node_expression_identifier {
+ ctx_string name; // f
+ ctx_string type; // u32
+};
+
+// (f *x* *y*)
+struct node_expression_parameter {
+ ctx_string name; // x or y
+ ctx_string type; // u32
+};
+
+// *(f x y)*
+struct node_expression {
+ struct node_expression_identifier *callee; // f
+ struct node_expression_parameter *parameters; // x y
+};
+
+/**
+ * Declarations
+ */
+
+// *u32:f* u32:x u32:y = (...)
+struct node_declaration_callee {
+ ctx_string name; // f
+ ctx_string type; // u32
+};
+
+// u32:f *u32:x* *u32:y* = (...)
+struct node_declaration_parameter {
+ ctx_string name; // x or y
+ ctx_string type; // u32
+};
+
+// *u32:f u32:x u32:y* = (...) OR
+// *u32:a* = ...
+struct node_declaration {
+ struct node_declaration_callee callee; // f
+ struct node_declaration_parameter *parameters; // x y OR NULL
+};
+
+struct node {
+ enum node_type type;
+ /* struct node *next; */
+ void *data;
+};
+
+struct tree {
+ struct node *node;
+};
+
+struct tree *tree_create(void);
+void tree_destroy(struct tree *tree);
+
+void treeify(struct ctx *ctx);
+
+#endif
diff --git a/src/context.c b/src/context.c
index c0898d8..ff0291a 100644
--- a/src/context.c
+++ b/src/context.c
@@ -1,6 +1,10 @@
+#include <assert.h>
+#include <stdio.h>
+#include <stdlib.h>
+
#include <context.h>
-#include <lib.h>
#include <tokenize.h>
+#include <treeify.h>
struct ctx *context_create(const char *path)
{
@@ -24,6 +28,9 @@ struct ctx *context_create(const char *path)
ctx->raw[ctx->size] = 0;
+ ctx->tree.head = tree_create();
+ ctx->tree.current = NULL;
+
return ctx;
}
@@ -41,6 +48,9 @@ void context_destroy(struct ctx *ctx)
if (ctx->tokens)
free(ctx->tokens);
+ if (ctx->tree.head)
+ tree_destroy(ctx->tree.head);
+
free(ctx);
}
diff --git a/src/inc/context.h b/src/inc/context.h
deleted file mode 100644
index 603f42e..0000000
--- a/src/inc/context.h
+++ /dev/null
@@ -1,23 +0,0 @@
-#ifndef CONTEXT_H
-#define CONTEXT_H
-
-#include <lib.h>
-
-struct ctx {
- size_t line;
- size_t column;
- const char *path;
-
- char *raw;
- char *data;
- size_t size;
-
- size_t token_count;
- struct token *tokens;
-};
-
-struct ctx *context_create(const char *path);
-void context_destroy(struct ctx *ctx);
-void context_rewind(struct ctx *ctx);
-
-#endif
diff --git a/src/inc/lib.h b/src/inc/lib.h
deleted file mode 100644
index 0402990..0000000
--- a/src/inc/lib.h
+++ /dev/null
@@ -1,33 +0,0 @@
-#ifndef LIB_H
-#define LIB_H
-
-// Compatibility imports (mainly for Melvix)
-
-#ifdef __melvix__
-
-#include <arg.h>
-#include <assert.h>
-#include <def.h>
-#include <mem.h>
-#include <print.h>
-#include <str.h>
-#include <sys.h>
-typedef size_t u32;
-
-#elif defined(__linux__) || defined(unix) || defined(__unix__) || defined(__unix) || \
- defined(__APPLE__) || defined(__FreeBSD)
-
-#include <assert.h>
-#include <ctype.h>
-#include <stdarg.h>
-#include <stdbool.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include <sys/param.h>
-
-#else
-#error "Unknown operating system"
-#endif
-
-#endif
diff --git a/src/inc/lint.h b/src/inc/lint.h
deleted file mode 100644
index 9c3f808..0000000
--- a/src/inc/lint.h
+++ /dev/null
@@ -1,8 +0,0 @@
-#ifndef LINT_H
-#define LINT_H
-
-#include <context.h>
-
-void lint(struct ctx *ctx);
-
-#endif
diff --git a/src/inc/log.h b/src/inc/log.h
