|
@@ -1,11 +1,5 @@
|
|
// Amalgamated source file
|
|
// Amalgamated source file
|
|
#include "upb.h"
|
|
#include "upb.h"
|
|
-/*
|
|
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2008-2012 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- */
|
|
|
|
|
|
|
|
|
|
|
|
#include <stdlib.h>
|
|
#include <stdlib.h>
|
|
@@ -1701,12 +1695,6 @@ upb_fielddef *upb_oneof_iter_field(const upb_oneof_iter *iter) {
|
|
void upb_oneof_iter_setdone(upb_oneof_iter *iter) {
|
|
void upb_oneof_iter_setdone(upb_oneof_iter *iter) {
|
|
upb_inttable_iter_setdone(iter);
|
|
upb_inttable_iter_setdone(iter);
|
|
}
|
|
}
|
|
-/*
|
|
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2014 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- */
|
|
|
|
|
|
|
|
|
|
|
|
#include <stdlib.h>
|
|
#include <stdlib.h>
|
|
@@ -1980,14 +1968,9 @@ upb_alloc_func *upb_seededalloc_getallocfunc(upb_seededalloc *a) {
|
|
return seeded_alloc;
|
|
return seeded_alloc;
|
|
}
|
|
}
|
|
/*
|
|
/*
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2011-2012 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- *
|
|
|
|
- * TODO(haberman): it's unclear whether a lot of the consistency checks should
|
|
|
|
- * assert() or return false.
|
|
|
|
- */
|
|
|
|
|
|
+** TODO(haberman): it's unclear whether a lot of the consistency checks should
|
|
|
|
+** assert() or return false.
|
|
|
|
+*/
|
|
|
|
|
|
|
|
|
|
#include <stdlib.h>
|
|
#include <stdlib.h>
|
|
@@ -2668,24 +2651,21 @@ bool upb_byteshandler_setendstr(upb_byteshandler *h,
|
|
return true;
|
|
return true;
|
|
}
|
|
}
|
|
/*
|
|
/*
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2012 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- *
|
|
|
|
- * Our key invariants are:
|
|
|
|
- * 1. reference cycles never span groups
|
|
|
|
- * 2. for ref2(to, from), we increment to's count iff group(from) != group(to)
|
|
|
|
- *
|
|
|
|
- * The previous two are how we avoid leaking cycles. Other important
|
|
|
|
- * invariants are:
|
|
|
|
- * 3. for mutable objects "from" and "to", if there exists a ref2(to, from)
|
|
|
|
- * this implies group(from) == group(to). (In practice, what we implement
|
|
|
|
- * is even stronger; "from" and "to" will share a group if there has *ever*
|
|
|
|
- * been a ref2(to, from), but all that is necessary for correctness is the
|
|
|
|
- * weaker one).
|
|
|
|
- * 4. mutable and immutable objects are never in the same group.
|
|
|
|
- */
|
|
|
|
|
|
+** upb::RefCounted Implementation
|
|
|
|
+**
|
|
|
|
+** Our key invariants are:
|
|
|
|
+** 1. reference cycles never span groups
|
|
|
|
+** 2. for ref2(to, from), we increment to's count iff group(from) != group(to)
|
|
|
|
+**
|
|
|
|
+** The previous two are how we avoid leaking cycles. Other important
|
|
|
|
+** invariants are:
|
|
|
|
+** 3. for mutable objects "from" and "to", if there exists a ref2(to, from)
|
|
|
|
+** this implies group(from) == group(to). (In practice, what we implement
|
|
|
|
+** is even stronger; "from" and "to" will share a group if there has *ever*
|
|
|
|
+** been a ref2(to, from), but all that is necessary for correctness is the
|
|
|
|
+** weaker one).
|
|
|
|
+** 4. mutable and immutable objects are never in the same group.
|
|
|
|
+*/
|
|
|
|
|
|
|
|
|
|
#include <setjmp.h>
|
|
#include <setjmp.h>
|
|
@@ -3514,12 +3494,6 @@ bool upb_refcounted_freeze(upb_refcounted *const*roots, int n, upb_status *s,
|
|
}
|
|
}
|
|
return freeze(roots, n, s, maxdepth);
|
|
return freeze(roots, n, s, maxdepth);
|
|
}
|
|
}
|
|
-/*
|
|
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2013 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- */
|
|
|
|
|
|
|
|
|
|
|
|
#include <stdlib.h>
|
|
#include <stdlib.h>
|
|
@@ -3605,12 +3579,6 @@ const upb_shim_data *upb_shim_getdata(const upb_handlers *h, upb_selector_t s,
|
|
|
|
|
|
return (const upb_shim_data*)upb_handlers_gethandlerdata(h, s);
|
|
return (const upb_shim_data*)upb_handlers_gethandlerdata(h, s);
|
|
}
|
|
}
|
|
-/*
|
|
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2008-2012 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- */
|
|
|
|
|
|
|
|
|
|
|
|
#include <stdlib.h>
|
|
#include <stdlib.h>
|
|
@@ -4041,13 +4009,10 @@ const upb_def *upb_symtab_iter_def(const upb_symtab_iter *iter) {
|
|
return upb_value_getptr(upb_strtable_iter_value(&iter->iter));
|
|
return upb_value_getptr(upb_strtable_iter_value(&iter->iter));
|
|
}
|
|
}
|
|
/*
|
|
/*
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2009 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- *
|
|
|
|
- * Implementation is heavily inspired by Lua's ltable.c.
|
|
|
|
- */
|
|
|
|
|
|
+** upb_table Implementation
|
|
|
|
+**
|
|
|
|
+** Implementation is heavily inspired by Lua's ltable.c.
