- 25 12月, 2016 2 次提交
-
-
由 Ryuta Kamizono 提交于
Follow up to 5b14129d. http://edgeapi.rubyonrails.org/classes/ActiveRecord/Attribute.html
-
由 Akira Matsuda 提交于
-
- 24 12月, 2016 2 次提交
-
-
由 Akira Matsuda 提交于
-
由 Akira Matsuda 提交于
-
- 23 12月, 2016 2 次提交
-
-
由 Akira Matsuda 提交于
-
由 Ryuta Kamizono 提交于
`append_record` was added at 15ddd517 for not double adding the record. But adding `append_record` (checking `@target.include?(record)`) caused performance regression #27434. Instead of checking not double adding the record, add a record to target before any callbacks loads the record. Fixes #27434.
-
- 21 12月, 2016 1 次提交
-
-
由 Julia Lopez 提交于
-
- 16 12月, 2016 2 次提交
-
-
由 Kir Shatrov 提交于
Right now, when fixture is not a Hash we throw an error message saying "fixture is not a hash". This is not very user friendly because it's not saying which fixture is invalid.
-
由 Richard Monette 提交于
make sql statements frozen dup if arel is not our string expect runtime error dont wrap runtime error in invalid log errors will now be treated as runtime errors update changelog
-
- 11 12月, 2016 1 次提交
-
-
由 Ryuta Kamizono 提交于
It is enough to distinguish only the trailing `unsigned` and `unsigned zerofill`.
-
- 10 12月, 2016 2 次提交
-
-
由 Ryuta Kamizono 提交于
-
由 Stefan Budeanu 提交于
Race conditions can occur when an ActiveRecord is destroyed twice or destroyed and updated. The callbacks should only be triggered once, similar to a SQL database trigger.
-
- 09 12月, 2016 5 次提交
-
-
由 Matthew Draper 提交于
-
由 utilum 提交于
-
由 Prathamesh Sonpatki 提交于
- If aliased, then use the aliased attribute name.
-
由 Prathamesh Sonpatki 提交于
- If aliased, then use the aliased attribute name. - Fixes #26417.
-
由 Sean Griffin 提交于
While these method aren't public API, they are a non-trivial internal API which warrant a bit of explanation.
-
- 07 12月, 2016 1 次提交
-
-
由 Alex Kitchens 提交于
-
- 06 12月, 2016 10 次提交
-
-
由 Ryuta Kamizono 提交于
Raise `ActiveRecord::RangeError` when values that executed are out of range.
-
由 Ryuta Kamizono 提交于
Raise `ActiveRecord::NotNullViolation` when a record cannot be inserted or updated because it would violate a not null constraint.
-
由 Ryuta Kamizono 提交于
Actually the `primary_key` definition is not used but the inconsistency is confusing. Actual definition is `bigint auto_increment PRIMARY KEY` so `UNSIGNED` and `(8)` is unnecessary. See also #21607.
-
由 Ryuta Kamizono 提交于
-
由 Ryuta Kamizono 提交于
Because defunct wrapper class is kept for compatibility.
-
由 Ryuta Kamizono 提交于
Follow up to #27272.
-
由 Pavel Pravosud 提交于
-
由 Jon McCartie 提交于
-
由 Ryuta Kamizono 提交于
Using `:auto_increment` option for abstracting the DB-specific auto incremental types. It is worth to ease to implement the compatibility layer.
-
由 Sean Griffin 提交于
SQLite's default query interface ignores anything after the first semicolon in a query. This is actually quite common behavior in database drivers, especially when dealing with code paths for prepared statements (which we are). While this should only affect SQLite, as I'm not aware of any drivers which don't support multi-insert. Even if this does affect other third party drivers though, I'd prefer not to assume that more than one query can be executed per call to `execute`. Fixes #26948. Close #27242.
-
- 05 12月, 2016 1 次提交
-
-
由 Ryuta Kamizono 提交于
-
- 02 12月, 2016 2 次提交
-
-
由 Edgars Beigarts 提交于
-
由 Sean Griffin 提交于
PG's type map assumes that all Ruby floats are going to a field with an OID of type float4 or float8, and generates text which is invalid syntax for other types. Since the gem can handle floats properly without this encoder (albeit slightly slower), we can continue to use that as we have in prior versions of Rails. Fixes #27246
-
- 01 12月, 2016 1 次提交
-
-
由 Sean Griffin 提交于
While working on updating Paper Trail for 5.1 compatibility, I noticed that I was required to pass a second argument to `attribute`. I didn't intend for this to be the case, as `attribute :foo` is totally reasonable shorthand for "I want `attr_accessor :foo`, but also have it work with things like `.attributes` and `ActiveRecord::Dirty`"
-
- 30 11月, 2016 1 次提交
-
-
由 Sean Griffin 提交于
I had pointed the messages at the new behavior, not the old.
-
- 28 11月, 2016 2 次提交
-
-
由 Kir Shatrov 提交于
-
由 Diego Plentz 提交于
-
- 27 11月, 2016 2 次提交
-
-
由 Ryuta Kamizono 提交于
Fixes #27125.
-
由 Matthew Draper 提交于
It'll be re-cleared when it's rebuilt in with_new_connections_blocked's ensure, but we still need to clear it inside this synchronize -- we've disconnected connections that may be available in the queue, and while other threads are not allowed to make *new* connections, they are still allowed to take existing ones from there. This was incorrectly removed in d314646c.
-
- 26 11月, 2016 1 次提交
-
- 25 11月, 2016 1 次提交
-
-
由 Akira Matsuda 提交于
-
- 24 11月, 2016 1 次提交
-
-
由 Matthew Draper 提交于
Two methods block new connections; we were already doing the right thing for clear_reloadable_connections, but it's better placed in with_new_connections_blocked, where it can work for disconnect too.
-