Bitshares-core: Dust settlement orders could mess up charts too

[HarukaMa]

Extension of issue #449 .

See this settlement order execution:
http://open-explorer.io/#/operations/1.11.112947572

which in turn produces the chart:

Should bitshares-core do something to correct this kind of data as well?

[oxarbitrage]

@abitmore what do you think ?

[abitmore]

The fill_price stored in the db might be incorrect. Will check.

{"id":"1.11.112947571","op":[4,{"fee":{"amount":0,"asset_id":"1.3.0"},"order_id":"1.8.26863","account_id":"1.2.546699","pays":{"amount":7,"asset_id":"1.3.0"},"receives":{"amount":5,"asset_id":"1.3.113"},"fill_price":{"base":{"amount":5,"asset_id":"1.3.113"},"quote":{"amount":7,"asset_id":"1.3.0"}},"is_maker":true}],"result":[0,{}],"block_num":23346221,"trx_in_block":5,"op_in_trx":1,"virtual_op":53599}

{"id":"1.11.112947572","op":[4,{"fee":{"amount":0,"asset_id":"1.3.0"},"order_id":"1.4.3719","account_id":"1.2.398303","pays":{"amount":5,"asset_id":"1.3.113"},"receives":{"amount":7,"asset_id":"1.3.0"},"fill_price":{"base":{"amount":5,"asset_id":"1.3.113"},"quote":{"amount":7,"asset_id":"1.3.0"}},"is_maker":false}],"result":[0,{}],"block_num":23346221,"trx_in_block":5,"op_in_trx":1,"virtual_op":53600}

[abitmore]

PR #571 submitted, please check.

With the patch, the object turns to:

{"id":"1.11.112947571","op":[4,{"fee":{"amount":0,"asset_id":"1.3.0"},"order_id":"1.8.26863","account_id":"1.2.546699","pays":{"amount":7,"asset_id":"1.3.0"},"receives":{"amount":5,"asset_id":"1.3.113"},"fill_price":{"base":{"amount":"29831273553","asset_id":"1.3.113"},"quote":{"amount":"50308802508","asset_id":"1.3.0"}},"is_maker":true}],"result":[0,{}],"block_num":23346221,"trx_in_block":5,"op_in_trx":1,"virtual_op":61402}

the new fill_price is (29831273553/10000)/(50308802508/100000) ~= 5.93

[oxarbitrage]

check difference in charts, the following unpatched version chart of CNY/BTS from open explorer daily(http://open-explorer.io/#/markets/BTS/CNY):

Now a patched version from pull #571 :

patch fix the charting.

[abitmore]

PR merged.

[landry314]

The new update in GUI 180201 has a much WORSE dust-trade-wick issue than the last release for some reason.

When will this merge be forked in? The sooner the better as this is a huge issue.

[oxarbitrage]

this fix is already available in the master branch of bitshares-core. we did that today and we will add a release tag tomorrow.

[oxarbitrage]

i am reopening this issue as it seems some change we made after this brought again the peaks in the charts.

i tested locally when i sent the 2 pics and the peaks were gone, now i am using current master release to update the explorer and even after a full replay(deleted chain data and started all again) i am still getting those high values:

http://open-explorer.io/#/markets/BTS/CNY -> change to 1 day

any idea what can be causing this ? i will investigate further too.

[abitmore]

I think the last change in PR #571 is actually wrong, which is made after you tested. Will create a new PR soon.

Moving to next release.

[abitmore]

Sorry, shouldn't be caused by that.

[abitmore]

@oxarbitrage I found no issue from my node as well as a public node bts.ai.la:

$ curl -d '{"id":1,"method":"call","params":["history","get_market_history",["1.3.0","1.3.113",86400,"2018-01-14T00:00:00","2018-01-14T00:00:00"]]}' https://bts.ai.la/ws;echo
{"id":1,"jsonrpc":"2.0","result":[{"id":"5.1.5834971","key":{"base":"1.3.0","quote":"1.3.113","seconds":86400,"open":"2018-01-14T00:00:00"},"high_base":245073039,"high_quote":74750000,"low_base":"57240618275","low_quote":"25444418225","open_base":226757369,"open_quote":100000000,"close_base":201110464,"close_quote":72399800,"base_volume":"5623538890791","quote_volume":"2032413699764"}]}

The peak is low_price = 10 * low_quote / low_base = 10 * 25444418225 / 57240618275 = 4.4445

But data from bitshares.openledger,info is wrong, I think it's running a really old node (maybe 171105a since it has no fill_price field)

$ curl -d '{"id":1,"method":"call","params":["history","get_market_history",["1.3.0","1.3.113",86400,"2018-01-14T00:00:00","2018-01-14T00:00:00"]]}' https://bitshares.openledger.info/ws;echo
{"id":1,"jsonrpc":"2.0","result":[{"id":"5.1.5834971","key":{"base":"1.3.0","quote":"1.3.113","seconds":86400,"open":"2018-01-14T00:00:00"},"high_base":1,"high_quote":0,"low_base":1,"low_quote":1,"open_base":11250284,"open_quote":4961375,"close_base":2045262,"close_quote":736295,"base_volume":"5623538890791","quote_volume":"2032413699764"}]}

I'll look for another node which is running 171212 for more data.

[oxarbitrage]

thanks @abitmore i am synchronizing a few chains with different versions to get more info. ill update.

[abitmore]

Found one running 171212:

$ curl -d '{"id":1,"method":"call","params":["history","get_market_history",["1.3.0","1.3.113",86400,"2018-01-14T00:00:00","2018-01-14T00:00:00"]]}' https://bit.btsabc.org/ws;echo
{"id":1,"jsonrpc":"2.0","result":[{"id":"5.1.1555910","key":{"base":"1.3.0","quote":"1.3.113","seconds":86400,"open":"2018-01-14T00:00:00"},"high_base":245073039,"high_quote":74750000,"low_base":1,"low_quote":1,"open_base":226757369,"open_quote":100000000,"close_base":201110464,"close_quote":72399800,"base_volume":"5623538890791","quote_volume":"2032413699764"}]}

The peak is low_price = 10 * low_quote / low_base = 10 * 1/ 1= 10, which is same as your data.

[oxarbitrage]

yes, hold on ... i think my binary for the witness node is in 171212, my mistake it seems ... ill confirm.

[oxarbitrage]

false alarm, the binary was not updated. sorry to make you waste time on it @abitmore

working perfect now: http://open-explorer.io/#/markets/BTS/CNY