|
|
|
|
+*/
|
|
|
|
|
|
|
|
|
|
#include <stdlib.h>
|
|
#include <stdlib.h>
|
|
@@ -4931,12 +4896,6 @@ uint32_t MurmurHash2(const void * key, size_t len, uint32_t seed) {
|
|
#undef MIX
|
|
#undef MIX
|
|
|
|
|
|
#endif /* UPB_UNALIGNED_READS_OK */
|
|
#endif /* UPB_UNALIGNED_READS_OK */
|
|
-/*
|
|
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2009-2012 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- */
|
|
|
|
|
|
|
|
#include <errno.h>
|
|
#include <errno.h>
|
|
#include <stdarg.h>
|
|
#include <stdarg.h>
|
|
@@ -5860,17 +5819,12 @@ static upb_inttable reftables[212] = {
|
|
#endif
|
|
#endif
|
|
|
|
|
|
/*
|
|
/*
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2008-2009 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- *
|
|
|
|
- * XXX: The routines in this file that consume a string do not currently
|
|
|
|
- * support having the string span buffers. In the future, as upb_sink and
|
|
|
|
- * its buffering/sharing functionality evolve there should be an easy and
|
|
|
|
- * idiomatic way of correctly handling this case. For now, we accept this
|
|
|
|
- * limitation since we currently only parse descriptors from single strings.
|
|
|
|
- */
|
|
|
|
|
|
+** XXX: The routines in this file that consume a string do not currently
|
|
|
|
+** support having the string span buffers. In the future, as upb_sink and
|
|
|
|
+** its buffering/sharing functionality evolve there should be an easy and
|
|
|
|
+** idiomatic way of correctly handling this case. For now, we accept this
|
|
|
|
+** limitation since we currently only parse descriptors from single strings.
|
|
|
|
+*/
|
|
|
|
|
|
|
|
|
|
#include <errno.h>
|
|
#include <errno.h>
|
|
@@ -6518,21 +6472,18 @@ const upb_handlers *upb_descreader_newhandlers(const void *owner) {
|
|
return h;
|
|
return h;
|
|
}
|
|
}
|
|
/*
|
|
/*
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2013 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- *
|
|
|
|
- * Code to compile a upb::Handlers into bytecode for decoding a protobuf
|
|
|
|
- * according to that specific schema and destination handlers.
|
|
|
|
- *
|
|
|
|
- * Compiling to bytecode is always the first step. If we are using the
|
|
|
|
- * interpreted decoder we leave it as bytecode and interpret that. If we are
|
|
|
|
- * using a JIT decoder we use a code generator to turn the bytecode into native
|
|
|
|
- * code, LLVM IR, etc.
|
|
|
|
- *
|
|
|
|
- * Bytecode definition is in decoder.int.h.
|
|
|
|
- */
|
|
|
|
|
|
+** protobuf decoder bytecode compiler
|
|
|
|
+**
|
|
|
|
+** Code to compile a upb::Handlers into bytecode for decoding a protobuf
|
|
|
|
+** according to that specific schema and destination handlers.
|
|
|
|
+**
|
|
|
|
+** Compiling to bytecode is always the first step. If we are using the
|
|
|
|
+** interpreted decoder we leave it as bytecode and interpret that. If we are
|
|
|
|
+** using a JIT decoder we use a code generator to turn the bytecode into native
|
|
|
|
+** code, LLVM IR, etc.
|
|
|
|
+**
|
|
|
|
+** Bytecode definition is in decoder.int.h.
|
|
|
|
+*/
|
|
|
|
|
|
#include <stdarg.h>
|
|
#include <stdarg.h>
|
|
|
|
|
|
@@ -7502,24 +7453,19 @@ void upb_pbdecodermethodopts_setlazy(upb_pbdecodermethodopts *opts, bool lazy) {
|
|
opts->lazy = lazy;
|
|
opts->lazy = lazy;
|
|
}
|
|
}
|
|
/*
|
|
/*
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2008-2013 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- *
|
|
|
|
- * This file implements a VM for the interpreted (bytecode) decoder.
|
|
|
|
- *
|
|
|
|
- * Bytecode must previously have been generated using the bytecode compiler in
|
|
|
|
- * compile_decoder.c. This decoder then walks through the bytecode op-by-op to
|
|
|
|
- * parse the input.
|
|
|
|
- *
|
|
|
|
- * Decoding is fully resumable; we just keep a pointer to the current bytecode
|
|
|
|
- * instruction and resume from there. A fair amount of the logic here is to
|
|
|
|
- * handle the fact that values can span buffer seams and we have to be able to
|
|
|
|
- * be capable of suspending/resuming from any byte in the stream. This
|
|
|
|
- * sometimes requires keeping a few trailing bytes from the last buffer around
|
|
|
|
- * in the "residual" buffer.
|
|
|
|
- */
|
|
|
|
|
|
+** upb::Decoder (Bytecode Decoder VM)
|
|
|
|
+**
|
|
|
|
+** Bytecode must previously have been generated using the bytecode compiler in
|
|
|
|
+** compile_decoder.c. This decoder then walks through the bytecode op-by-op to
|
|
|
|
+** parse the input.
|
|
|
|
+**
|
|
|
|
+** Decoding is fully resumable; we just keep a pointer to the current bytecode
|
|
|
|
+** instruction and resume from there. A fair amount of the logic here is to
|
|
|
|
+** handle the fact that values can span buffer seams and we have to be able to
|
|
|
|
+** be capable of suspending/resuming from any byte in the stream. This
|
|
|
|
+** sometimes requires keeping a few trailing bytes from the last buffer around
|
|
|
|
+** in the "residual" buffer.
|
|
|
|
+*/
|
|
|
|
|
|
#include <inttypes.h>
|
|
#include <inttypes.h>
|
|
#include <stddef.h>
|
|
#include <stddef.h>
|
|
@@ -8529,63 +8475,60 @@ bool upb_pbdecoder_setmaxnesting(upb_pbdecoder *d, size_t max) {
|
|
return true;
|
|
return true;
|
|
}
|
|
}
|
|
/*
|
|
/*
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2014 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- *
|
|
|
|
- * Since we are implementing pure handlers (ie. without any out-of-band access
|
|
|
|
- * to pre-computed lengths), we have to buffer all submessages before we can
|
|
|
|
- * emit even their first byte.
|
|
|
|
- *
|
|
|
|
- * Not knowing the size of submessages also means we can't write a perfect
|
|
|
|
- * zero-copy implementation, even with buffering. Lengths are stored as
|
|
|
|
- * varints, which means that we don't know how many bytes to reserve for the
|
|
|
|
- * length until we know what the length is.
|
|
|
|
- *
|
|
|
|
- * This leaves us with three main choices:
|
|
|
|
- *
|
|
|
|
- * 1. buffer all submessage data in a temporary buffer, then copy it exactly
|
|
|
|
- * once into the output buffer.
