Protocol Buffers - Google's data interchange format (grpc依赖) https://developers.google.com/protocol-buffers/
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

481 lines
13 KiB

// Protocol Buffers - Google's data interchange format
// Copyright 2023 Google LLC. All rights reserved.
// https://developers.google.com/protocol-buffers/
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
//
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above
// copyright notice, this list of conditions and the following disclaimer
// in the documentation and/or other materials provided with the
// distribution.
// * Neither the name of Google LLC nor the names of its
// contributors may be used to endorse or promote products derived from
// this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "upb/text/encode.h"
#include <ctype.h>
#include <float.h>
#include <inttypes.h>
#include <stdarg.h>
#include <string.h>
#include "upb/collections/internal/map_sorter.h"
#include "upb/collections/map.h"
#include "upb/lex/round_trip.h"
#include "upb/port/vsnprintf_compat.h"
#include "upb/reflection/message.h"
#include "upb/wire/eps_copy_input_stream.h"
#include "upb/wire/reader.h"
#include "upb/wire/types.h"
// Must be last.
#include "upb/port/def.inc"
typedef struct {
char *buf, *ptr, *end;
size_t overflow;
int indent_depth;
int options;
const upb_DefPool* ext_pool;
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
_upb_mapsorter sorter;
} txtenc;
static void txtenc_msg(txtenc* e, const upb_Message* msg,
const upb_MessageDef* m);
static void txtenc_putbytes(txtenc* e, const void* data, size_t len) {
size_t have = e->end - e->ptr;
if (UPB_LIKELY(have >= len)) {
memcpy(e->ptr, data, len);
e->ptr += len;
} else {
if (have) {
memcpy(e->ptr, data, have);
e->ptr += have;
}
e->overflow += (len - have);
}
}
static void txtenc_putstr(txtenc* e, const char* str) {
txtenc_putbytes(e, str, strlen(str));
}
static void txtenc_printf(txtenc* e, const char* fmt, ...) {
size_t n;
size_t have = e->end - e->ptr;
va_list args;
va_start(args, fmt);
n = _upb_vsnprintf(e->ptr, have, fmt, args);
va_end(args);
if (UPB_LIKELY(have > n)) {
e->ptr += n;
} else {
e->ptr = UPB_PTRADD(e->ptr, have);
e->overflow += (n - have);
}
}
static void txtenc_indent(txtenc* e) {
if ((e->options & UPB_TXTENC_SINGLELINE) == 0) {
int i = e->indent_depth;
while (i-- > 0) {
txtenc_putstr(e, " ");
}
}
}
static void txtenc_endfield(txtenc* e) {
if (e->options & UPB_TXTENC_SINGLELINE) {
txtenc_putstr(e, " ");
} else {
txtenc_putstr(e, "\n");
}
}
static void txtenc_enum(int32_t val, const upb_FieldDef* f, txtenc* e) {
const upb_EnumDef* e_def = upb_FieldDef_EnumSubDef(f);
const upb_EnumValueDef* ev = upb_EnumDef_FindValueByNumber(e_def, val);
if (ev) {
txtenc_printf(e, "%s", upb_EnumValueDef_Name(ev));
} else {
txtenc_printf(e, "%" PRId32, val);
}
}
static void txtenc_string(txtenc* e, upb_StringView str, bool bytes) {
const char* ptr = str.data;
const char* end = ptr + str.size;
txtenc_putstr(e, "\"");
while (ptr < end) {
switch (*ptr) {
case '\n':
txtenc_putstr(e, "\\n");
break;
case '\r':
txtenc_putstr(e, "\\r");
break;
case '\t':
txtenc_putstr(e, "\\t");
break;
case '\"':
txtenc_putstr(e, "\\\"");
break;
case '\'':
txtenc_putstr(e, "\\'");
break;
case '\\':
txtenc_putstr(e, "\\\\");
break;
default:
if ((bytes || (uint8_t)*ptr < 0x80) && !isprint(*ptr)) {
txtenc_printf(e, "\\%03o", (int)(uint8_t)*ptr);
} else {
txtenc_putbytes(e, ptr, 1);
}
5 years ago
break;
}
ptr++;
}
txtenc_putstr(e, "\"");
}
static void txtenc_field(txtenc* e, upb_MessageValue val,
const upb_FieldDef* f) {
txtenc_indent(e);
const upb_CType type = upb_FieldDef_CType(f);
const bool is_ext = upb_FieldDef_IsExtension(f);
