THRIFT-5670: lib: cpp: Fix wrong timeout error message

[stiga-huang]

We've seen a strange timeout error message like this:

TSocket::read() THRIFT_EAGAIN (timed out) after %f ms: Unknown error 30000

The log comes from this line:

GlobalOutput.perror("TSocket::read() THRIFT_EAGAIN (timed out) after %f ms", recvTimeout_);

The second argument of GlobalOutput.perror() is an errno. We should not use recvTimeout_ on it:

thrift/lib/cpp/src/thrift/TOutput.h

Line 40 in a2a1f53

void perror(const char* message, int errno_copy);

class TOutput {
  ...
  void perror(const char* message, int errno_copy);  // <-- The second arg is errno
  ...
  void printf(const char* message, ...);
};
THRIFT_EXPORT extern TOutput GlobalOutput;

The implementation of perror():

void TOutput::perror(const char* message, int errno_copy) {
  std::string out = message + std::string(": ") + strerror_s(errno_copy);
  f_(out.c_str());
}

It converts the errno to the error string, and 30000 is an unknown errno.

What we want is to print the timeout threshold. We should use GlobalOutput.printf() instead.

This PR fixes the wrong usage of GlobalOutput.perror(). The error message will be

TSocket::read() THRIFT_EAGAIN (timed out) after 30000 ms

• Did you create an Apache Jira ticket? (not required for trivial changes)
• If a ticket exists: Does your pull request title follow the pattern "THRIFT-NNNN: describe my issue"?
• Did you squash your changes to a single commit? (not required, but preferred)
• Did you do your best to avoid breaking changes? If one was needed, did you label the Jira ticket with "Breaking-Change"?
• If your change does not involve any code, include [skip ci] anywhere in the commit message to free up build resources.

[stiga-huang]

@emmenlau Could you help to review this?

[stiga-huang]

Ping @emmenlau @Jens-G

[stiga-huang]

Updated the PR description. Hopefully it's more clear about the issue.

[emmenlau]

LGTM.