|
|
|
|
- *
|
|
|
|
- * 2. attempt to buffer data directly into the output buffer, estimating how
|
|
|
|
- * many bytes each length will take. When our guesses are wrong, use
|
|
|
|
- * memmove() to grow or shrink the allotted space.
|
|
|
|
- *
|
|
|
|
- * 3. buffer directly into the output buffer, allocating a max length
|
|
|
|
- * ahead-of-time for each submessage length. If we overallocated, we waste
|
|
|
|
- * space, but no memcpy() or memmove() is required. This approach requires
|
|
|
|
- * defining a maximum size for submessages and rejecting submessages that
|
|
|
|
- * exceed that size.
|
|
|
|
- *
|
|
|
|
- * (2) and (3) have the potential to have better performance, but they are more
|
|
|
|
- * complicated and subtle to implement:
|
|
|
|
- *
|
|
|
|
- * (3) requires making an arbitrary choice of the maximum message size; it
|
|
|
|
- * wastes space when submessages are shorter than this and fails
|
|
|
|
- * completely when they are longer. This makes it more finicky and
|
|
|
|
- * requires configuration based on the input. It also makes it impossible
|
|
|
|
- * to perfectly match the output of reference encoders that always use the
|
|
|
|
- * optimal amount of space for each length.
|
|
|
|
- *
|
|
|
|
- * (2) requires guessing the the size upfront, and if multiple lengths are
|
|
|
|
- * guessed wrong the minimum required number of memmove() operations may
|
|
|
|
- * be complicated to compute correctly. Implemented properly, it may have
|
|
|
|
- * a useful amortized or average cost, but more investigation is required
|
|
|
|
- * to determine this and what the optimal algorithm is to achieve it.
|
|
|
|
- *
|
|
|
|
- * (1) makes you always pay for exactly one copy, but its implementation is
|
|
|
|
- * the simplest and its performance is predictable.
|
|
|
|
- *
|
|
|
|
- * So for now, we implement (1) only. If we wish to optimize later, we should
|
|
|
|
- * be able to do it without affecting users.
|
|
|
|
- *
|
|
|
|
- * The strategy is to buffer the segments of data that do *not* depend on
|
|
|
|
- * unknown lengths in one buffer, and keep a separate buffer of segment pointers
|
|
|
|
- * and lengths. When the top-level submessage ends, we can go beginning to end,
|
|
|
|
- * alternating the writing of lengths with memcpy() of the rest of the data.
|
|
|
|
- * At the top level though, no buffering is required.
|
|
|
|
- */
|
|
|
|
|
|
+** upb::Encoder
|
|
|
|
+**
|
|
|
|
+** Since we are implementing pure handlers (ie. without any out-of-band access
|
|
|
|
+** to pre-computed lengths), we have to buffer all submessages before we can
|
|
|
|
+** emit even their first byte.
|
|
|
|
+**
|
|
|
|
+** Not knowing the size of submessages also means we can't write a perfect
|
|
|
|
+** zero-copy implementation, even with buffering. Lengths are stored as
|
|
|
|
+** varints, which means that we don't know how many bytes to reserve for the
|
|
|
|
+** length until we know what the length is.
|
|
|
|
+**
|
|
|
|
+** This leaves us with three main choices:
|
|
|
|
+**
|
|
|
|
+** 1. buffer all submessage data in a temporary buffer, then copy it exactly
|
|
|
|
+** once into the output buffer.
|
|
|
|
+**
|
|
|
|
+** 2. attempt to buffer data directly into the output buffer, estimating how
|
|
|
|
+** many bytes each length will take. When our guesses are wrong, use
|
|
|
|
+** memmove() to grow or shrink the allotted space.
|
|
|
|
+**
|
|
|
|
+** 3. buffer directly into the output buffer, allocating a max length
|
|
|
|
+** ahead-of-time for each submessage length. If we overallocated, we waste
|
|
|
|
+** space, but no memcpy() or memmove() is required. This approach requires
|
|
|
|
+** defining a maximum size for submessages and rejecting submessages that
|
|
|
|
+** exceed that size.
|
|
|
|
+**
|
|
|
|
+** (2) and (3) have the potential to have better performance, but they are more
|
|
|
|
+** complicated and subtle to implement:
|
|
|
|
+**
|
|
|
|
+** (3) requires making an arbitrary choice of the maximum message size; it
|
|
|
|
+** wastes space when submessages are shorter than this and fails
|
|
|
|
+** completely when they are longer. This makes it more finicky and
|
|
|
|
+** requires configuration based on the input. It also makes it impossible
|
|
|
|
+** to perfectly match the output of reference encoders that always use the
|
|
|
|
+** optimal amount of space for each length.
|
|
|
|
+**
|
|
|
|
+** (2) requires guessing the the size upfront, and if multiple lengths are
|
|
|
|
+** guessed wrong the minimum required number of memmove() operations may
|
|
|
|
+** be complicated to compute correctly. Implemented properly, it may have
|
|
|
|
+** a useful amortized or average cost, but more investigation is required
|
|
|
|
+** to determine this and what the optimal algorithm is to achieve it.
|
|
|
|
+**
|
|
|
|
+** (1) makes you always pay for exactly one copy, but its implementation is
|
|
|
|
+** the simplest and its performance is predictable.
|
|
|
|
+**
|
|
|
|
+** So for now, we implement (1) only. If we wish to optimize later, we should
|
|
|
|
+** be able to do it without affecting users.
|
|
|
|
+**
|
|
|
|
+** The strategy is to buffer the segments of data that do *not* depend on
|
|
|
|
+** unknown lengths in one buffer, and keep a separate buffer of segment pointers
|
|
|
|
+** and lengths. When the top-level submessage ends, we can go beginning to end,
|
|
|
|
+** alternating the writing of lengths with memcpy() of the rest of the data.
|
|
|
|
+** At the top level though, no buffering is required.
