Merge pull request #114 from pfalcon/streams-bootstrap

Define buffer and stream protocols, and other starting bits of io.* framework, with io.FileIO-like implementation for Unix

py/obj.h

@@ -84,6 +84,41 @@ typedef struct _mp_method_t {
     mp_const_obj_t fun;
 } mp_method_t;
 
+// Buffer protocol
+typedef struct _buffer_info_t {
+    // if we'd bother to support various versions of structure
+    // (with different number of fields), we can distinguish
+    // them with ver = sizeof(struct). Cons: overkill for *micro*?
+    //int ver; // ?
+
+    void *buf;
+    machine_int_t len;
+
+    // Rationale: have array.array and have SIMD operations on them
+    // Cons: users can pass item size to processing functions themselves,
+    // though that's not "plug&play"
+    // int itemsize;
+
+    // Rationale: to load arbitrary-sized sprites directly to LCD
+    // Cons: a bit adhoc usecase
+    // int stride;
+} buffer_info_t;
+#define BUFFER_READ  (1)
+#define BUFFER_WRITE (2)
+#define BUFFER_RW (BUFFER_READ | BUFFER_WRITE)
+typedef struct _mp_buffer_p_t {
+    machine_int_t (*get_buffer)(mp_obj_t obj, buffer_info_t *bufinfo, int flags);
+} mp_buffer_p_t;
+
+// Stream protocol
+typedef struct _mp_stream_p_t {
+    // On error, functions should return -1 and fill in *errcode (values are
+    // implementation-dependent, but will be exposed to user, e.g. via exception).
+    machine_int_t (*read)(mp_obj_t obj, void *buf, machine_uint_t size, int *errcode);
+    machine_int_t (*write)(mp_obj_t obj, const void *buf, machine_uint_t size, int *errcode);
+    // add seek() ?
+} mp_stream_p_t;
+
 struct _mp_obj_type_t {
     mp_obj_base_t base;
     const char *name;
@@ -98,6 +133,12 @@ struct _mp_obj_type_t {
     mp_fun_1_t getiter;
     mp_fun_1_t iternext;
 
+    // Alternatively, pointer(s) to interfaces to save space
+    // in mp_obj_type_t at the expense of extra pointer and extra dereference
+    // when actually used.
+    mp_buffer_p_t buffer_p;
+    mp_stream_p_t stream_p;
+
     const mp_method_t *methods;
 
     /*

py/py.mk

@@ -68,6 +68,7 @@ PY_O_BASENAME = \
 	objstr.o \
 	objtuple.o \
 	objtype.o \
+	stream.o \
 	builtin.o \
 	builtinimport.o \
 	vm.o \

py/stream.c

@@ -0,0 +1,55 @@
+#include <string.h>
+
+#include "nlr.h"
+#include "misc.h"
+#include "mpconfig.h"
+#include "mpqstr.h"
+#include "obj.h"
+#include "stream.h"
+
+// This file defines generic Python stream read/write methods which
+// dispatch to the underlying stream interface of an object.
+
+static mp_obj_t stream_read(mp_obj_t self_in, mp_obj_t arg) {
+    struct _mp_obj_base_t *o = (struct _mp_obj_base_t *)self_in;
+    if (o->type->stream_p.read == NULL) {
+        // CPython: io.UnsupportedOperation, OSError subclass
+        nlr_jump(mp_obj_new_exception_msg(MP_QSTR_OSError, "Operation not supported"));
+    }
+
+    machine_int_t sz = mp_obj_get_int(arg);
+    // +1 because so far we mark end of string with \0
+    char *buf = m_new(char, sz + 1);
+    int error;
+    machine_int_t out_sz = o->type->stream_p.read(self_in, buf, sz, &error);
+    if (out_sz == -1) {
+        nlr_jump(mp_obj_new_exception_msg_1_arg(MP_QSTR_OSError, "[Errno %d]", (const char *)error));
+    } else {
+        buf[out_sz] = 0;
+        return mp_obj_new_str(qstr_from_str_take(buf, /*out_sz,*/ sz + 1));
+    }
+}
+
+static mp_obj_t stream_write(mp_obj_t self_in, mp_obj_t arg) {
+    struct _mp_obj_base_t *o = (struct _mp_obj_base_t *)self_in;
+    if (o->type->stream_p.write == NULL) {
+        // CPython: io.UnsupportedOperation, OSError subclass
+        nlr_jump(mp_obj_new_exception_msg(MP_QSTR_OSError, "Operation not supported"));
+    }
+
+    const char *buf = qstr_str(mp_obj_get_qstr(arg));
+    machine_int_t sz = strlen(buf);
+    int error;
+    machine_int_t out_sz = o->type->stream_p.write(self_in, buf, sz, &error);
+    if (out_sz == -1) {
+        nlr_jump(mp_obj_new_exception_msg_1_arg(MP_QSTR_OSError, "[Errno %d]", (const char *)error));
+    } else {
+        // http://docs.python.org/3/library/io.html#io.RawIOBase.write
+        // "None is returned if the raw stream is set not to block and no single byte could be readily written to it."
+        // Do they mean that instead of 0 they return None?
+        return MP_OBJ_NEW_SMALL_INT(out_sz);
+    }
+}
+
+MP_DEFINE_CONST_FUN_OBJ_2(mp_stream_read_obj, stream_read);
+MP_DEFINE_CONST_FUN_OBJ_2(mp_stream_write_obj, stream_write);

