Attila Szakacs

Hi @hebaodanroot Thanks for opening this issue ticket! :) For further investigation, can you share the following? 1. Do you have an example input log, which reproduces the problem? 2....

I have opened a PR, which intends to fix the problem. The PR has not been reviewed yet, so there is no guarantee of its behavior, but for testing purposes,...

If we reached `log-msg-size()` we would have seen error or notice messages about it. To investigate why the character parsing failed, we need an example incoming message, which reproduces the...

Okay, let me prepare some additional logging just for this investigation :)

I have added some temporary logs for this investigation only. For the code, see: https://github.com/alltilla/syslog-ng/commit/f423e0a19b3080263bc0e7b5adca4e582d664473 You do not have to enable trace logging with this change, these new messages will...

Sorry I do not understand what you are asking, can you please rephrase your question? I think @MrAnno's answer explained what syslog-ng is able to do in your situation.

I cannot say anything more: syslog-ng cannot handle an incoming log message with mixed encodings, neither could any text editor. Either you fix the message on the sender side, or...

I have reviewed the PR up until the commit: `filter-cmp: implement type-aware comparison of values`. Maybe I misunderstand something, but I think this implementation does not mimic JavaScript's behavior. E.g....

Maybe we could split off the generic number parts of this PR. I would definitely approve that change, and it would help out in a lot of other generic number...

> hmm... I concentrated so much on JavaScript behavior when the types are NOT the same that I missed that if they are the same, the behaviour is not useful...