const char* full = upb_FieldDef_FullName(f);
const char* name = upb_FieldDef_Name(f);
if (type == kUpb_CType_Message) {
if (is_ext) {
txtenc_printf(e, "[%s] {", full);
} else {
txtenc_printf(e, "%s {", name);
}
txtenc_endfield(e);
e->indent_depth++;
txtenc_msg(e, val.msg_val, upb_FieldDef_MessageSubDef(f));
e->indent_depth--;
txtenc_indent(e);
txtenc_putstr(e, "}");
txtenc_endfield(e);
return;
}
if (is_ext) {
txtenc_printf(e, "[%s]: ", full);
} else {
txtenc_printf(e, "%s: ", name);
}
switch (type) {
case kUpb_CType_Bool:
txtenc_putstr(e, val.bool_val ? "true" : "false");
break;
case kUpb_CType_Float: {
char buf[32];
_upb_EncodeRoundTripFloat(val.float_val, buf, sizeof(buf));
txtenc_putstr(e, buf);
break;
}
case kUpb_CType_Double: {
char buf[32];
_upb_EncodeRoundTripDouble(val.double_val, buf, sizeof(buf));
txtenc_putstr(e, buf);
break;
}
case kUpb_CType_Int32:
txtenc_printf(e, "%" PRId32, val.int32_val);
break;
case kUpb_CType_UInt32:
txtenc_printf(e, "%" PRIu32, val.uint32_val);
break;
case kUpb_CType_Int64:
txtenc_printf(e, "%" PRId64, val.int64_val);
break;
case kUpb_CType_UInt64:
txtenc_printf(e, "%" PRIu64, val.uint64_val);
break;
case kUpb_CType_String:
txtenc_string(e, val.str_val, false);
break;
case kUpb_CType_Bytes:
txtenc_string(e, val.str_val, true);
break;
case kUpb_CType_Enum:
txtenc_enum(val.int32_val, f, e);
break;
default:
UPB_UNREACHABLE();
}
txtenc_endfield(e);
}
/*
* Arrays print as simple repeated elements, eg.
*
* foo_field: 1
* foo_field: 2
* foo_field: 3
*/
static void txtenc_array(txtenc* e, const upb_Array* arr,
const upb_FieldDef* f) {
size_t i;
size_t size = upb_Array_Size(arr);
for (i = 0; i < size; i++) {
txtenc_field(e, upb_Array_Get(arr, i), f);
}
}
static void txtenc_mapentry(txtenc* e, upb_MessageValue key,
upb_MessageValue val, const upb_FieldDef* f) {
const upb_MessageDef* entry = upb_FieldDef_MessageSubDef(f);
const upb_FieldDef* key_f = upb_MessageDef_Field(entry, 0);
const upb_FieldDef* val_f = upb_MessageDef_Field(entry, 1);
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
txtenc_indent(e);
txtenc_printf(e, "%s {", upb_FieldDef_Name(f));
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
txtenc_endfield(e);
e->indent_depth++;
txtenc_field(e, key, key_f);
txtenc_field(e, val, val_f);
e->indent_depth--;
txtenc_indent(e);
txtenc_putstr(e, "}");
txtenc_endfield(e);
}
/*
* Maps print as messages of key/value, etc.
*
* foo_map: {
* key: "abc"
* value: 123
* }
* foo_map: {
* key: "def"
* value: 456
* }
*/
static void txtenc_map(txtenc* e, const upb_Map* map, const upb_FieldDef* f) {
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
if (e->options & UPB_TXTENC_NOSORT) {
size_t iter = kUpb_Map_Begin;
upb_MessageValue key, val;
while (upb_Map_Next(map, &key, &val, &iter)) {
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
txtenc_mapentry(e, key, val, f);
}
} else {
const upb_MessageDef* entry = upb_FieldDef_MessageSubDef(f);
const upb_FieldDef* key_f = upb_MessageDef_Field(entry, 0);
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
_upb_sortedmap sorted;
upb_MapEntry ent;
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
_upb_mapsorter_pushmap(&e->sorter, upb_FieldDef_Type(key_f), map, &sorted);
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
while (_upb_sortedmap_next(&e->sorter, map, &sorted, &ent)) {
upb_MessageValue key, val;
memcpy(&key, &ent.data.k, sizeof(key));
memcpy(&val, &ent.data.v, sizeof(val));
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
txtenc_mapentry(e, key, val, f);
}
_upb_mapsorter_popmap(&e->sorter, &sorted);
}
}
#define CHK(x) \
do { \
if (!(x)) { \
return false; \
} \
} while (0)
/*
* Unknown fields are printed by number.
