This makes both the C (.h) and C++ (.hpp) files read nicer
and keeps the core of upb C-only.
Existing users of the C++ wrappers will have to add manual
#includes of the .hpp files.
New code is smaller (in both source size and compiled size) and faster.
# Speed
The decoder speeds up on all machines I tested, though the amount of speedup varies. I was only able to test Intel CPUs.
### Linux Desktop
```
CPU: Intel(R) Core(TM) i7-8700K CPU @ 3.70GHz
OS: Linux
name old time/op new time/op delta
CreateArena 4.72ns ± 0% 4.93ns ± 0% +4.47% (p=0.000 n=11+11)
ParseDescriptor 12.4µs ± 1% 9.1µs ± 1% -26.65% (p=0.000 n=11+11)
```
### Mac Laptop
```
CPU: Intel(R) Core(TM) i7-8850H CPU @ 2.60GHz
OS: macOS
name old time/op new time/op delta
CreateArena 5.33ns ± 3% 5.58ns ± 2% +4.69% (p=0.000 n=12+12)
ParseDescriptor 15.0µs ± 2% 11.9µs ± 2% -20.20% (p=0.000 n=12+12)
```
### Linux Workstation
```
CPU: Intel(R) Xeon(R) Gold 6154 CPU @ 3.00GHz
OS: Linux
name old time/op new time/op delta
CreateArena 5.29ns ± 0% 5.52ns ± 0% +4.37% (p=0.000 n=10+12)
ParseDescriptor 18.6µs ± 0% 16.4µs ± 0% -11.54% (p=0.000 n=12+12)
```
# Size
A few source files grow marginally because of some arena functionality moved inline. But `upb/decode.c` shrinks by 30% on Linux:
```
VM SIZE
--------------
+2.1% +283 upb/json_decode.c
+24% +205 upb/msg.c
+8.4% +115 upb/upb.c
+0.9% +28 upb/reflection.c
[ = ] 0 upb/def.c
[ = ] 0 upb/encode.c
[ = ] 0 upb/json_encode.c
[ = ] 0 upb/table.c
-30.3% -1.51Ki upb/decode.c
-0.7% -738 TOTAL
```
* Removed reflection and other extraneous things from the core library.
* Added missing files and ran buildifier.
* New CMakeLists.txt.
* Made table its own cc_library() for internal usage.
* Split upb::Arena/upb::Allocator from upb::Environment.
This will allow arenas and allocators to be used
independently of environments, which will be important
for an upcoming change (a message representation).
Overall this design feels cleaner that the previous
Environment/SeededAllocator design.
As part of this change, moved all allocations in upb
to use a global allocator instead of hard-coding
malloc/free. This will allow injecting OOM faults
for more robust testing.
One place that doesn't use the global allocator is
the tracked ref code. Instead of its previous approach
of CHECK_OOM() after every malloc() or table insert, it
simply uses an allocator that does this automatically.
I moved Allocator/Arena/Environment into upb.h.
This seems principled since these are the only types
in upb whose size is directly exposed to users, since
they form the basis of memory allocation strategy.
* Cleaned up some header includes and fixed more malloc -> upb_gmalloc().
* Changes from PR review.
* Don't use UINTPTR_MAX or UINT64_MAX.
* Punt on adding line/file for now.
* We actually can't store (uint64_t)-1, update comment and test.
It is entirely optional: MessageDef/EnumDef can still exist
on their own. But this can represent a def's file when it is
desirable to do so (eg. for code generators).
This approach will require that we change the way we handle
extensions. But I think it will be a good change overall.
Specifically, we previously handled extensions by duplicating
the extended message and then adding the extension as a regular
field to the duplicated message. This required also duplicating
any messages that could reach the extended message.
In the new world we will need a way of declaring and looking up
extensions separately from the message being extended.
This change also involves some notable changes to the generated
code:
- files are now called foo.upbdefs.h instead of foo.upb.h.
This reflects the fact that we might possibly generate several
different output files for a .proto file, and this one is just
for defs.
- we no longer generate selectors in the .h file.
- the upbdefs.c no longer vends a SymbolTable. Now it vends the
individual messages (and possibly a FileDef later). I think this
will compose better once we can generate files where one
generated files imports another.
We also make the descriptor reader vend a list of FileDefs now.
This is the best conceptual match for parsing a FileDescriptorSet.
https://google.github.io/styleguide/cppguide.html#Constructor_Initializer_Lists
```
// If the signature and initializer list are not all on one line,
// you must wrap before the colon and indent 4 spaces:
MyClass::MyClass(int var)
: some_var_(var), some_other_var_(var + 1) {
DoSomething();
}
```
and
```
// As with any other code block, the close curly can be on the same
// line as the open curly, if it fits.