|
|
|
|
+*/
|
|
|
|
|
|
|
|
|
|
#include <stdlib.h>
|
|
#include <stdlib.h>
|
|
@@ -9095,12 +9038,6 @@ upb_pb_encoder *upb_pb_encoder_create(upb_env *env, const upb_handlers *h,
|
|
}
|
|
}
|
|
|
|
|
|
upb_sink *upb_pb_encoder_input(upb_pb_encoder *e) { return &e->input_; }
|
|
upb_sink *upb_pb_encoder_input(upb_pb_encoder *e) { return &e->input_; }
|
|
-/*
|
|
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2010-2012 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- */
|
|
|
|
|
|
|
|
|
|
|
|
#include <stdio.h>
|
|
#include <stdio.h>
|
|
@@ -9189,10 +9126,7 @@ bool upb_load_descriptor_file_into_symtab(upb_symtab *symtab, const char *fname,
|
|
return success;
|
|
return success;
|
|
}
|
|
}
|
|
/*
|
|
/*
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2009 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
|
|
+ * upb::pb::TextPrinter
|
|
*
|
|
*
|
|
* OPT: This is not optimized at all. It uses printf() which parses the format
|
|
* OPT: This is not optimized at all. It uses printf() which parses the format
|
|
* string every time, and it allocates memory for every put.
|
|
* string every time, and it allocates memory for every put.
|
|
@@ -9529,12 +9463,6 @@ upb_sink *upb_textprinter_input(upb_textprinter *p) { return &p->input_; }
|
|
void upb_textprinter_setsingleline(upb_textprinter *p, bool single_line) {
|
|
void upb_textprinter_setsingleline(upb_textprinter *p, bool single_line) {
|
|
p->single_line_ = single_line;
|
|
p->single_line_ = single_line;
|
|
}
|
|
}
|
|
-/*
|
|
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2011 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- */
|
|
|
|
|
|
|
|
|
|
|
|
/* Index is descriptor type. */
|
|
/* Index is descriptor type. */
|
|
@@ -9662,28 +9590,25 @@ upb_decoderet upb_vdecode_max8_wright(upb_decoderet r) {
|
|
|
|
|
|
#line 1 "upb/json/parser.rl"
|
|
#line 1 "upb/json/parser.rl"
|
|
/*
|
|
/*
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2014 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- *
|
|
|
|
- * A parser that uses the Ragel State Machine Compiler to generate
|
|
|
|
- * the finite automata.
|
|
|
|
- *
|
|
|
|
- * Ragel only natively handles regular languages, but we can manually
|
|
|
|
- * program it a bit to handle context-free languages like JSON, by using
|
|
|
|
- * the "fcall" and "fret" constructs.
|
|
|
|
- *
|
|
|
|
- * This parser can handle the basics, but needs several things to be fleshed
|
|
|
|
- * out:
|
|
|
|
- *
|
|
|
|
- * - handling of unicode escape sequences (including high surrogate pairs).
|
|
|
|
- * - properly check and report errors for unknown fields, stack overflow,
|
|
|
|
- * improper array nesting (or lack of nesting).
|
|
|
|
- * - handling of base64 sequences with padding characters.
|
|
|
|
- * - handling of push-back (non-success returns from sink functions).
|
|
|
|
- * - handling of keys/escape-sequences/etc that span input buffers.
|
|
|
|
- */
|
|
|
|
|
|
+** upb::json::Parser (upb_json_parser)
|
|
|
|
+**
|
|
|
|
+** A parser that uses the Ragel State Machine Compiler to generate
|
|
|
|
+** the finite automata.
|
|
|
|
+**
|
|
|
|
+** Ragel only natively handles regular languages, but we can manually
|
|
|
|
+** program it a bit to handle context-free languages like JSON, by using
|
|
|
|
+** the "fcall" and "fret" constructs.
|
|
|
|
+**
|
|
|
|
+** This parser can handle the basics, but needs several things to be fleshed
|
|
|
|
+** out:
|
|
|
|
+**
|
|
|
|
+** - handling of unicode escape sequences (including high surrogate pairs).
|
|
|
|
+** - properly check and report errors for unknown fields, stack overflow,
|
|
|
|
+** improper array nesting (or lack of nesting).
|
|
|
|
+** - handling of base64 sequences with padding characters.
|
|
|
|
+** - handling of push-back (non-success returns from sink functions).
|
|
|
|
+** - handling of keys/escape-sequences/etc that span input buffers.
|
|
|
|
+*/
|
|
|
|
|
|
#include <stdio.h>
|
|
#include <stdio.h>
|
|
#include <stdint.h>
|
|
#include <stdint.h>
|
|
@@ -9731,7 +9656,7 @@ struct upb_json_parser {
|
|
upb_jsonparser_frame *top;
|
|
upb_jsonparser_frame *top;
|
|
upb_jsonparser_frame *limit;
|
|
upb_jsonparser_frame *limit;
|
|
|
|
|
|
- upb_status *status;
|
|
|
|
|
|
+ upb_status status;
|
|
|
|
|
|
/* Ragel's internal parsing stack for the parsing state machine. */
|
|
/* Ragel's internal parsing stack for the parsing state machine. */
|
|
int current_state;
|
|
int current_state;
|
|
@@ -9778,7 +9703,8 @@ static upb_selector_t parser_getsel(upb_json_parser *p) {
|
|
|
|
|
|
static bool check_stack(upb_json_parser *p) {
|
|
static bool check_stack(upb_json_parser *p) {
|
|
if ((p->top + 1) == p->limit) {
|
|
if ((p->top + 1) == p->limit) {
|
|
- upb_status_seterrmsg(p->status, "Nesting too deep");
|
|
|
|
|
|
+ upb_status_seterrmsg(&p->status, "Nesting too deep");
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
|
|
|
|
@@ -9860,9 +9786,10 @@ static bool base64_push(upb_json_parser *p, upb_selector_t sel, const char *ptr,
|
|
char output[3];
|
|
char output[3];
|
|
|
|
|
|
if (limit - ptr < 4) {
|
|
if (limit - ptr < 4) {
|
