- From: Daniel Stenberg <daniel@haxx.se>
- Date: Wed, 10 Feb 2021 08:30:42 +0100 (CET)
- To: Adrien de Croy <adrien@qbik.com>
- cc: Wesley Oliver <wesley.olis@gmail.com>, HTTP Working Group <ietf-http-wg@w3.org>
On Tue, 9 Feb 2021, Adrien de Croy wrote: > Browsers probably should fix their treatment of truncated content. That's *exactly* what I thought when I grabbed that work item in the Firefox code back in 2014. I went into that task with confidence based on the knowledge that curl is not that lax on these areas. Yeah, I'm that naive! =) Several months later, after having tried lots of things (that broke the web experience for a lot of peeps), I mostly gave up [1]. > close on connection during chunked download without 0\r\n\r\n should raise > some kind of flag. Back then at least, there were widely used web components that generated such contents with enough users that we got a fair amount of bug reports when we tried it. More specifically they often left out the final zero length chunk. > Same with connection close with C-L prior to advertised number of bytes > arriving. It's been a while since I did this of course, but the wrong C-L case was still then fairly common for gzip-compressed content. At least common enough that we deemed we couldn't impose this behavior strictly on users. I don't know if others have revisited these things since I left Mozilla in late 2018. [1] = https://daniel.haxx.se/blog/2014/10/26/stricter-http-1-1-framing-good-bye/ -- / daniel.haxx.se
Received on Wednesday, 10 February 2021 07:30:57 UTC