Added assembler and linker in bfc.c instead of Makefile, Added options to choose output filename and if output is assembly or executable

3 files changed, 280 insertions, 64 deletions

diff --git a/.clang-format b/.clang-format
new file mode 100644
index 0000000..f8821e3
--- /dev/null
+++ b/.clang-format
@@ -0,0 +1,69 @@
+---
+# BasedOnStyle: LLVM
+---
+
+Language: Cpp
+
+# PPIndentWidth: 1  # clang-format 13 only
+
+IndentWidth: 4
+TabWidth: 4
+UseTab: Never
+
+ColumnLimit: 100
+
+DerivePointerAlignment: false
+PointerAlignment: Right
+
+AlignConsecutiveDeclarations: Consecutive
+AlignConsecutiveAssignments: None
+AlignConsecutiveBitFields: Consecutive
+# AlignArrayOfStructures: Left  # clang-format 13 only
+AlignEscapedNewlines: Left
+AlignTrailingComments: true
+AlignAfterOpenBracket: Align
+
+AlwaysBreakAfterDefinitionReturnType: All
+AlwaysBreakAfterReturnType: All
+
+SpaceBeforeParens: ControlStatements
+
+AllowShortFunctionsOnASingleLine: None
+AllowShortBlocksOnASingleLine: Never
+AllowShortIfStatementsOnASingleLine: Never
+
+AllowAllArgumentsOnNextLine: true
+AllowAllParametersOfDeclarationOnNextLine: true
+
+AlwaysBreakBeforeMultilineStrings: false
+
+BinPackArguments: false
+BinPackParameters: false
+
+BreakBeforeBraces: Allman
+BreakBeforeTernaryOperators: true
+
+IncludeBlocks: Merge
+
+KeepEmptyLinesAtTheStartOfBlocks: false
+MaxEmptyLinesToKeep: 1
+
+PenaltyBreakBeforeFirstCallParameter: 1
+PenaltyBreakString: 1
+PenaltyExcessCharacter: 1000
+
+SpaceAfterCStyleCast: false
+SpaceBeforeAssignmentOperators: true
+SpaceBeforeSquareBrackets: false
+SpaceInEmptyParentheses: false
+SpacesInCStyleCastParentheses: false
+SpacesInParentheses: false
+SpacesBeforeTrailingComments: 2
+SpacesInSquareBrackets: false
+SpaceAroundPointerQualifiers: Default
+
+AlignOperands: Align
+
+Cpp11BracedListStyle: true
+
+# vim:ft=yaml
diff --git a/Makefile b/Makefile
index 07a867d..8d1c0b8 100644
--- a/Makefile
+++ b/Makefile
@@ -1,26 +1,13 @@
 CC = gcc
 CCFLAGS = -Wall -Wextra -std=c99
-NASM = nasm
-
-ifeq ($(shell uname),Linux)
-	NASMFLAGS = -f elf64
-endif
-ifeq ($(shell uname),Darwin)
-	NASMFLAGS = -f macho64
-endif
-ifeq ($(NASMFLAGS),)
-	$(error architecture: $(shell uname) not recognized)
-endif
 
 SRC = bfc.c
 NAME = bfc
 
-ASMDIR = asm
 BINDIR = bin
 
 EXAMPLESDIR = examples
 EXAMPLES = $(shell find $(EXAMPLESDIR) -name '*.bf' -type f)
-EXAMPLES_ASM = $(EXAMPLES:$(EXAMPLESDIR)/%.bf=$(ASMDIR)/%.asm)
 EXAMPLES_NAME = $(EXAMPLES:$(EXAMPLESDIR)/%.bf=$(BINDIR)/%)
 
 all: $(NAME) $(EXAMPLES_NAME)
@@ -28,26 +15,21 @@ all: $(NAME) $(EXAMPLES_NAME)
 $(NAME): bfc.c
 	$(CC) $(CCFLAGS) -o $@ $<
 
-$(BINDIR)/%: $(ASMDIR)/%.asm $(BINDIR)
-	$(NASM) $(NASMFLAGS) -o $@.o $<
-	ld -o $@ $@.o
-
-$(ASMDIR)/%.asm: $(EXAMPLESDIR)/%.bf $(ASMDIR)
-	./$(NAME) < $< > $@
+$(BINDIR)/%: $(EXAMPLESDIR)/%.bf $(BINDIR)
+	./$(NAME) -o $@ $<
 
 $(BINDIR):
 	mkdir -pv $@
 
-$(ASMDIR):
-	mkdir -pv $@
-
 clean:
-	- rm -rv $(ASMDIR)
 	- rm -rv $(BINDIR)
 	- rm $(NAME)
 
 re: clean all
 
+format:
+	clang-format -i bfc.c
+
 # weird GNU make behavior where it rm's every asm files at the end
 # (https://stackoverflow.com/questions/47447369)
 .PRECIOUS: $(EXAMPLES_ASM)
diff --git a/bfc.c b/bfc.c
index 30098d1..6fff9cd 100644
--- a/bfc.c
+++ b/bfc.c
@@ -1,52 +1,190 @@
+#define _XOPEN_SOURCE 500
+#include <ctype.h>
+#include <errno.h>
 #include <getopt.h>
+#include <stdarg.h>
+#include <stdbool.h>
 #include <stddef.h>
-#include <stdlib.h>
 #include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/wait.h>
+#include <unistd.h>
 
