fix folly::stringPrintf when ref empty str

[pengweisong]

What type of PR is this?

• bug
• feature
• enhancement

What problem(s) does this PR solve?

Issue(s) number:

#4926

Description:

when the sym_ in PropertyExpression is empty, return reference of string and pass it to folly::stringFormat will crash.

(gdb) bt
#0  0x00007f26e967f079 in vfprintf () from /lib64/libc.so.6
#1  0x00007f26e96aa179 in vsnprintf () from /lib64/libc.so.6
#2  0x0000000002c2a2b3 in ?? ()
#3  0x0000000002c2b737 in folly::stringVPrintf[abi:cxx11](char const*, __va_list_tag*) ()
#4  0x0000000002c2b8df in folly::stringPrintf[abi:cxx11](char const*, ...) ()
#5  0x00000000014ce1ca in nebula::storage::UpdateResNode<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >::doExecute(int, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) ()
#6  0x00000000014bd030 in nebula::storage::RelNode<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >::execute(int, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) ()
#7  0x00000000014bcd86 in nebula::storage::RelNode<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >::doExecute(int, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) ()
#8  0x00000000014bd030 in nebula::storage::RelNode<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >::execute(int, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) ()
#9  0x00000000014bbfb0 in nebula::storage::UpdateVertexProcessor::doProcess(nebula::storage::cpp2::UpdateVertexRequest const&) ()
#10 0x00000000028b3a27 in virtual thunk to apache::thrift::concurrency::FunctionRunner::run() ()
#11 0x0000000002a0ea38 in apache::thrift::concurrency::ThreadManager::Impl::Worker::run() ()
#12 0x0000000002a10b3e in apache::thrift::concurrency::PthreadThread::threadMain(void*) ()
#13 0x00007f26e9a07ea5 in start_thread () from /lib64/libpthread.so.0
#14 0x00007f26e9730b0d in clone () from /lib64/libc.so.6

How do you solve it?

just use exp->toString(), in which will test empty, and then copy.

Special notes for your reviewer, ex. impact of this fix, design document, etc:

Checklist:

Tests:

• Unit test(positive and negative cases)
• Function test
• Performance test
• N/A

Affects:

• Documentation affected (Please add the label if documentation needs to be modified.)
• Incompatibility (If it breaks the compatibility, please describe it and add the label.）
• If it's needed to cherry-pick (If cherry-pick to some branches is required, please label the destination version(s).)
• Performance impacted: Consumes more CPU/Memory

Release notes:

Please confirm whether to be reflected in release notes and how to describe:

ex. Fixed the bug .....

[critical27]

@jievince, does graphd add something like constant folding?

UPDATE  VERTEX ON `aa` 'a:1005' SET `name` = 'test' YIELD 'a:1005' AS `aId`, `name`;

The root cause is that in YIELD 'a:1005' AS aId, previously it is a PropertyExpression or at least storaged assume it is. But it is a now a ConstantExpression, see the debug info below

two expression in returnPropsExp_ is 'a:1005' AS aId and name

[jievince]

'a:1005' is a ConstantExpression generated by the parser.
The graph validator doesn't check the non-property related expr in the yield clause of update statement.
Do we allow such usage?

[critical27]

Ok, I see. You could just remove the static_cast, and use exp->toString here. That could solve it. @pengweisong

[pengweisong]

Ok, I see. You could just remove the static_cast, and use exp->toString here. That could solve it. @pengweisong

OK, fixed, and takes some time to compile and verify the sentences in the issue.

[HarrisChu]

could we add some tck cases?

[pengweisong]

could we add some tck cases?

OK, I'll try.