Vinnie Falco wrote:
On Tue, Jul 11, 2017 at 10:21 AM, Phil Endecott via Boost
wrote: I was hoping one of those links would tell me what exactly failed....
https://travis-ci.org/vinniefalco/Beast/jobs/252431572#L6391
e.g. https://github.com/vinniefalco/Beast/blob/29cd9d6a9e1e7217798be06ded299ffac5...
OK, I think that is an "overlong encoding" issue, where the UTF-8 would decode to a codepoint that should be represented by a shorter sequence. The UTF-8 decoder that I used as a starting point for this happily accepts such input. There are probably other issues. For example, I don't think the "skip pending" loop does the right thing when the buffer ends and the pending bytes end at the same time. I wasn't expecting you to actually try to use this code! Regards, Phil.