simplify: impossible, because of the structure of linearise.

Question

simplify: impossible, because of the structure of linearise.

o1lo01ol1o opened this issue 5 years ago · comments

I have a test of a postgres schema that looks a lot like the simple CRUD server in the examples. However, when I run parallel commands, the suit fails with simplify: impossible, because of the structure of linearise.. Where should I look to determine the cause of this issue?

Tim Pierson commented 5 years ago

6.0

Stevan A · Answer 1 · Wed May 29 2019 04:05:16 GMT+0800 (China Standard Time)

Which version of the library are you using?

Stevan A · Answer 2 · Wed May 29 2019 04:55:44 GMT+0800 (China Standard Time)

I fixed a related bug in: b87a74b#diff-12b81713cdb3d2bc53339c2e995fd733 , but that fix should be part of 0.6.0, so I guess there must be a similar bug still lurking there.

Can you share your test?

Tim Pierson · Answer 3 · Wed May 29 2019 05:08:52 GMT+0800 (China Standard Time)

Can you share your test?

I could potentially rename a section if you could tell me which would be most helpful; unfortunately, I can't share the code as is.

Stevan A · Answer 4 · Wed May 29 2019 05:38:11 GMT+0800 (China Standard Time)

Can you try using the feat/simplify-show-ce branch? It should give you a better error message, which hopefully will be enough to figure out what the problem is.

Tim Pierson · Answer 5 · Thu May 30 2019 00:30:45 GMT+0800 (China Standard Time)

I've been running against that branch; I can't seem to throw the error again.

However, the tests seem to proceed much more slowly. For example, previously I would generate about 1000 database inserts in a few minutes; currently I'm averaging about 1 insert a minute. (Inserts have the same frequency rating as all other queries.)

Stevan A · Answer 6 · Thu May 30 2019 16:52:04 GMT+0800 (China Standard Time)

I cannot think of any change to the library, between v0.6.0 and feat/simplify-show-ce, that would have made things that much slower.

I'm closing this, as the original problem seems to have been fixed. Feel free to open an issue about the performance degradation, but please provide a small example which illustrates the problem otherwise it will be difficult to debug.