Cut off lower number of expected clients at 0
We should avoid negative numbers in clients.csv and in per-graph CSV files based on it.
Take this graph as example:
The corresponding CSV file is:
date,users,downturns,upturns,lower,upper
2013-08-15,2419,FALSE,FALSE,1509,3400
2013-08-16,2417,FALSE,FALSE,1530,3526
2013-08-17,2583,FALSE,FALSE,1888,3593
2013-08-18,2625,FALSE,FALSE,2099,3451
2013-08-19,2631,FALSE,FALSE,1950,3865
2013-08-20,2998,FALSE,FALSE,-307,8058
2013-08-21,3429,FALSE,FALSE,-2158,12097
2013-08-22,4161,FALSE,FALSE,-5815,19493
2013-08-23,4910,FALSE,FALSE,-11092,29505
2013-08-24,5387,FALSE,FALSE,-14669,35722
2013-08-25,5899,FALSE,FALSE,-16103,38944
2013-08-26,6115,FALSE,FALSE,-18394,42370
2013-08-27,6486,FALSE,FALSE,-8958,29056
2013-08-28,7213,FALSE,FALSE,-4772,23752
2013-08-29,7950,FALSE,FALSE,-1111,19506
2013-08-30,8844,FALSE,FALSE,1660,17111
2013-08-31,9966,FALSE,FALSE,2088,19165We can just cut the lower number at 0. Like in the graph.
We could do it in write_userstats_relay_country() or in the clients module. The first is at lot easier to fix.
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information