Adds support for statically typed insert statements. #133
Conversation
See the new test in `PreparedStatementsTest` for an example of the problem being fixed. Previously the test would fail on the call to `setLong` because the prepared statement result from the stub server did not contain column type information.
|
I have fixed the failing tests by changing the This seems pretty terrible to me, but I'm beginning to suspect fully supporting priming columns in both ways together would require some architectural changes or at the least a breaking change in the API. If anyone has any better ideas on how to accomplish this please let me know. |
|
I have added support for asserting on activity. I am still very unhappy with this solution, but it gets the job done and fixes my use case. I would love to see a better solution, but I really do need something that works more than I need clean code at this juncture. :/ |
|
Bleh. My fix for asserting on activity doesn't actually work because the map.values() doesn't give the results in a well defined order, so you can't reliably assert on it. :( |
…mn metadata. I am not very happy with the solution, but I lack the context for the overall architecture of this project to fix this problem in a better way. Right now I suspect that things would fall apart if you tried to use both the column metadata and variable types.
|
I am now using this branch for my Cassandra testing since, even with the caveats, it is better than nothing. What are the chances of this getting merged into mainline? If they are low, then I'll look into making my fork a bit more permanent (setup CI, etc.). If they are high I'll continue waiting. |
|
thanks for your work on this @Zoltu I should have time to look at this properly at the weekend |
|
Closing this per the comment on the issue |
See the new test in
PreparedStatementsTestfor an example of the problem being fixed. Previously the test would fail on the call tosetLongbecause the prepared statement result from the stub server did not contain column type information.Fixes #132
This is not ready for merge, but I am a bit stuck. The original bug has been fixed, but some tests are now failing and I don't fully understand why (nor do I understand what they were testing). In particular, it appears the column_types were working for
selectstatements and now they aren't. By not sending back column information in the prepare result (just commenting out thecolumns.forEach) fixes the old tests... but of course breaks the new test.