|
- upb_status_seterrf(p->status,
|
|
|
|
|
|
+ upb_status_seterrf(&p->status,
|
|
"Base64 input for bytes field not a multiple of 4: %s",
|
|
"Base64 input for bytes field not a multiple of 4: %s",
|
|
upb_fielddef_name(p->top->f));
|
|
upb_fielddef_name(p->top->f));
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
|
|
|
|
@@ -9886,9 +9813,10 @@ static bool base64_push(upb_json_parser *p, upb_selector_t sel, const char *ptr,
|
|
otherchar:
|
|
otherchar:
|
|
if (nonbase64(ptr[0]) || nonbase64(ptr[1]) || nonbase64(ptr[2]) ||
|
|
if (nonbase64(ptr[0]) || nonbase64(ptr[1]) || nonbase64(ptr[2]) ||
|
|
nonbase64(ptr[3]) ) {
|
|
nonbase64(ptr[3]) ) {
|
|
- upb_status_seterrf(p->status,
|
|
|
|
|
|
+ upb_status_seterrf(&p->status,
|
|
"Non-base64 characters in bytes field: %s",
|
|
"Non-base64 characters in bytes field: %s",
|
|
upb_fielddef_name(p->top->f));
|
|
upb_fielddef_name(p->top->f));
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
} if (ptr[2] == '=') {
|
|
} if (ptr[2] == '=') {
|
|
uint32_t val;
|
|
uint32_t val;
|
|
@@ -9926,10 +9854,11 @@ otherchar:
|
|
}
|
|
}
|
|
|
|
|
|
badpadding:
|
|
badpadding:
|
|
- upb_status_seterrf(p->status,
|
|
|
|
|
|
+ upb_status_seterrf(&p->status,
|
|
"Incorrect base64 padding for field: %s (%.*s)",
|
|
"Incorrect base64 padding for field: %s (%.*s)",
|
|
upb_fielddef_name(p->top->f),
|
|
upb_fielddef_name(p->top->f),
|
|
4, ptr);
|
|
4, ptr);
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
|
|
|
|
@@ -9976,7 +9905,8 @@ static bool accumulate_realloc(upb_json_parser *p, size_t need) {
|
|
|
|
|
|
mem = upb_env_realloc(p->env, p->accumulate_buf, old_size, new_size);
|
|
mem = upb_env_realloc(p->env, p->accumulate_buf, old_size, new_size);
|
|
if (!mem) {
|
|
if (!mem) {
|
|
- upb_status_seterrmsg(p->status, "Out of memory allocating buffer.");
|
|
|
|
|
|
+ upb_status_seterrmsg(&p->status, "Out of memory allocating buffer.");
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
|
|
|
|
@@ -9999,7 +9929,8 @@ static bool accumulate_append(upb_json_parser *p, const char *buf, size_t len,
|
|
}
|
|
}
|
|
|
|
|
|
if (!checked_add(p->accumulated_len, len, &need)) {
|
|
if (!checked_add(p->accumulated_len, len, &need)) {
|
|
- upb_status_seterrmsg(p->status, "Integer overflow.");
|
|
|
|
|
|
+ upb_status_seterrmsg(&p->status, "Integer overflow.");
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
|
|
|
|
@@ -10077,7 +10008,8 @@ static bool multipart_text(upb_json_parser *p, const char *buf, size_t len,
|
|
switch (p->multipart_state) {
|
|
switch (p->multipart_state) {
|
|
case MULTIPART_INACTIVE:
|
|
case MULTIPART_INACTIVE:
|
|
upb_status_seterrmsg(
|
|
upb_status_seterrmsg(
|
|
- p->status, "Internal error: unexpected state MULTIPART_INACTIVE");
|
|
|
|
|
|
+ &p->status, "Internal error: unexpected state MULTIPART_INACTIVE");
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
|
|
|
|
case MULTIPART_ACCUMULATE:
|
|
case MULTIPART_ACCUMULATE:
|
|
@@ -10336,7 +10268,8 @@ static bool parse_number(upb_json_parser *p) {
|
|
return true;
|
|
return true;
|
|
|
|
|
|
err:
|
|
err:
|
|
- upb_status_seterrf(p->status, "error parsing number: %s", buf);
|
|
|
|
|
|
+ upb_status_seterrf(&p->status, "error parsing number: %s", buf);
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
multipart_end(p);
|
|
multipart_end(p);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
@@ -10345,9 +10278,10 @@ static bool parser_putbool(upb_json_parser *p, bool val) {
|
|
bool ok;
|
|
bool ok;
|
|
|
|
|
|
if (upb_fielddef_type(p->top->f) != UPB_TYPE_BOOL) {
|
|
if (upb_fielddef_type(p->top->f) != UPB_TYPE_BOOL) {
|
|
- upb_status_seterrf(p->status,
|
|
|
|
|
|
+ upb_status_seterrf(&p->status,
|
|
"Boolean value specified for non-bool field: %s",
|
|
"Boolean value specified for non-bool field: %s",
|
|
upb_fielddef_name(p->top->f));
|
|
upb_fielddef_name(p->top->f));
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
|
|
|
|
@@ -10398,9 +10332,10 @@ static bool start_stringval(upb_json_parser *p) {
|
|
multipart_startaccum(p);
|
|
multipart_startaccum(p);
|
|
return true;
|
|
return true;
|
|
} else {
|
|
} else {
|
|
- upb_status_seterrf(p->status,
|
|
|
|
|
|
+ upb_status_seterrf(&p->status,
|
|
"String specified for non-string/non-enum field: %s",
|
|
"String specified for non-string/non-enum field: %s",
|
|
upb_fielddef_name(p->top->f));
|
|
upb_fielddef_name(p->top->f));
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
@@ -10438,7 +10373,8 @@ static bool end_stringval(upb_json_parser *p) {
|
|
upb_selector_t sel = parser_getsel(p);
|
|
upb_selector_t sel = parser_getsel(p);
|
|
upb_sink_putint32(&p->top->sink, sel, int_val);
|
|
upb_sink_putint32(&p->top->sink, sel, int_val);
|
|
} else {
|
|
} else {
|
|
- upb_status_seterrf(p->status, "Enum value unknown: '%.*s'", len, buf);
|
|
|
|
|
|
+ upb_status_seterrf(&p->status, "Enum value unknown: '%.*s'", len, buf);
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
}
|
|
}
|
|
|
|
|
|
break;
|
|
break;
|
|
@@ -10446,7 +10382,8 @@ static bool end_stringval(upb_json_parser *p) {
|
|
|
|
|
|
default:
|
|
default:
|
|
assert(false);
|
