#readpartial is not respecting max length argument

[akelmanson]

Hi there,

According to IO#readpartial documentation, this method "reads at most maxlen bytes from the I/O stream", but HTTP::Response::Body#readpartial is returning more bytes than expected.

HTTP.get('http://www.rubydoc.info/github/httprb/http').body.readpartial(10).length

oj calls this method within C code and this misbehavior leads to Segmentation fault.

Thank you!

[ixti]

Can you please provide a simple example that will reproduce the issue? Because Body delegates #readpartial to IO at the end in fact with given maxlen.

Edit: Oh, I'm sorry misread your issue in a first place. you have example that shows that exactly...

[ixti]

I confirm issue exists:

response = HTTP.get('http://www.rubydoc.info/github/httprb/http')
5.times { puts response.body.readpartial(10).length }
# 5401
# 10
# 10
# 10
# 10

[ixti]

The issue is because we read response headers first. And once they are fully fetched some of body might happen to be written in the chunk waiting for the first readpartial call.

[tarcieri]

@bidiu please ask on the Google Group:

https://groups.google.com/forum/#!forum/httprb

[tarcieri]

@bidiu no worries, just FYI that's the proper place to ask

[bidiu]

@tarcieri Thanks!

[bidiu]

@tarcieri No need to ask now, just checked the source of Body#each, figured it out: D