Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.
The documentation says "remove + generate", so I'd say this is then buggy behavior.
I looked briefly at the code and ... it's a bit questionable if it would work but in some certain combinations of calls.
To either replace or remove the header is conditioned additionally by _siputils_pcv_status == PCV_PARSED. From what I could deduce, that's not set, unless you tried to get that header and it was parsed... I have no idea why replace/remove would be conditioned by that. (But when it works, then removes both the value and the header name, I think.)
Adding the header is done with a lump at a position which seems a bit... arbitrary to me: msg->parsed - msg->buf, so where the parsing stopped before. Lumps were always scary to me, so I'd guess maybe some deleted headers pushes it in the message body. Also: the added part is just the header value, skipping the header name part, albeit that's written in a buffer. So matches what you see in the message. I'm wondering how it would work for a simple generate...
The last issue is why it's the same value, which I don't know if it still matters given 1&2. There are some parsing-back extra complications after generating, so that might be the issue. Or lumps issues again?
Maybe @neutrino38 can pitch in?
P.S. I'd suggest an alternative path, to only generate a PCV value in C-code and return it to the script (or if there is a UUID/unique/etc generator function in the script that'd be enough), then use other script functions to add the header with that value wherever you want, remove older values, etc.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.