ipnsfs

[whyrusleeping]

In this PR i implement a 'mock' in memory mutable filesystem to be used as the backend for the fuse ipns interface.

This is PR number 6 of the ipns set

[cryptix]

reference? to what?

[whyrusleeping]

oxDEADC0DE

[jbenet]

let's remove it

[jbenet]

@whyrusleeping network-test failed again: https://build.protocol-dev.com/job/network-test/2012/console Error: blockstore: block not found.

the common failures of network-test are:

• it actually succeeds but gets marked as failure (not sure why)
• it runs too long on a starved VM and gets killed early.

other than that network-test is reliable. this failure is not common for network-test. something is going on. either in this PR, or in code recently merged.

[whyrusleeping]

@jbenet it also failed over in @chriscool's latest PR

[jbenet]

Oh uh. This is bad news. Master may be broken. We should bisect 

—
Sent from Mailbox

On Mon, Mar 16, 2015 at 6:18 PM, Jeromy Johnson notifications@github.com
wrote:

@jbenet it also failed over in @chriscool's latest PR

Reply to this email directly or view it on GitHub:
#923 (comment)

[jbenet]

@jbenet it also failed over in @chriscool's latest PR

No, the failure in #931 is: https://build.protocol-dev.com/job/network-test/2009/console which is the "succeeded but mistakenly marked as failed."

Ok, i've tested master 10x times (on jupiter) and it failed 0/10 times.
I've tested this PR 5x times (on jupiter) and it failed 5/5 times.

There's something wrong in this PR. @whyrusleeping can you not repro the failure locally? Look at the test console output: https://build.protocol-dev.com/job/network-test/2012/console

[whyrusleeping]

Thats really weird, it looks like only client_1 ever starts doing anything..