| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| | | | | | | | | |
|
| |\ \ \ \ \ \ \ \ |
|
| | |\ \ \ \ \ \ \ \
| | | | | | | | | | |
| | | | | | | | | | | |
CI with Ruby 2.5.0
|
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
Refer https://github.com/ruby/bigdecimal/pull/86
|
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
Address #517
|
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
Refer https://github.com/rubygems/rubygems/issues/2123
|
| | |/ / / / / / / /
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
https://www.ruby-lang.org/en/news/2017/12/25/ruby-2-5-0-released/
|
| | |\ \ \ \ \ \ \ \
| | | | | | | | | | |
| | | | | | | | | | | |
Lateral expressions for PostgreSQL
|
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
Support for PostgreSQL lateral expressions. This is treated as an unary
function applied to a query expression.
Lateral is a separate function to provide interoperability with aliases
and unions. These are also separate node types that wrap
SelectStatements. The lateral option would need to be implemented in
these nodes separately if lateral was an option of SelectStatement.
When building the query, an alias can be given as an argument. This
enables building a lateral query with an table alias without using
either Nodes::TableAlias or Nodes::Lateral directly.
|
| | | | | | | | | | | |
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
Remove Unused `require`
|
| | |/ / / / / / / / /
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
It looks like they are left from old design
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
Reduce `Reduce`
|
| | |/ / / / / / / / /
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
Remove `Arel::Visitors::Reduce` because it almost completely duplicates
`Arel::Visitors::Visitor`
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
Introduce Expression Node
|
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | | |
This requires a little cut and paste from the Binary node, but it is
used in different parts of sql
|
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | | |
SQL is very powerful. Many nodes can be used as a sub expression or
query. grouping all of these possible nodes together
|
| | |\ \ \ \ \ \ \ \ \ \
| | | |/ / / / / / / / /
| | |/| | | | | | | | | |
Explicitly mentions the #and operator
|
| | |/ / / / / / / / / |
|
| | | | | | | | | | | |
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
Better titles
|
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | | |
* s/crazy/advanced/ to avoid ableist words
* last sections didn't have titles to separate topics
I have written a bit more examples of "less trivial" Arel code here, perhaps we can add some of that here:
https://github.com/rstacruz/cheatsheets/pull/92/files
|
| | |\ \ \ \ \ \ \ \ \ \
| | | |/ / / / / / / / /
| | |/| | | | | | | | | |
Allow count nodes to have math functions
|
| | |/ / / / / / / / / |
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | | |
jcoleman/fix_incorrect_typecasting_of_raw_sql_strings
Type-castable attributes should not try to cast SqlLiteral nodes
|
| | |/ / / / / / / / / |
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
Unused variable
|
| | | | | | | | | | | | |
|
| | |\ \ \ \ \ \ \ \ \ \
| | | | | | | | | | | | |
| | | | | | | | | | | | | |
Support BindParams in subqueries
|
| | |/ / / / / / / / / /
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | | |
When an Arel AST contains a SelectManager as a component, e.g. as a CTE
expression, the SelectManager is converted to SQL with `#to_sql`. This
uses a new collector that leads to invalid expressions when using
bind parameters. For example, when generating PostgreSQL queries, the
bind parameter number starts from one again. When using the
SubstituteBinds collector, binds in the subquery are not substituted.
This commit changes the ToSql visitor to visit the SelectManager ast
itself.
|
| | |\ \ \ \ \ \ \ \ \ \
| | | |/ / / / / / / / /
| | |/| | | | | | | | | |
Update .travis.yml
|
| | |/ / / / / / / / / |
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
Add required_ruby_version to gemspec
|
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | |
| | | | | | | | | | | | |
travis.yml made it follow Rails (AR). The summary is as follows.
- Drop MRI 2.2.1 or lower and updated with latest MRI versions
- Drop rbx-2. This build has failed since before
- Update JRuby version to 9.1.12.0
|
| | |/ / / / / / / / / |
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
Remove encoding utf-8 magic comment
|
| | |/ / / / / / / / / |
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
Remove Unused variable - offset_bind
|
| | |/ / / / / / / / /
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
Fixes warning in Rails tests:
```
/home/travis/build/rails/rails/vendor/bundle/ruby/2.4.0/bundler/gems/arel-b9ca36f09d5e/lib/arel/visitors/oracle.rb:32: warning: assigned but unused variable - offset_bind
``
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
CI against ruby 2.4.1
|
| | | | | | | | | | | | |
|
| | |\ \ \ \ \ \ \ \ \ \
| | | |/ / / / / / / / /
| | |/| | | | | | | | | |
Oracle visitor needs two more `add_bind` argument fix
|
| | |/ / / / / / / / /
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
Here `offset` is bound twice. It used to be `:b1` appeared twice.
This commit changes the second `offset` name changed from `:b1` to `:b3`
which should not affect the actual bind values for offset.
|
| | | | | | | | | | | |
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
Add missing hash, eql?, == to various node classes
|
| | |/ / / / / / / / /
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
Some of the nodes classes are missing either one or many of the common comparison methods #hash, #eql? and #==.
This makes comparision and working with the ast sometimes painful, as equality or operations likes array differences (which uses a hash behind the scene) produces unexpected results.
A test has been added that ensures that every descendants of Node:
* have all 3 methods
* that all 3 methods were defined from the same class
* that the class defining all 3 is also a descendant of Node, to avoid the default ones that rely on identity only
|
| | |\ \ \ \ \ \ \ \ \
| | | | | | | | | | | |
| | | | | | | | | | | | |
Remove unused `bind_values`
|
| | |/ / / / / / / / /
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
Since https://github.com/rails/rails/commit/213796f, `bind_values` is no
longer used from Active Record.
|
| | | | | | | | | | | |
|
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | |
| | | | | | | | | | | |
We need `value` to have a writer for `StatementCache` to work when
prepared statements are disabled. This is something I'd like to revert
eventually, either by disabling that form of caching in that case or
re-introducing something like the old `Bind` collector.
The addition of `nil?` is to make `IS NULL` be inserted correctly,
similar to what we already do with quoted and casted nodes
|