It was breaking a non-conforming http client I was using, not breaking anything anymore. I'd be happy to leave as is, since any conformant http client should be ignoring content-length when a non-identity transfer-encoding is also present.

@diasdavid saw something weird today. on https://ipfs.io homepage:

https://ipfs.io/styles/bootstrap/css/bootstrap.min.css Failed to load resource: net::ERR_CONTENT_LENGTH_MISMATCH
https://ipfs.io/styles/img/main/alpha-sunrise-2000x1682.png Failed to load resource: net::ERR_CONTENT_LENGTH_MISMATCH
chrome-extension://boadgeojelhgndaghljhdicfkmllpafd/cast_sender.js Failed to load resource: net::ERR_ADDRESS_UNREACHABLE
html5player-new.js:309 SVG's SMIL animations (<animate>, <set>, etc.) are deprecated and will be removed. Please use CSS animations or Web animations instead.
https://static.doubleclick.net/instream/ad_status.js Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
2chrome-extension://boadgeojelhgndaghljhdicfkmllpafd/cast_sender.js Failed to load resource: net::ERR_ADDRESS_UNREACHABLE
3chrome-extension://dliochdbjfkdbacpmhlcpmleaejidimm/cast_sender.js Failed to load resource: net::ERR_ADDRESS_UNREACHABLE
3chrome-extension://hfaagokkkhdbgiakmmlclaapfelnkoah/cast_sender.js Failed to load resource: net::ERR_ADDRESS_UNREACHABLE
3chrome-extension://fmfcbgogabcbclcofgocippekhfcmgfj/cast_sender.js Failed to load resource: net::ERR_ADDRESS_UNREACHABLE
3chrome-extension://enhhojjnijigcajfphajepfemndkmdlo/cast_sender.js Failed to load resource: net::ERR_ADDRESS_UNREACHABLE
3chrome-extension://eojlgccfgnjlphjnlopmadngcgmmdgpk/cast_sender.js Failed to load resource: net::ERR_ADDRESS_UNREACHABLE
Chrome - Version 45.0.2454.101 

but then it stopped. may be worth merging to avoid this.

though i wonder what this was..... how would the length vary?

We had decided in another PR that we should set the header X-Content-Length or similar, so clients who want the info have it, but so that we also dont break things that make poor assumptions.

We should also always return correct length-- have seen it off sometimes

On Thursday, October 8, 2015, Jeromy Johnson notifications@github.com
wrote:

We had decided in another PR that we should set the header
X-Content-Length or similar, so clients who want the info have it, but so
that we also dont break things that make poor assumptions.

—
Reply to this email directly or view it on GitHub
#1776 (comment).

@wemeetagain @jbenet update here?

Not sure about the incorrect content length being sent, but I'll shortly update the PR to return the X-Content-Length header instead of simply suppressing the content length.

updated~, OP edited to reflect changes.

maybe this is a good way to avoid breaking things if the thing is incorrect header.

before merging, feedback @mappum @diasdavid ?

It has been a long time since I saw that error. Setting up Content-Length LGTM as it is the standardised header for that purpose

@diasdavid this is not using Content-Length but rather making sure we don't use it. @mappum thoughts?

I think switching to X-Content-Length is a smart move. But I'm not sure we need to provide content-length at all for this; cat shouldn't be returning it .

LGTM

I believe this has been resolved in a different PR, i no longer see both content length and chunked encoding together. If it is still an issue please reopen this PR, or file an issue with repro steps