deleted file mode 100644
index 4698f9c..0000000
--- a/src/inc/log.h
+++ /dev/null
@@ -1,9 +0,0 @@
-#ifndef LOG_H
-#define LOG_H
-
-#include <context.h>
-
-__attribute__((noreturn)) void errln(struct ctx *ctx, const char *fmt, ...);
-__attribute__((noreturn)) void err(const char *fmt, ...);
-
-#endif
diff --git a/src/inc/preprocess.h b/src/inc/preprocess.h
deleted file mode 100644
index e57af10..0000000
--- a/src/inc/preprocess.h
+++ /dev/null
@@ -1,8 +0,0 @@
-#ifndef PREPROCESS_H
-#define PREPROCESS_H
-
-#include <context.h>
-
-void preprocess(struct ctx *ctx);
-
-#endif
diff --git a/src/inc/tokenize.h b/src/inc/tokenize.h
deleted file mode 100644
index 557da9c..0000000
--- a/src/inc/tokenize.h
+++ /dev/null
@@ -1,35 +0,0 @@
-#ifndef TOKENIZE_H
-#define TOKENIZE_H
-
-#include <context.h>
-
-#define TOKENS_MAX 4096
-
-enum token_type {
- UNKNOWN,
-
- TYPE,
- TYPEDELIM,
- PARAM,
-
- IDENT,
- OPERATOR,
-
- LPAREN,
- RPAREN,
- EQUAL,
-
- NEWLINE,
- EOL,
- END,
-};
-
-struct token {
- enum token_type type;
- size_t start, end;
-};
-
-void tokens_print(struct ctx *ctx);
-void tokenize(struct ctx *ctx);
-
-#endif
diff --git a/src/lint.c b/src/lint.c
index 6b7fd44..5ff1864 100644
--- a/src/lint.c
+++ b/src/lint.c
@@ -1,4 +1,5 @@
-#include <lib.h>
+#include <stddef.h>
+
#include <lint.h>
#include <log.h>
#include <tokenize.h>
diff --git a/src/log.c b/src/log.c
index ee06313..5820a67 100644
--- a/src/log.c
+++ b/src/log.c
@@ -1,4 +1,7 @@
-#include <lib.h>
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+
#include <log.h>
static void context_print(FILE *fd, struct ctx *ctx)
diff --git a/src/main.c b/src/main.c
index 74cc3d6..0e6807d 100644
--- a/src/main.c
+++ b/src/main.c
@@ -3,6 +3,7 @@
#include <log.h>
#include <preprocess.h>
#include <tokenize.h>
+#include <treeify.h>
int main(int argc, char *argv[])
{
@@ -13,7 +14,7 @@ int main(int argc, char *argv[])
preprocess(ctx);
tokenize(ctx);
lint(ctx);
- tokens_print(ctx);
+ treeify(ctx);
context_destroy(ctx);
diff --git a/src/preprocess.c b/src/preprocess.c
index dec55b7..8d9394a 100644
--- a/src/preprocess.c
+++ b/src/preprocess.c
@@ -1,4 +1,8 @@
-#include <lib.h>
+#include <assert.h>
+#include <math.h>
+#include <stddef.h>
+#include <string.h>
+
#include <log.h>
#include <preprocess.h>
@@ -29,7 +33,7 @@ void preprocess(struct ctx *ctx)
} else if (cur == '\0') {
break;
} else if (cur == '#' && ctx->column == 1) {
- if (strncmp(ctx->raw + i + 1, "inc ", MIN(4, ctx->size - i)) == 0) {
+ if (strncmp(ctx->raw + i + 1, "inc ", fmin(4, ctx->size - i)) == 0) {
// TODO: Add include features
} else if (*(ctx->raw + i + 1) == '#') {
// Comment
diff --git a/src/tokenize.c b/src/tokenize.c
index 69f17c9..179b251 100644
--- a/src/tokenize.c
+++ b/src/tokenize.c
@@ -1,4 +1,9 @@
-#include <lib.h>
+#include <assert.h>
+#include <ctype.h>
+#include <stdbool.h>
+#include <stddef.h>
+#include <stdio.h>
+
#include <log.h>
#include <tokenize.h>
@@ -72,8 +77,8 @@ static void token_add(struct ctx *ctx, enum token_type type, size_t start, size_
{
struct token token = { 0 };
token.type = type;
- token.start = start;
- token.end = end;
+ token.string.start = start;
+ token.string.end = end;
assert(++ctx->token_count < TOKENS_MAX);
ctx->tokens[ctx->token_count] = token;
