Joshua Reed
Joshua Reed
https://github.com/mkleehammer/pyodbc/issues/489 I see this exact error randomly instead of the garbled messages in issue 489. I believe this is a result of getting an error message half full of arbitrary...
It looks to me like Formatter is a class in his local helper lib. When you run pip install Formatter it fails as formatter was only compat with Python 2....
Further in reviewing... Since as Victor mentions: ```self.conn.setdecoding(pyodbc.SQL_WMETADATA, encoding="utf-8")``` Fixes the issue, I doubly think this is related to #489 The reason being if you half the size of a...
Also he mentions it's random. The system error which is thrown occurs randomly as the characters that are pulled (when they shouldn't be) and ran through unicode_decode are random. So,...
Finally, can we get some motion on #489? Would a PR to repair #489 be more well received if I was able to provide a reproducible test case that proves...
Yes please! I would love a compat flag or the like. ATM, there isn't much of a workaround as trying to catch and better parse the issue only works sometimes....
At the least if I could get back a raw message in some way I can easily enough reduce the length by two. Unfortunately on occasion the UnicodeDecodeError is thrown...
Can confirm that removing the `*2` that is `*sizeof(ODBCCHAR)` from the below errors.cpp snippet works around this as I just removed it and installed from source. So, how can we...
@v-chojas ... I have upgraded to the latest. Compiled unixODBC from source. This bug is still present. Simply rebuilding pyodbc with the *sizeof(ODBCCHAR) removed from: https://github.com/mkleehammer/pyodbc/blob/master/src/errors.cpp#L295 solves this, but I'd...