|
assert(false);
|
|
- upb_status_seterrmsg(p->status, "Internal error in JSON decoder");
|
|
|
|
|
|
+ upb_status_seterrmsg(&p->status, "Internal error in JSON decoder");
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
ok = false;
|
|
ok = false;
|
|
break;
|
|
break;
|
|
}
|
|
}
|
|
@@ -10476,7 +10413,8 @@ static bool parse_mapentry_key(upb_json_parser *p) {
|
|
|
|
|
|
p->top->f = upb_msgdef_itof(p->top->m, UPB_MAPENTRY_KEY);
|
|
p->top->f = upb_msgdef_itof(p->top->m, UPB_MAPENTRY_KEY);
|
|
if (p->top->f == NULL) {
|
|
if (p->top->f == NULL) {
|
|
- upb_status_seterrmsg(p->status, "mapentry message has no key");
|
|
|
|
|
|
+ upb_status_seterrmsg(&p->status, "mapentry message has no key");
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
switch (upb_fielddef_type(p->top->f)) {
|
|
switch (upb_fielddef_type(p->top->f)) {
|
|
@@ -10499,8 +10437,9 @@ static bool parse_mapentry_key(upb_json_parser *p) {
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
} else {
|
|
} else {
|
|
- upb_status_seterrmsg(p->status,
|
|
|
|
|
|
+ upb_status_seterrmsg(&p->status,
|
|
"Map bool key not 'true' or 'false'");
|
|
"Map bool key not 'true' or 'false'");
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
multipart_end(p);
|
|
multipart_end(p);
|
|
@@ -10518,7 +10457,8 @@ static bool parse_mapentry_key(upb_json_parser *p) {
|
|
break;
|
|
break;
|
|
}
|
|
}
|
|
default:
|
|
default:
|
|
- upb_status_seterrmsg(p->status, "Invalid field type for map key");
|
|
|
|
|
|
+ upb_status_seterrmsg(&p->status, "Invalid field type for map key");
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
|
|
|
|
@@ -10573,7 +10513,8 @@ static bool handle_mapentry(upb_json_parser *p) {
|
|
p->top->is_mapentry = true; /* set up to pop frame after value is parsed. */
|
|
p->top->is_mapentry = true; /* set up to pop frame after value is parsed. */
|
|
p->top->mapfield = mapfield;
|
|
p->top->mapfield = mapfield;
|
|
if (p->top->f == NULL) {
|
|
if (p->top->f == NULL) {
|
|
- upb_status_seterrmsg(p->status, "mapentry message has no value");
|
|
|
|
|
|
+ upb_status_seterrmsg(&p->status, "mapentry message has no value");
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
|
|
|
|
@@ -10593,7 +10534,8 @@ static bool end_membername(upb_json_parser *p) {
|
|
if (!f) {
|
|
if (!f) {
|
|
/* TODO(haberman): Ignore unknown fields if requested/configured to do
|
|
/* TODO(haberman): Ignore unknown fields if requested/configured to do
|
|
* so. */
|
|
* so. */
|
|
- upb_status_seterrf(p->status, "No such field: %.*s\n", (int)len, buf);
|
|
|
|
|
|
+ upb_status_seterrf(&p->status, "No such field: %.*s\n", (int)len, buf);
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
|
|
|
|
@@ -10669,9 +10611,10 @@ static bool start_subobject(upb_json_parser *p) {
|
|
|
|
|
|
return true;
|
|
return true;
|
|
} else {
|
|
} else {
|
|
- upb_status_seterrf(p->status,
|
|
|
|
|
|
+ upb_status_seterrf(&p->status,
|
|
"Object specified for non-message/group field: %s",
|
|
"Object specified for non-message/group field: %s",
|
|
upb_fielddef_name(p->top->f));
|
|
upb_fielddef_name(p->top->f));
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
@@ -10697,9 +10640,10 @@ static bool start_array(upb_json_parser *p) {
|
|
assert(p->top->f);
|
|
assert(p->top->f);
|
|
|
|
|
|
if (!upb_fielddef_isseq(p->top->f)) {
|
|
if (!upb_fielddef_isseq(p->top->f)) {
|
|
- upb_status_seterrf(p->status,
|
|
|
|
|
|
+ upb_status_seterrf(&p->status,
|
|
"Array specified for non-repeated field: %s",
|
|
"Array specified for non-repeated field: %s",
|
|
upb_fielddef_name(p->top->f));
|
|
upb_fielddef_name(p->top->f));
|
|
|
|
+ upb_env_reporterror(p->env, &p->status);
|
|
return false;
|
|
return false;
|
|
}
|
|
}
|
|
|
|
|
|
@@ -10736,7 +10680,11 @@ static void start_object(upb_json_parser *p) {
|
|
static void end_object(upb_json_parser *p) {
|
|
static void end_object(upb_json_parser *p) {
|
|
if (!p->top->is_map) {
|
|
if (!p->top->is_map) {
|
|
upb_status status;
|
|
upb_status status;
|
|
|
|
+ upb_status_clear(&status);
|
|
upb_sink_endmsg(&p->top->sink, &status);
|
|
upb_sink_endmsg(&p->top->sink, &status);
|
|
|
|
+ if (!upb_ok(&status)) {
|
|
|
|
+ upb_env_reporterror(p->env, &status);
|
|
|
|
+ }
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
|
|
@@ -10762,11 +10710,11 @@ static void end_object(upb_json_parser *p) {
|
|
* final state once, when the closing '"' is seen. */
|
|
* final state once, when the closing '"' is seen. */
|
|
|
|
|
|
|
|
|
|
-#line 1198 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1218 "upb/json/parser.rl"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
-#line 1110 "upb/json/parser.c"
|
|
|
|
|
|
+#line 1130 "upb/json/parser.c"
|
|
static const char _json_actions[] = {
|
|
static const char _json_actions[] = {
|
|
0, 1, 0, 1, 2, 1, 3, 1,
|
|
0, 1, 0, 1, 2, 1, 3, 1,
|
|
5, 1, 6, 1, 7, 1, 8, 1,
|
|
5, 1, 6, 1, 7, 1, 8, 1,
|
|
@@ -10915,7 +10863,7 @@ static const int json_en_value_machine = 27;
|
|
static const int json_en_main = 1;
|
|
static const int json_en_main = 1;
|
|
|
|
|
|
|
|
|
|
-#line 1201 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1221 "upb/json/parser.rl"