py/stream.h

@@ -0,0 +1,2 @@
+extern const mp_obj_fun_native_t mp_stream_read_obj;
+extern const mp_obj_fun_native_t mp_stream_write_obj;

unix/Makefile

@@ -16,6 +16,7 @@ LDFLAGS = -lm
 # source files
 SRC_C = \
 	main.c \
+	file.c \
 
 OBJ = $(addprefix $(BUILD)/, $(SRC_C:.c=.o)) $(PY_O)
 LIB = -lreadline

unix/file.c

@@ -0,0 +1,117 @@
+#include <string.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <errno.h>
+
+#include "nlr.h"
+#include "misc.h"
+#include "mpconfig.h"
+#include "mpqstr.h"
+#include "obj.h"
+#include "stream.h"
+
+typedef struct _mp_obj_fdfile_t {
+    mp_obj_base_t base;
+    int fd;
+} mp_obj_fdfile_t;
+
+static void fdfile_print(void (*print)(void *env, const char *fmt, ...), void *env, mp_obj_t self_in) {
+    mp_obj_fdfile_t *self = self_in;
+    print(env, "<io.FileIO %d>", self->fd);
+}
+
+static machine_int_t fdfile_read(mp_obj_t o_in, void *buf, machine_uint_t size, int *errcode) {
+    mp_obj_fdfile_t *o = o_in;
+    machine_int_t r = read(o->fd, buf, size);
+    if (r == -1) {
+        *errcode = errno;
+    }
+    return r;
+}
+
+static machine_int_t fdfile_write(mp_obj_t o_in, const void *buf, machine_uint_t size, int *errcode) {
+    mp_obj_fdfile_t *o = o_in;
+    machine_int_t r = write(o->fd, buf, size);
+    if (r == -1) {
+        *errcode = errno;
+    }
+    return r;
+}
+
+static mp_obj_t fdfile_close(mp_obj_t self_in) {
+    mp_obj_fdfile_t *self = self_in;
+    close(self->fd);
+    return mp_const_none;
+}
+static MP_DEFINE_CONST_FUN_OBJ_1(fdfile_close_obj, fdfile_close);
+
+static mp_obj_t fdfile_make_new(mp_obj_t type_in, int n_args, const mp_obj_t *args) {
+    mp_obj_fdfile_t *o = m_new_obj(mp_obj_fdfile_t);
+    o->base.type = type_in;
+
+    if (MP_OBJ_IS_SMALL_INT(args[0])) {
+        o->fd = MP_OBJ_SMALL_INT_VALUE(args[0]);
+        return o;
+    }
+
+    const char *fname = qstr_str(mp_obj_get_qstr(args[0]));
+    const char *mode_s;
+    if (n_args > 1) {
+        mode_s = qstr_str(mp_obj_get_qstr(args[1]));
+    } else {
+        mode_s = "r";
+    }
+
+    int mode = 0;
+    while (*mode_s) {
+        switch (*mode_s++) {
+            // Note: these assume O_RDWR = O_RDONLY | O_WRONLY
+            case 'r':
+                mode |= O_RDONLY;
+                break;
+            case 'w':
+                mode |= O_WRONLY | O_CREAT | O_TRUNC;
+                break;
+            case 'a':
+                mode |= O_APPEND;
+                break;
+            case '+':
+                mode |= O_RDWR;
+                break;
+        }
+    }
+
+    o->fd = open(fname, mode, 0644);
+    if (o->fd == -1) {
+        nlr_jump(mp_obj_new_exception_msg_1_arg(MP_QSTR_OSError, "[Errno %d]", (const char *)errno));
+    }
+    return o;
+}
+
+static const mp_method_t rawfile_type_methods[] = {
+        { "read", &mp_stream_read_obj },
+        { "write", &mp_stream_write_obj },
+        { "close", &fdfile_close_obj },
+        { NULL, NULL },
+};
+
+static const mp_obj_type_t rawfile_type = {
+    { &mp_const_type },
+    "io.FileIO",
+    .print = fdfile_print,
+    .make_new = fdfile_make_new,
+    .getiter = NULL,
+    .iternext = NULL,
+    .stream_p = {
+        .read = fdfile_read,
+        .write = fdfile_write,
+    },
+    .methods = rawfile_type_methods,
+};
+
+// Factory function for I/O stream classes
+mp_obj_t mp_builtin_open(int n_args, const mp_obj_t *args) {
+    // TODO: analyze mode and buffering args and instantiate appropriate type
+    return fdfile_make_new((mp_obj_t)&rawfile_type, n_args, args);
+}
+MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(mp_builtin_open_obj, 1, 2, mp_builtin_open);

unix/main.c

@@ -20,6 +20,8 @@
 #include <readline/history.h>
 #endif
 
+extern const mp_obj_fun_native_t mp_builtin_open_obj;
+
 static void execute_from_lexer(mp_lexer_t *lex, mp_parse_input_kind_t input_kind, bool is_repl) {
     if (lex == NULL) {
         return;
@@ -206,6 +208,7 @@ int main(int argc, char **argv) {
     rt_init();
 
     rt_store_name(qstr_from_str_static("test"), test_obj_new(42));
+    rt_store_name(qstr_from_str_static("open"), (mp_obj_t)&mp_builtin_open_obj);
 
     /*
     printf("bytes:\n");