Re: Bug: Getting dice from random.org for the second time

[Philippe Michel]

On Wed, Aug 12, 2020 at 05:09:44PM +0200, Christian Anthon wrote:

> Definitely a bug of some kind. The dice generator either runs out
> prematurely or decides that 0 is a valid rolid. I added a bit of debug code
> and it ended around 450 rolls each time and decided to return 0...

I confess ; I had been lazy and changed the batch size to 50 to check 
faster... With it back to 500 I see the same error than Peter.

After rebuilding with RANDOMORG_DEBUG set to 1, I get a popup with 
something like:

Random rolls received: 233155434624322265542633155145126431145541114223451435642
65551565515156631526333351513342233543262155255516114651145513263135353261523241
44546433422653135141444546614251431231636535313213554551424521224135253445614526
15443664451264523662522552136111142655314662256113354626334415Random rolls recei
ved: 422134663136156533445661522555423255452655414616465514653664534156643361522
45356656635145226423234544324216365235624163135555343115123661115365646244156465
442233253323236435332626562332214365152234362532556336422222616236
 
GNUbg rolls 4, You rolls 2.

4-2, not 2-3. It looks like RandomOrgCallBack() is called twice and the 
second group of digits clobbers the first one instead of being appended 
to it.

FWIW, I was a bit surprised how fast it went through "500" dice. 
Possibly, for me it failed at less than 450, either the lentgh of the 
second string or that of the longer of them, between 200 and 300.


Anyway, the issue is indeed not getting dice for the second time (it 
doesn't even get there) but dropping some of those we got initially and 
unexpectedly running out of them.

It's strange this was not noticed earlier. Maybe random.org changed 
something in the way it sends numbers, or libcurl in the way it pulls 
them, in smaller sets than it used to.