https://bz.apache.org/bugzilla/show_bug.cgi?id=56729
--- Comment #28 from Michael Kaufmann <[email protected]> --- > You possibly should test it... Yes, you are right. I have tested the latest patch and it indeed fixes this bug! Sorry that I have made wrong assumptions. > After r1621453 and r1641376 (which BTW I assume you applied already) That's correct, I have tested including these two changes. > thus the racy log_transaction hook reqtimeout_after_body() is not needed > anymore. Good refactoring! (minor issue: the old hook name "log_transaction" is still mentioned in a comment) > I can't see how the state could be wrong (from moq_reqt POV...). Maybe some future or custom MPM does nonblocking+speculative reads in the keep-alive state. So perhaps the corresponding "if" block that handles this case should be moved to the top of the function. But this is probably more a theoretical issue. Yann, I really appreciate the time and effort that you have put into fixing this bug. Keep up the good work! -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