*
* 1001: 123
* 1002: "hello"
* 1006: 0xdeadbeef
* 1003: {
* 1: 111
* }
*/
static const char* txtenc_unknown(txtenc* e, const char* ptr,
upb_EpsCopyInputStream* stream,
int groupnum) {
// We are guaranteed that the unknown data is valid wire format, and will not
// contain tag zero.
uint32_t end_group = groupnum > 0
? ((groupnum << kUpb_WireReader_WireTypeBits) |
kUpb_WireType_EndGroup)
: 0;
while (!upb_EpsCopyInputStream_IsDone(stream, &ptr)) {
uint32_t tag;
CHK(ptr = upb_WireReader_ReadTag(ptr, &tag));
if (tag == end_group) return ptr;
txtenc_indent(e);
txtenc_printf(e, "%d: ", (int)upb_WireReader_GetFieldNumber(tag));
switch (upb_WireReader_GetWireType(tag)) {
case kUpb_WireType_Varint: {
uint64_t val;
CHK(ptr = upb_WireReader_ReadVarint(ptr, &val));
txtenc_printf(e, "%" PRIu64, val);
break;
}
case kUpb_WireType_32Bit: {
uint32_t val;
ptr = upb_WireReader_ReadFixed32(ptr, &val);
txtenc_printf(e, "0x%08" PRIu32, val);
break;
}
case kUpb_WireType_64Bit: {
uint64_t val;
ptr = upb_WireReader_ReadFixed64(ptr, &val);
txtenc_printf(e, "0x%016" PRIu64, val);
break;
}
case kUpb_WireType_Delimited: {
int size;
char* start = e->ptr;
size_t start_overflow = e->overflow;
CHK(ptr = upb_WireReader_ReadSize(ptr, &size));
CHK(upb_EpsCopyInputStream_CheckDataSizeAvailable(stream, ptr, size));
// Speculatively try to parse as message.
txtenc_putstr(e, "{");
txtenc_endfield(e);
// EpsCopyInputStream can't back up, so create a sub-stream for the
// speculative parse.
upb_EpsCopyInputStream sub_stream;
const char* sub_ptr = upb_EpsCopyInputStream_GetAliasedPtr(stream, ptr);
upb_EpsCopyInputStream_Init(&sub_stream, &sub_ptr, size, true);
e->indent_depth++;
if (txtenc_unknown(e, sub_ptr, &sub_stream, -1)) {
ptr = upb_EpsCopyInputStream_Skip(stream, ptr, size);
e->indent_depth--;
txtenc_indent(e);
txtenc_putstr(e, "}");
} else {
// Didn't work out, print as raw bytes.
e->indent_depth--;
e->ptr = start;
e->overflow = start_overflow;
const char* str = ptr;
ptr = upb_EpsCopyInputStream_ReadString(stream, &str, size, NULL);
assert(ptr);
txtenc_string(e, (upb_StringView){.data = str, .size = size}, true);
}
break;
}
case kUpb_WireType_StartGroup:
txtenc_putstr(e, "{");
txtenc_endfield(e);
e->indent_depth++;
CHK(ptr = txtenc_unknown(e, ptr, stream,
upb_WireReader_GetFieldNumber(tag)));
e->indent_depth--;
txtenc_indent(e);
txtenc_putstr(e, "}");
break;
default:
return NULL;
}
txtenc_endfield(e);
}
return end_group == 0 && !upb_EpsCopyInputStream_IsError(stream) ? ptr : NULL;
}
#undef CHK
static void txtenc_msg(txtenc* e, const upb_Message* msg,
const upb_MessageDef* m) {
size_t iter = kUpb_Message_Begin;
const upb_FieldDef* f;
upb_MessageValue val;
while (upb_Message_Next(msg, m, e->ext_pool, &f, &val, &iter)) {
if (upb_FieldDef_IsMap(f)) {
txtenc_map(e, val.map_val, f);
} else if (upb_FieldDef_IsRepeated(f)) {
txtenc_array(e, val.array_val, f);
} else {
txtenc_field(e, val, f);
}
}
if ((e->options & UPB_TXTENC_SKIPUNKNOWN) == 0) {
size_t size;
const char* ptr = upb_Message_GetUnknown(msg, &size);