|
|
|
|
|
|
size_t parse(void *closure, const void *hd, const char *buf, size_t size,
|
|
size_t parse(void *closure, const void *hd, const char *buf, size_t size,
|
|
const upb_bufhandle *handle) {
|
|
const upb_bufhandle *handle) {
|
|
@@ -10937,7 +10885,7 @@ size_t parse(void *closure, const void *hd, const char *buf, size_t size,
|
|
capture_resume(parser, buf);
|
|
capture_resume(parser, buf);
|
|
|
|
|
|
|
|
|
|
-#line 1281 "upb/json/parser.c"
|
|
|
|
|
|
+#line 1301 "upb/json/parser.c"
|
|
{
|
|
{
|
|
int _klen;
|
|
int _klen;
|
|
unsigned int _trans;
|
|
unsigned int _trans;
|
|
@@ -11012,118 +10960,118 @@ _match:
|
|
switch ( *_acts++ )
|
|
switch ( *_acts++ )
|
|
{
|
|
{
|
|
case 0:
|
|
case 0:
|
|
-#line 1113 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1133 "upb/json/parser.rl"
|
|
{ p--; {cs = stack[--top]; goto _again;} }
|
|
{ p--; {cs = stack[--top]; goto _again;} }
|
|
break;
|
|
break;
|
|
case 1:
|
|
case 1:
|
|
-#line 1114 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1134 "upb/json/parser.rl"
|
|
{ p--; {stack[top++] = cs; cs = 10; goto _again;} }
|
|
{ p--; {stack[top++] = cs; cs = 10; goto _again;} }
|
|
break;
|
|
break;
|
|
case 2:
|
|
case 2:
|
|
-#line 1118 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1138 "upb/json/parser.rl"
|
|
{ start_text(parser, p); }
|
|
{ start_text(parser, p); }
|
|
break;
|
|
break;
|
|
case 3:
|
|
case 3:
|
|
-#line 1119 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1139 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(end_text(parser, p)); }
|
|
{ CHECK_RETURN_TOP(end_text(parser, p)); }
|
|
break;
|
|
break;
|
|
case 4:
|
|
case 4:
|
|
-#line 1125 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1145 "upb/json/parser.rl"
|
|
{ start_hex(parser); }
|
|
{ start_hex(parser); }
|
|
break;
|
|
break;
|
|
case 5:
|
|
case 5:
|
|
-#line 1126 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1146 "upb/json/parser.rl"
|
|
{ hexdigit(parser, p); }
|
|
{ hexdigit(parser, p); }
|
|
break;
|
|
break;
|
|
case 6:
|
|
case 6:
|
|
-#line 1127 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1147 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(end_hex(parser)); }
|
|
{ CHECK_RETURN_TOP(end_hex(parser)); }
|
|
break;
|
|
break;
|
|
case 7:
|
|
case 7:
|
|
-#line 1133 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1153 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(escape(parser, p)); }
|
|
{ CHECK_RETURN_TOP(escape(parser, p)); }
|
|
break;
|
|
break;
|
|
case 8:
|
|
case 8:
|
|
-#line 1139 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1159 "upb/json/parser.rl"
|
|
{ p--; {cs = stack[--top]; goto _again;} }
|
|
{ p--; {cs = stack[--top]; goto _again;} }
|
|
break;
|
|
break;
|
|
case 9:
|
|
case 9:
|
|
-#line 1142 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1162 "upb/json/parser.rl"
|
|
{ {stack[top++] = cs; cs = 19; goto _again;} }
|
|
{ {stack[top++] = cs; cs = 19; goto _again;} }
|
|
break;
|
|
break;
|
|
case 10:
|
|
case 10:
|
|
-#line 1144 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1164 "upb/json/parser.rl"
|
|
{ p--; {stack[top++] = cs; cs = 27; goto _again;} }
|
|
{ p--; {stack[top++] = cs; cs = 27; goto _again;} }
|
|
break;
|
|
break;
|
|
case 11:
|
|
case 11:
|
|
-#line 1149 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1169 "upb/json/parser.rl"
|
|
{ start_member(parser); }
|
|
{ start_member(parser); }
|
|
break;
|
|
break;
|
|
case 12:
|
|
case 12:
|
|
-#line 1150 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1170 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(end_membername(parser)); }
|
|
{ CHECK_RETURN_TOP(end_membername(parser)); }
|
|
break;
|
|
break;
|
|
case 13:
|
|
case 13:
|
|
-#line 1153 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1173 "upb/json/parser.rl"
|
|
{ end_member(parser); }
|
|
{ end_member(parser); }
|
|
break;
|
|
break;
|
|
case 14:
|
|
case 14:
|
|
-#line 1159 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1179 "upb/json/parser.rl"
|
|
{ start_object(parser); }
|
|
{ start_object(parser); }
|
|
break;
|
|
break;
|
|
case 15:
|
|
case 15:
|
|
-#line 1162 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1182 "upb/json/parser.rl"
|
|
{ end_object(parser); }
|
|
{ end_object(parser); }
|
|
break;
|
|
break;
|
|
case 16:
|
|
case 16:
|
|
-#line 1168 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1188 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(start_array(parser)); }
|
|
{ CHECK_RETURN_TOP(start_array(parser)); }
|
|
break;
|
|
break;
|
|
case 17:
|
|
case 17:
|
|
-#line 1172 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1192 "upb/json/parser.rl"
|
|
{ end_array(parser); }
|
|
{ end_array(parser); }
|
|
break;
|
|
break;
|
|
case 18:
|
|
case 18:
|
|
-#line 1177 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1197 "upb/json/parser.rl"
|
|
{ start_number(parser, p); }
|
|
{ start_number(parser, p); }
|
|
break;
|
|
break;
|
|
case 19:
|
|
case 19:
|
|
-#line 1178 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1198 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(end_number(parser, p)); }