@@ -86,7 +91,7 @@ static void token_add(struct ctx *ctx, enum token_type type, size_t start, size_
}
}
-static void token_print(struct ctx *ctx, struct token *token)
+void token_print(struct ctx *ctx, struct token *token)
{
assert(token->type != UNKNOWN);
@@ -96,17 +101,11 @@ static void token_print(struct ctx *ctx, struct token *token)
return;
}
- for (size_t i = token->start; i < token->end; i++)
+ for (size_t i = token->string.start; i < token->string.end; i++)
printf("%c", ctx->data[i]);
printf("'\n");
}
-void tokens_print(struct ctx *ctx)
-{
- for (size_t i = 1; i < ctx->token_count; i++)
- token_print(ctx, &ctx->tokens[i]);
-}
-
void tokenize(struct ctx *ctx)
{
for (size_t i = 0; i < ctx->size; i++) {
diff --git a/src/treeify.c b/src/treeify.c
new file mode 100644
index 0000000..32a12c2
--- /dev/null
+++ b/src/treeify.c
@@ -0,0 +1,109 @@
+#include <assert.h>
+#include <stdlib.h>
+
+#include <log.h>
+#include <tokenize.h>
+#include <treeify.h>
+
+static void expected(enum token_type type)
+{
+ err("Expected token of type %d while creating tree!\n", type);
+}
+
+static void unexpected(struct ctx *ctx, struct token *token)
+{
+ token_print(ctx, token);
+ err("Unexpected token while creating tree!\n");
+}
+
+static struct token *next(struct token *token, size_t i)
+{
+ return token + i;
+}
+
+static struct token *parse_declaration(struct ctx *ctx, struct token *token)
+{
+ if (next(token, 1)->type != TYPEDELIM)
+ unexpected(ctx, next(token, 1));
+ if (next(token, 2)->type != PARAM)
+ unexpected(ctx, next(token, 2));
+
+ // Search for equal sign
+ struct token *iterator = token;
+ while ((iterator = next(iterator, 1))) {
+ if (iterator->type == EQUAL)
+ break;
+ if (iterator->type == EOL || iterator->type == END)
+ expected(EQUAL);
+ }
+
+ struct node_declaration *node = malloc(sizeof(*node));
+ node->callee.type = token->string;
+ node->callee.name = next(token, 2)->string;
+
+ // Magic
+ size_t diff = iterator - token - 3;
+ assert(diff % 3 == 0);
+ node->parameters = malloc((diff / 3 + 1) * sizeof(*node->parameters));
+ for (size_t i = 0; i < diff / 3; i++) {
+ struct token *param = token + (i + 1) * 3;
+ assert(param->type == TYPE);
+ assert(next(param, 2)->type == PARAM);
+ node->parameters[i].type = param->string;
+ node->parameters[i].name = next(param, 2)->string;
+ }
+
+ return next(iterator, 1);
+}
+
+static struct token *parse(struct ctx *ctx, struct token *token)
+{
+ switch (token->type) {
+ case TYPE:
+ return parse_declaration(ctx, token);
+ case UNKNOWN:
+ case TYPEDELIM:
+ case PARAM:
+ case IDENT:
+ case OPERATOR:
+ case LPAREN:
+ case RPAREN:
+ case EQUAL:
+ case NEWLINE:
+ case EOL:
+ case END:
+ default:
+ unexpected(ctx, token);
+ }
+
+ return NULL;
+}
+
+struct tree *tree_create(void)
+{
+ struct tree *tree = malloc(sizeof(*tree));
+ tree->node = NULL;
+ return tree;
+}
+
+void tree_destroy(struct tree *tree)
+{
+ // TODO: Destroy nodes
+ free(tree);
+}
+
+void treeify(struct ctx *ctx)
+{
+ struct token *token = &ctx->tokens[1];
+ while (token) {
+ assert(token->type != UNKNOWN);
+ if (token->type == NEWLINE || token->type == EOL) {
+ token = next(token, 1);
+ continue;
+ }
+ if (token->type == END)
+ break;
+
+ token = parse(ctx, token);
+ }
+}
--
cgit v1.2.3