-static const size_t buffer_size = 256;
-/* static const char* asm_filename = "out.asm"; */
+#ifdef __linux__
+#define NASM_FORMAT "elf64"
+#endif
+#ifdef __APPLE__
+#define NASM_FORMAT "macho64"
+#endif
+#ifndef NASM_FORMAT
+#error "architecture not supported"
+#endif
 
-#define LABEL_COUNT_STACK_SIZE 256
+#define LABEL_COUNT_STACK_SIZE 4096
 static size_t label_stack[LABEL_COUNT_STACK_SIZE + 1] = {0};
 static size_t label_stack_frame = 0;
 
-int main(int argc, char *argv[])
+static bool   compile_to_exec = true;
+static char  *output_filename = NULL;
+static size_t buffer_len = 256;
+static size_t buffer_elem_bytes = 1;
+
+static char *elem_bytes_asm_str = NULL;
+static char *elem_bytes_asm_str_res = NULL;
+
+#define TMP_FILEPATH_TEMPLATE "/tmp/bfc_asm_XXXXXX"
+#define TMP_OBJ_FILEPATH_TEMPLATE (TMP_FILEPATH_TEMPLATE ".o")
+static char tmp_filepath[] = TMP_FILEPATH_TEMPLATE;
+static char tmp_obj_filepath[] = TMP_OBJ_FILEPATH_TEMPLATE;
+
+static void
+die(const char *fmt, ...)
+{
+    va_list ap;
+    va_start(ap, fmt);
+    fputs("bfc: ", stderr);
+    vfprintf(stderr, fmt, ap);
+    fputc('\n', stderr);
+    va_end(ap);
+    exit(EXIT_FAILURE);
+}
+
+static size_t
+xatoul(char *s)
+{
+    if (!isdigit(*s))
+        die("'%s' is not a valid number", s);
+    char  *end;
+    size_t num = strtoul(s, &end, 10);
+    if (*end != '\0')
+        die("'%s' is not a valid number", s);
+    return num;
+}
+
+static void
+run_subprocess(char *name, char *const argv[])
 {
-    // int option;
-
-    /* while ((option = getopt(argc, argv, "So:b:")) != -1) */
-    /* { */
-    /*     switch (option) */
-    /*     { */
-    /*         case 'S': */
-    /*             assembly = true; */
-    /*         case 'o': */
-    /*             output_filename = optstring */
-    /*         case 'b': */
-    /*             buffer_length = atoi(optstring); */
-    /*         case 'e': */
-    /*             // byte, word, dword or qword */
-    /*             buffer_elem = x; */
-    /*     } */
-    /* } */
+    pid_t pid = fork();
+    if (pid == -1)
+        die("%s", strerror(errno));
+    if (pid == 0)
+    {
+        execvp(name, argv);
+        die("%s", strerror(errno));
+    }
+    waitpid(pid, &pid, 0);
+    int status = WEXITSTATUS(pid);
+    if (status != 0)
+        die("%s process failed with %d\n", name, status);
+}
+
+static void
+remove_tmp_file(void)
+{
+    if (strcmp(tmp_filepath, TMP_FILEPATH_TEMPLATE) != 0)
+        remove(tmp_filepath);
+    if (strcmp(tmp_obj_filepath, TMP_OBJ_FILEPATH_TEMPLATE) != 0)
+        remove(tmp_obj_filepath);
+}
+
+int
+main(int argc, char *argv[])
+{
+    int option;
+
+    while ((option = getopt(argc, argv, "So:b:e:h")) != -1)
+    {
+        switch (option)
+        {
+        case 'S':
+            compile_to_exec = false;
+            break;
+        case 'o':
+            output_filename = optarg;
+            break;
+        case 'b':
+            buffer_len = xatoul(optarg);
+            break;
+        case 'e':
+            buffer_elem_bytes = xatoul(optarg);
+            switch (buffer_elem_bytes)
+            {
+            case 1:
+                elem_bytes_asm_str = "byte";
+                elem_bytes_asm_str_res = "resb";
+                break;
+            case 2:
+                elem_bytes_asm_str = "word";
+                elem_bytes_asm_str_res = "resw";
+                break;
+            case 4:
+                elem_bytes_asm_str = "dword";
+                elem_bytes_asm_str_res = "resd";
+                break;
+            case 8:
+                elem_bytes_asm_str = "qword";
+                elem_bytes_asm_str_res = "resq";
+                break;
+            default:
+                die("%lu: is not a valid number of bytes for buffer element", buffer_elem_bytes);
+            }
+            break;
+        case 'h':
+            puts("Usage: bfc [-h] [-S] [-o file] [-b buffer_len] [-e "
+                 "buffer_elem_bytes] [INPUT_FILE]");
+            puts("");
+            puts("INPUT_FILE  Brainfuck source (read stdin if not present)");
+            puts("");
+            puts("-h          Print this message");
+            puts("-S          Output assembly instead of compiled executable");
+            puts("-o          Output filename ('-' for stdout)");
+            puts("-b          Length of the buffer available the program");
+            puts("-e          Number of bytes in each element of the program's "
+                 "buffer");
+            puts("            (can only be: 1, 2, 4 or 8)");
+            exit(EXIT_SUCCESS);
+        }
+    }
+    if (optind > argc + 1)
+        die("Unexpected argument: %s\n", argv[optind + 1]);
+
+    if (output_filename == NULL)
+        output_filename = compile_to_exec ? "a.out" : "a.asm";
+
+    atexit(remove_tmp_file);
+
     FILE *input_file = NULL;
-    if (argc == 2)
-        input_file = fopen(argv[1], "r");
-    else
+    if (optind == argc)
         input_file = stdin;
+    else
+    {
+        input_file = fopen(argv[optind], "r");
+        if (input_file == NULL)
+            die("%s: %s\n", strerror(errno), argv[optind]);
+    }
+
+    FILE *output_file = NULL;
+    if (strcmp(output_filename, "-") == 0)
+        output_file = stdout;
+    else
+    {
+        if (compile_to_exec)
+            output_file = fdopen(mkstemp(tmp_filepath), "w");
+        else
+            output_file = fopen(output_filename, "w");
+        if (output_file == NULL)
+            die("%s: %s\n", strerror(errno), argv[optind]);
+    }
 
