PG set stringtype=unspecified to handle string/uuid edge case#1087
PG set stringtype=unspecified to handle string/uuid edge case#1087imustafin wants to merge 7 commits intojruby:61-stablefrom
Conversation
| end | ||
|
|
||
| def extract_raw_bind_values(binds) | ||
| binds.map(&:value_for_database) |
There was a problem hiding this comment.
Is this something only missing for postgresql?
There was a problem hiding this comment.
I'm really not sure. This change is not really relevant to the purpose of this PR.
Also, #1088 (comment) reports that this should not be an issue in Rails 6.1.4
There was a problem hiding this comment.
Ah yeah I guess we will not worry about that one. I am trying to get our CI working again. Looks like some changes in bundler and retrieving via git was messed up. Once I get it ok I will rerun this.
|
I updated this against 61-stable since against master I am seeing a lot of errors which most likely is a problem with rails HEAD. |
|
@imustafin unfortunately we have a few consistent errors but this particular one looks like it added some new one related to prepared arity mismatch. Any chance stringtype is involved here? @dr-itz If the above is not really an issue for this PR do you have an opinion here? Also @kares :) https://app.travis-ci.com/github/jruby/activerecord-jdbc-adapter/jobs/527824796 |
|
@enebo which tests are now failing comparing to 61-stable? I can't seem to find them. I'll take a look at this PR (hopefully by Monday). Do you plan to make CI green soon on 61-stable? |
|
@imustafin in the link above it is the tests like this: I did not see these in the head build for 61-stable. |
Co-authored-by: Karol Bucek <kares@users.noreply.github.com>
|
@imustafin yeah. I just merged @kares review comment since it makes sense to honor settings. This will start a new CI run and we can see if we still see those couple of errors with this change or they are intermittent. |
|
What is the status of this PR? |
Fixes #1086.
I'm not sure if changing stringtype is a good idea, but tests pass. I am now testing these changes in our projects.
I think that standard Rails with pg is working similarly by letting postgres to infer the needed types.