|
|
{ CHECK_RETURN_TOP(end_number(parser, p)); }
|
|
break;
|
|
break;
|
|
case 20:
|
|
case 20:
|
|
-#line 1180 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1200 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(start_stringval(parser)); }
|
|
{ CHECK_RETURN_TOP(start_stringval(parser)); }
|
|
break;
|
|
break;
|
|
case 21:
|
|
case 21:
|
|
-#line 1181 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1201 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(end_stringval(parser)); }
|
|
{ CHECK_RETURN_TOP(end_stringval(parser)); }
|
|
break;
|
|
break;
|
|
case 22:
|
|
case 22:
|
|
-#line 1183 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1203 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(parser_putbool(parser, true)); }
|
|
{ CHECK_RETURN_TOP(parser_putbool(parser, true)); }
|
|
break;
|
|
break;
|
|
case 23:
|
|
case 23:
|
|
-#line 1185 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1205 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(parser_putbool(parser, false)); }
|
|
{ CHECK_RETURN_TOP(parser_putbool(parser, false)); }
|
|
break;
|
|
break;
|
|
case 24:
|
|
case 24:
|
|
-#line 1187 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1207 "upb/json/parser.rl"
|
|
{ /* null value */ }
|
|
{ /* null value */ }
|
|
break;
|
|
break;
|
|
case 25:
|
|
case 25:
|
|
-#line 1189 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1209 "upb/json/parser.rl"
|
|
{ CHECK_RETURN_TOP(start_subobject(parser)); }
|
|
{ CHECK_RETURN_TOP(start_subobject(parser)); }
|
|
break;
|
|
break;
|
|
case 26:
|
|
case 26:
|
|
-#line 1190 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1210 "upb/json/parser.rl"
|
|
{ end_subobject(parser); }
|
|
{ end_subobject(parser); }
|
|
break;
|
|
break;
|
|
case 27:
|
|
case 27:
|
|
-#line 1195 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1215 "upb/json/parser.rl"
|
|
{ p--; {cs = stack[--top]; goto _again;} }
|
|
{ p--; {cs = stack[--top]; goto _again;} }
|
|
break;
|
|
break;
|
|
-#line 1467 "upb/json/parser.c"
|
|
|
|
|
|
+#line 1487 "upb/json/parser.c"
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
|
|
@@ -11136,10 +11084,11 @@ _again:
|
|
_out: {}
|
|
_out: {}
|
|
}
|
|
}
|
|
|
|
|
|
-#line 1222 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1242 "upb/json/parser.rl"
|
|
|
|
|
|
if (p != pe) {
|
|
if (p != pe) {
|
|
- upb_status_seterrf(parser->status, "Parse error at %s\n", p);
|
|
|
|
|
|
+ upb_status_seterrf(&parser->status, "Parse error at %s\n", p);
|
|
|
|
+ upb_env_reporterror(parser->env, &parser->status);
|
|
} else {
|
|
} else {
|
|
capture_suspend(parser, &p);
|
|
capture_suspend(parser, &p);
|
|
}
|
|
}
|
|
@@ -11176,19 +11125,20 @@ static void json_parser_reset(upb_json_parser *p) {
|
|
|
|
|
|
/* Emit Ragel initialization of the parser. */
|
|
/* Emit Ragel initialization of the parser. */
|
|
|
|
|
|
-#line 1520 "upb/json/parser.c"
|
|
|
|
|
|
+#line 1541 "upb/json/parser.c"
|
|
{
|
|
{
|
|
cs = json_start;
|
|
cs = json_start;
|
|
top = 0;
|
|
top = 0;
|
|
}
|
|
}
|
|
|
|
|
|
-#line 1261 "upb/json/parser.rl"
|
|
|
|
|
|
+#line 1282 "upb/json/parser.rl"
|
|
p->current_state = cs;
|
|
p->current_state = cs;
|
|
p->parser_top = top;
|
|
p->parser_top = top;
|
|
accumulate_clear(p);
|
|
accumulate_clear(p);
|
|
p->multipart_state = MULTIPART_INACTIVE;
|
|
p->multipart_state = MULTIPART_INACTIVE;
|
|
p->capture = NULL;
|
|
p->capture = NULL;
|
|
p->accumulated = NULL;
|
|
p->accumulated = NULL;
|
|
|
|
+ upb_status_clear(&p->status);
|
|
}
|
|
}
|
|
|
|
|
|
|
|
|
|
@@ -11214,8 +11164,8 @@ upb_json_parser *upb_json_parser_create(upb_env *env, upb_sink *output) {
|
|
upb_sink_reset(&p->top->sink, output->handlers, output->closure);
|
|
upb_sink_reset(&p->top->sink, output->handlers, output->closure);
|
|
p->top->m = upb_handlers_msgdef(output->handlers);
|
|
p->top->m = upb_handlers_msgdef(output->handlers);
|
|
|
|
|
|
- /* If this fails, uncomment and increase the value in parser.h.
|
|
|
|
- * fprintf(stderr, "%zd\n", upb_env_bytesallocated(env) - size_before); */
|
|
|
|
|
|
+ /* If this fails, uncomment and increase the value in parser.h. */
|
|
|
|
+ /* fprintf(stderr, "%zd\n", upb_env_bytesallocated(env) - size_before); */
|
|
assert(upb_env_bytesallocated(env) - size_before <= UPB_JSON_PARSER_SIZE);
|
|
assert(upb_env_bytesallocated(env) - size_before <= UPB_JSON_PARSER_SIZE);
|
|
return p;
|
|
return p;
|
|
}
|
|
}
|
|
@@ -11224,14 +11174,9 @@ upb_bytessink *upb_json_parser_input(upb_json_parser *p) {
|
|
return &p->input_;
|
|
return &p->input_;
|
|
}
|
|
}
|
|
/*
|
|
/*
|
|
- * upb - a minimalist implementation of protocol buffers.
|
|
|
|
- *
|
|
|
|
- * Copyright (c) 2014 Google Inc. See LICENSE for details.
|
|
|
|
- * Author: Josh Haberman <jhaberman@gmail.com>
|
|
|
|
- *
|
|
|
|
- * This currently uses snprintf() to format primitives, and could be optimized
|
|
|
|
- * further.
|
|
|
|
- */
|
|
|
|
|
|
+** This currently uses snprintf() to format primitives, and could be optimized
|
|
|
|
+** further.
|
|
|
|
+*/
|
|
|
|
|
|
|
|
|
|
#include <stdlib.h>
|
|
#include <stdlib.h>
|