-    /* FILE* output_file = fopen(asm_filename, "w"); */
-    FILE *output_file = stdout;
-    fprintf(
-        output_file,
-        "global _start\n\n"
-        "section .bss\n"
-        "\tbuffer: resb %zu\n\n"
-        "section .text\n"
-        "_start:\n"
-        "\tmov rbx, buffer\n",
-        buffer_size
-    );
+    fprintf(output_file,
+            "global _start\n\n"
+            "section .bss\n"
+            "\tbuffer: resb %zu\n\n"
+            "section .text\n"
+            "_start:\n"
+            "\tmov rbx, buffer\n",
+            buffer_len);
     size_t label_frame_id;
 
     char c;
@@ -82,17 +220,29 @@ int main(int argc, char *argv[])
             break;
         case '[':
             label_frame_id = label_stack[label_stack_frame];
-            fprintf(output_file, "label_open_%03zu_%03zu:          ; [\n", label_stack_frame, label_frame_id);
+            fprintf(output_file,
+                    "label_open_%03zu_%03zu:          ; [\n",
+                    label_stack_frame,
+                    label_frame_id);
             fprintf(output_file, "    cmp byte [rbx], 0        ; [\n");
-            fprintf(output_file, "    je  label_close_%03zu_%03zu  ; [\n", label_stack_frame, label_frame_id);
+            fprintf(output_file,
+                    "    je  label_close_%03zu_%03zu  ; [\n",
+                    label_stack_frame,
+                    label_frame_id);
             label_stack_frame++;
             break;
         case ']':
             label_stack_frame--;
             label_frame_id = label_stack[label_stack_frame];
-            fprintf(output_file, "label_close_%03zu_%03zu:         ; ]\n", label_stack_frame, label_frame_id);
+            fprintf(output_file,
+                    "label_close_%03zu_%03zu:         ; ]\n",
+                    label_stack_frame,
+                    label_frame_id);
             fprintf(output_file, "    cmp byte [rbx], 0        ; ]\n");
-            fprintf(output_file, "    jne label_open_%03zu_%03zu   ; ]\n", label_stack_frame, label_frame_id);
+            fprintf(output_file,
+                    "    jne label_open_%03zu_%03zu   ; ]\n",
+                    label_stack_frame,
+                    label_frame_id);
             label_stack[label_stack_frame]++;
             break;
         case ';':
@@ -103,10 +253,25 @@ int main(int argc, char *argv[])
             break;
         }
     }
+    if (input_file != stdin)
+        fclose(input_file);
     fprintf(output_file, "\n");
     fprintf(output_file, "    mov rdi, 0   ; exit\n");
     fprintf(output_file, "    mov rax, 60  ; exit\n");
     fprintf(output_file, "    syscall      ; exit\n");
+    if (output_file != stdout)
+        fclose(output_file);
+
+    if (compile_to_exec)
+    {
+        strcpy(tmp_obj_filepath, tmp_filepath);
+        strcat(tmp_obj_filepath, ".o");
+        char *const nasm_argv[] = {
+            "nasm", "-f", NASM_FORMAT, "-o", tmp_obj_filepath, tmp_filepath, NULL};
+        run_subprocess("nasm", nasm_argv);
+        char *const ld_argv[] = {"ld", "-o", output_filename, tmp_obj_filepath, NULL};
+        run_subprocess("ld", ld_argv);
+    }
 
     return 0;
 }