MyClass::MyClass(int var)
: some_var_(var) {}
```
$ make Q= googlepb
g++ -O3 -std=c++98 -pedantic -Wno-long-long -Wall -Wextra -Wpointer-arith -Wno-unused-private-field -I. -DNDEBUG -c -o obj/upb/bindings/googlepb/bridge.o upb/bindings/googlepb/bridge.cc
In file included from ./upb/handlers.h:22,
from ./upb/bindings/googlepb/bridge.h:42,
from upb/bindings/googlepb/bridge.cc:8:
./upb/def.h: In constructor ‘upb::Pointer<upb::Def>::Pointer(upb::Def*)’:
./upb/def.h:39: error: class ‘upb::Pointer<upb::Def>’ does not have any field named ‘PointerBase’
./upb/def.h:39: error: no matching function for call to ‘upb::PointerBase<upb::Def, upb::RefCounted>::PointerBase()’
./upb/upb.h:246: note: candidates are: upb::PointerBase<T, Base>::PointerBase(T*) [with T = upb::Def, Base = upb::RefCounted]
./upb/upb.h:244: note: upb::PointerBase<upb::Def, upb::RefCounted>::PointerBase(const upb::PointerBase<upb::Def, upb::RefCounted>&)
...
the generated code looks like:
template <> class Pointer<upb::Def> : public PointerBase<upb::Def, upb::RefCounted> {
public: explicit Pointer(upb::Def* ptr) : PointerBase(ptr) {}
};
..
which falls into https://gcc.gnu.org/bugzilla/show_bug.cgi?id=189
(http://stackoverflow.com/questions/8887864/template-base-constructor-call-in-member-initialization-list-error)
changing the generated code to:
template <> class Pointer<upb::Def> : public PointerBase<upb::Def, upb::RefCounted> {
public: explicit Pointer(upb::Def* ptr) : PointerBase<upb::Def, upb::RefCounted>(ptr) {}
};
makes it compile at least on 4.4.7 that we are testing with:
$ gcc -v
Using built-in specs.
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugzilla.redhat.com/bugzilla --enable-bootstrap --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions
--enable-gnu-unique-object --enable-languages=c,c++,objc,obj-c++,java,fortran,ada --enable-java-awt=gtk --disable-dssi --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-1.5.0.0/jre --enable-libgcj-multifile --enable-java-maintainer-mode --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --disable-libjava-multilib
--with-ppl --with-cloog --with-tune=generic --with-arch_32=i686 --build=x86_64-redhat-linux
Thread model: posix
gcc version 4.4.7 20120313 (Red Hat 4.4.7-4) (GCC)
A large part of this change contains surface-level
porting, like moving variable declarations to the
top of the block.
However there are a few more substantial things too:
- moved internal-only struct definitions to a separate
file (structdefs.int.h), for greater encapsulation
and ABI compatibility.
- removed the UPB_UPCAST macro, since it requires access
to the internal-only struct definitions. Replaced uses
with calls to inline, type-safe casting functions.
- removed the UPB_DEFINE_CLASS/UPB_DEFINE_STRUCT macros.
Class and struct definitions are now more explicit -- you
get to see the actual class/struct keywords in the source.
The casting convenience functions have been moved into
UPB_DECLARE_DERIVED_TYPE() and UPB_DECLARE_DERIVED_TYPE2().
- the new way that we duplicate base methods in derived types
is also more convenient and requires less duplication.
It is also less greppable, but hopefully that is not
too big a problem.
Compiler flags (-std=c89 -pedantic) should help to rigorously
enforce that the code is free of C99-isms.
A few functions are not available in C89 (strtoll). There
are temporary, hacky solutions in place.
- rewritten decoder; interpreted decoder is bytecode-based,
JIT decoder no longer falls back to the interpreter.
- C++ improvements: C++11-compatible iterators, upb::reffed_ptr
for RAII refcounting, better upcast/downcast support.
- removed the gross upb_value abstraction from public upb.h.
- Better error reporting for upb::Def setters.
- error reporting for upb::Handlers setters.
- made the start/endmsg handlers a little less special-cased.
Major changes:
- Got rid of all bytestream interfaces in favor of
using regular handlers.
- new Pipeline object represents a upb pipeline, does
bump allocation internally to manage memory.
- proto2 support now can handle extensions.
Many things have changed and been simplified.
The memory-management story for upb_def and upb_handlers
is much more robust; upb_def and upb_handlers should be
fairly stable interfaces now. There is still much work
to do for the runtime component (upb_sink).
Many improvements, too many to mention. One significant
perf regression warrants investigation:
omitfp.parsetoproto2_googlemessage1.upb_jit: 343 -> 252 (-26.53)
plain.parsetoproto2_googlemessage1.upb_jit: 334 -> 251 (-24.85)
25% regression for this benchmark is bad, but since I don't think
there's any fundamental design issue that caused it I'm going to
go ahead with the commit anyway. Can investigate and fix later.
Other benchmarks were neutral or showed slight improvement.
Joshua Haberman
Esun Kim
Bo Yang
HATATANI Shinta
borislav nikolov
Josh Haberman