condition that cause always passthrough bypass

[ykopel]

Fix for 497e475.
This commit contains a condition that will always be true.
This is because ua_raw_buffer_reader->read_avail() always will be grater than zero.
You haven't see it until now because of another bug (#1471) that cause is_transparent_passthrough_allowed() to be always false and this for it never reached to your lines.
This is why I suggest to just delete those lines meantime.

[SolidWallOfCode]

Have you tested this in practice? ua_buffer_reader->read_avail() should be zero in most cases, because all of the header has been read and the user agent shouldn't send more data after that if there no is keep alive set (which is also checked).

[ykopel]

I tested it in my env. And I see it is grater than zero.
There are several lines before it that assume that "read_avail" is grater than zero:

  // Reset the inactivity timeout if this is the first
  //   time we've been called.  The timeout had been set to
  //   the accept timeout by the ProxyClientTransaction
  //
  if ((ua_buffer_reader->read_avail() > 0) && (client_request_hdr_bytes == 0)) {
    milestones[TS_MILESTONE_UA_FIRST_READ] = Thread::get_hrtime();
    ua_session->set_inactivity_timeout(HRTIME_SECONDS(t_state.txn_conf->transaction_no_activity_timeout_in));
  }

[zwoop]

[approve ci]

[atsci]

Intel CC build successful! See https://ci.trafficserver.apache.org/job/icc-github/31/ for details.

[atsci]

Linux build successful! See https://ci.trafficserver.apache.org/job/linux-github/1493/ for details.

[atsci]

FreeBSD build successful! See https://ci.trafficserver.apache.org/job/freebsd-github/1599/ for details.

[atsci]

clang-analyzer build successful! See https://ci.trafficserver.apache.org/job/clang-analyzer-github/163/ for details.

[oknet]

It is fixed by “TS-5103: replace ua_raw_buffer_reader with ua_buffer_reader” and #1271
please check the code in master branch.

[oknet]

"TS-3100: Extend tr-pass to allow malformed HTTP GET requests to be blind tunneled." is added by @shinrich .

[zwoop]

@oknet Does that mean we should close this PR (without merging) ?

[oknet]

@zwoop yes, close without merging.