if (size != 0) {
char* start = e->ptr;
upb_EpsCopyInputStream stream;
upb_EpsCopyInputStream_Init(&stream, &ptr, size, true);
if (!txtenc_unknown(e, ptr, &stream, -1)) {
/* Unknown failed to parse, back up and don't print it at all. */
e->ptr = start;
}
}
}
}
size_t txtenc_nullz(txtenc* e, size_t size) {
size_t ret = e->ptr - e->buf + e->overflow;
if (size > 0) {
if (e->ptr == e->end) e->ptr--;
*e->ptr = '\0';
}
return ret;
}
3 years ago
size_t upb_TextEncode(const upb_Message* msg, const upb_MessageDef* m,
const upb_DefPool* ext_pool, int options, char* buf,
size_t size) {
txtenc e;
e.buf = buf;
e.ptr = buf;
e.end = UPB_PTRADD(buf, size);
e.overflow = 0;
e.indent_depth = 0;
e.options = options;
e.ext_pool = ext_pool;
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
_upb_mapsorter_init(&e.sorter);
txtenc_msg(&e, msg, m);
Added map sorting to binary and text encoders. For the binary encoder, sorting is off by default. For the text encoder, sorting is on by default. Both defaults can be explicitly overridden. This grows code size a bit. I think we could potentially shave this (and other map-related code size) by having the generated code inject a function pointer to the map-related parsing/serialization code if maps are present. FILE SIZE VM SIZE -------------- -------------- +86% +1.07Ki +71% +768 upb/msg.c [NEW] +391 [NEW] +344 _upb_mapsorter_pushmap [NEW] +158 [NEW] +112 _upb_mapsorter_cmpstr [NEW] +111 [NEW] +64 _upb_mapsorter_cmpbool [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpi64 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu32 [NEW] +110 [NEW] +64 _upb_mapsorter_cmpu64 -3.6% -8 -4.3% -8 _upb_map_new +9.5% +464 +9.2% +424 upb/text_encode.c [NEW] +656 [NEW] +616 txtenc_mapentry +15% +32 +20% +32 upb_text_encode -20.1% -224 -20.7% -224 txtenc_msg +5.7% +342 +5.3% +296 upb/encode.c [NEW] +344 [NEW] +304 encode_mapentry [NEW] +246 [NEW] +208 upb_encode_ex [NEW] +41 [NEW] +16 upb_encode_ex.ch +0.7% +8 +0.7% +8 encode_scalar -1.0% -32 -1.0% -32 encode_message [DEL] -38 [DEL] -16 upb_encode.ch [DEL] -227 [DEL] -192 upb_encode +2.0% +152 +2.2% +152 upb/decode.c +44% +128 +44% +128 [section .rodata] +3.4% +24 +3.4% +24 _GLOBAL_OFFSET_TABLE_ +0.6% +107 +0.3% +48 upb/def.c [NEW] +100 [NEW] +48 upb_fielddef_descriptortype +7.1% +7 [ = ] 0 upb_fielddef_defaultint32 +2.9% +24 +2.9% +24 [section .dynsym] +1.2% +24 [ = ] 0 [section .symtab] +3.2% +16 +3.2% +16 [section .plt] [NEW] +16 [NEW] +16 memcmp@plt +0.5% +16 +0.6% +16 tests/conformance_upb.c +1.5% +16 +1.6% +16 DoTestIo +0.1% +16 +0.1% +16 upb/json_decode.c +0.4% +16 +0.4% +16 jsondec_wellknown +3.0% +8 +3.0% +8 [section .got.plt] +3.0% +8 +3.0% +8 _GLOBAL_OFFSET_TABLE_ +1.6% +7 +1.6% +7 [section .dynstr] +1.8% +4 +1.8% +4 [section .hash] +0.5% +3 +0.5% +3 [LOAD #2 [RX]] +2.8% +2 +2.8% +2 [section .gnu.version] -60.0% -1.74Ki [ = ] 0 [Unmapped] +0.3% +496 +1.4% +1.74Ki TOTAL
4 years ago
_upb_mapsorter_destroy(&e.sorter);
return txtenc_nullz(&e, size);
}