You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
OutgoingMessage should not allow writing more bytes than Content-Length #39041
I don't think we currently error if the user specifices a content-length header and then writes more bytes to the response. I think this is something we could detect and error. Some http clients have a hard time handling this case (especially with pipelining) and would be nice if we could help with reducing the probability of this occuring.
if this will be a perf drop, what about a config flag for strict validation?
also this assumes that we know the size of the payload before sending it :) which from exp may not always be the case.
I think its the RECEIVER of the content that should validate the size against the headers. :)
Hi everyone !!!!!
Good Day!!
Its my first time contributing to an open source project.
Wanna know if this issue is still being worked upon.
And any insights for this issue would be very helpful.
Hey everyone,
I am looking forward to contribute in this , so can anyone help me on how to start with contributing as I have just started the open source contribution .
So please share the steps of how to start contributing here .
Activity
ronag commentedon Jun 15, 2021
@nodejs/http wdyt?
mcollina commentedon Jun 15, 2021
I'm in favor if this is not costing much too much throughput.
I think we should add a new option to enable the check (as a minor), and then make a breaking change to flip that default.
betochf commentedon Jun 16, 2021
@mcollina @ronag hey, I'm just getting started, do you mind if I try to pull this off?
daukadolt commentedon Jun 16, 2021
Looks like an interesting ticket, good luck with it @betochf 👍
Please do let me know though in case if I can take over 🙈
jodevsa commentedon Oct 30, 2021
@betochf Are you still on it?
manav014 commentedon Dec 11, 2021
May I start working on this issue? As I think I would be able to create a PR for the same.
manav014 commentedon Dec 13, 2021
@ronag may please help me with a specific test case which I can add as a test and also I can use that to test my changes.
mpodolsk commentedon Jan 3, 2022
if this will be a perf drop, what about a config flag for strict validation?
also this assumes that we know the size of the payload before sending it :) which from exp may not always be the case.
I think its the RECEIVER of the content that should validate the size against the headers. :)
there are very specific conditions when to send this header.
https://datatracker.ietf.org/doc/html/rfc7230#section-3.3.2
If we consider request payloads "inputs" its the receiving system that should validated those 🤔
LoveSharmaDEV commentedon Feb 26, 2022
Hi everyone !!!!!
Good Day!!
Its my first time contributing to an open source project.
Wanna know if this issue is still being worked upon.
And any insights for this issue would be very helpful.
74ankur09 commentedon Apr 1, 2022
Hey everyone,
I am looking forward to contribute in this , so can anyone help me on how to start with contributing as I have just started the open source contribution .
So please share the steps of how to start contributing here .
sidwebworks commentedon Aug 23, 2022
Hey @ronag is this issue still available for someone to take? I noticed some previously closed PRs due to some broken tests.
Can I take this one and get this resolved?
mcollina commentedon Aug 23, 2022
Go for it!