run Gittip 94

[chadwhitacre]

93

Brought to you from Momi Toby's Revolution Café & Art Bar in Hayes Valley, where DPR Ross Ulbricht allegedly ran Silk Road.

[chadwhitacre]

Here's a Quora answer that links to a paywalled WSJ article as its source for the Momi Toby's connection, and here's a quote from a local news story (cached because the original is 500 for me right now):

The FBI complaint listed Momi Toby's Revolution Cafe & Art Bar on Laguna Street in Hayes Valley as one of the locations Ulbricht worked from.

Regular customers at the café told KTVU they recognized the photo of Ulbricht.

He’d just sit, drinks coffee. Talked to people I guess, said San Francisco resident Stephen Toole. "I never spoke to the guy ever."
One man who didn’t want to appear on camera said they sometimes made small talk.

"He's a normal, everyday guy," the man said. "An average guy who was very unassuming."

Here's the FBI investigation. Section 41 (pp 27-28) describes "an Internet café on Laguna Street in San Francisco, California" where DPR logged into a Silk Road server on June 3, 2013, and how this café is "less than 500 feet" from a residential address on Hickory Street where Ross Ulbricht logged into his Gmail account (Google, Comcast, and "a certain server-hosting company" having been subpoened for an IP address trail).

Based on my [Christopher Tarbell, Special Agent, FBI] training and experience, this evidence places the administrator of Silk Road, that is, DPR, in the same approximate geographic location, on the same day, as Ulbricht.

[chadwhitacre]

Okay! Now that that's out of the way ... :-)

[chadwhitacre]

Marked suspicious:

https://www.gittip.com/ApoideaTheorem/

[chadwhitacre]

Unmarked after further review.

[chadwhitacre]

Reviewed 46 accounts.

[chadwhitacre]

Backup taken and verified. Droplet spun up and updated.

[chadwhitacre]

Script started (3 hours ago). Just starting payout loop.

[chadwhitacre]

sigh

pid-1197 thread-140375692850944 (MainThread) Did payout for 27548 participants.
pid-1197 thread-140375692850944 (MainThread) Traceback (most recent call last):
pid-1197 thread-140375692850944 (MainThread)   File "/home/whit537/www.gittip.com/gittip/cli.py", line 32, in payday
pid-1197 thread-140375692850944 (MainThread)     Payday(db).run()
pid-1197 thread-140375692850944 (MainThread)   File "/home/whit537/www.gittip.com/gittip/billing/payday.py", line 147, in run
pid-1197 thread-140375692850944 (MainThread)     self.db.self_check()
pid-1197 thread-140375692850944 (MainThread)   File "/home/whit537/www.gittip.com/gittip/models/__init__.py", line 18, in self_check
pid-1197 thread-140375692850944 (MainThread)     self._check_balances()
pid-1197 thread-140375692850944 (MainThread)   File "/home/whit537/www.gittip.com/gittip/models/__init__.py", line 91, in _check_balances
pid-1197 thread-140375692850944 (MainThread)     assert b == 0, "conflicting balances: {}".format(b)
pid-1197 thread-140375692850944 (MainThread) AssertionError: conflicting balances: 1

[chadwhitacre]

Log downloaded.

Droplet destroyed.

[chadwhitacre]

Tweet tweeted on Twitter.

[chadwhitacre]

MassPay run for 35 users and posted back to Gittip.

[chadwhitacre]

Escrow shuffled.

[chadwhitacre]

Leaving open to investigate conflicting balances traceback.

[chadwhitacre]

Picking up with the balance conflict.

I've located the user in question.

[chadwhitacre]

The user's balance manually checks out on their history page. Is there a bug in the self-check?

[chadwhitacre]

Got it! There is a bug in self-check.

=> select 0 as a union select -30 as a union select -30 as a union select 70.26 as a; 
┌───────┐
│   a   │
├───────┤
│   -30 │
│     0 │
│ 70.26 │
└───────┘
(3 rows)

[chadwhitacre]

In the unions, two of exactly the same value are collapsed into a single instance of that value. So, in the above example, instead of 0 - 30 - 30 + 70.26, we end up with 0 - 30 + 70.26.

Any chance you can take a look at this, @zwn?

[zwn]

@whit537 I will try.

[rummik]

Weird. MySQL does the same thing. I've never run into that before >.>;

[rummik]

select 0 as a union all select -30 as a union all select -30 as a union all select 70.26 as a;

   a   
-------
     0
   -30
   -30
 70.26
(4 rows)

[zwn]

@rummik Nice, thanks. The documentation agrees with the reality http://www.postgresql.org/docs/9.3/static/queries-union.html 😄

[chadwhitacre]

Self-check bug reticketed as #2191.