David Čepelík

I think > demonstrating that the \r has become part of the value of the header U. was what bothered me the most.

Hi, > But I can't shake the feeling that it's good to be lenient - especially where network parsers are concerned. This is a [good read](https://portswigger.net/daily-swig/new-differential-fuzzing-tool-reveals-novel-http-request-smuggling-techniques) explaining why that might...

Hi, what's the current status of this issue? I just ran into something similar. Is there any ETA on a fix?

Hi @kyleconroy, I would like to contribute this feature. It seems to be quite straightforward, but I have a few questions: - It seems to me that all I need...

> Thanks for catching that! You're very welcome. > I haven't tried bootstrapping on a musl system. I probably won't either. I'd run our make distcheck target to get a...

Thanks for the clarification @Mightyjo. Will take a look at BUILT_SOURCES. > you'll find commits in master that are informational if their PR built correctly at the end. Ah, OK!...

> Hi @dcepelik. Thanks for reporting! (And sorry for taking so long to respond.) Hi @k0ekk0ek, no problem! > We use the `resolve_path` to ensure we have the full path...

To resolve the original issue: - Would it make sense to pass the original file name to fopen (since that will resolve relative paths relative to current working directory anyway)...

> The include checks are done [here](https://github.com/NLnetLabs/simdzone/blob/9a595d54d287c6e61b6578589128cac406f5bf81/src/generic/format.h#L291). It does both, cyclic check and include depth. I prefer to leave that detection as-is. Ah! Of course. I forgot to look into...