* The approach is to compile <% %> into a method call that checks whether
the value returned from a block is a String. If it is, it concats to the buffer and
prints a deprecation warning.
* <%= %> uses exactly the same logic to compile the template, which first checks
to see whether it's compiling a block.
* This should have no impact on other uses of block in templates. For instance, in
<% [1,2,3].each do |i| %><%= i %><% end %>, the call to each returns an Array,
not a String, so the result is not concatenated
* In two cases (#capture and #cache), a String can be returned that should *never*
be concatenated. We have temporarily created a String subclass called NonConcattingString
which behaves (and is serialized) identically to String, but is not concatenated
by the code that handles deprecated <% %> block helpers. Once we remove support
for <% %> block helpers, we can remove NonConcattingString.
Obviously #key is a too common name to be included
in the AMo interface, #to_key fits better and also
relates nicely to #to_param. Thx wycats, koz and
josevalim for the suggestion.
AR's #to_key implementation now takes customized
primary keys into account and there's a testcase
for that too.
The #to_param AMo lint makes no assumptions on how
the method behaves in the presence of composite
primary keys. It leaves the decision wether to
provide a default, or to raise and thus signal to
the user that implementing this method will need
his special attention, up to the implementers. All
AMo cares about is that #to_param is implemented
and returns nil in case of a new_record?.
The default CompliantObject used in lint_test
provides a naive default implementation that just
joins all key attributes with '-'.
The #to_key default implementation in lint_test's
CompliantObject now returns [id] instead of [1].
This was previously causing the (wrong) tests I
added for AR's #to_key implementation to pass. The
#to_key tests added with this patch should be
better.
The CI failure was caused by my lack of knowledge
about the test:isolated task. The tests for the
record_identifier code in action_controller are
using fake non AR models and I forgot to stub the
#to_key method over there. This issue didn't come
up when running the test task, only test:isolated
revealed it. This patch fixes that.
All tests pass isolated or not, well, apart from
one previously unpended test in action_controller
that is unrelated to my patch.
This commit introduces two new methods that every
AMo compliant object must implement. Below are the
default implementations along with the implied
interface contract.
# Returns an Enumerable of all (primary) key
# attributes or nil if new_record? is true
def key
new_record? ? nil : [1]
end
# Returns a string representing the object's key
# suitable for use in URLs, or nil if new_record?
# is true
def to_param
key ? key.first.to_s : nil
end
1) The #key method
Previously rails' record_identifier code, which is
used in the #dom_id helper, relied on calling #id
on the record to provide a reasonable DOM id. Now
with rails3 being all ORM agnostic, it's not safe
anymore to assume that every record ever will have
an #id as its primary key attribute.
Having a #key method available on every AMo object
means that #dom_id can be implemented using
record.to_model.key # instead of
record.id
Using this we're able to take composite primary
keys into account (e.g. available in datamapper)
by implementing #dom_id using a newly added
record_key_for_dom_id(record)
method. The user can overwrite this method to
provide customized versions of the object's key
used in #dom_id.
Also, dealing with more complex keys that can
contain arbitrary strings, means that we need to
make sure that we only provide DOM ids that are
valid according to the spec. For this reason, this
patch sends the key provided through a newly added
sanitize_dom_id(candidate_id)
method, that makes sure we only produce valid HTML
The reason to not just add #dom_id to the AMo
interface was that it feels like providing a DOM
id should not be a model concern. Adding #dom_id
to the AMo interface would force these concern on
the model, while it's better left to be implemented
in a helper.
Now one could say the same is true for #to_param,
and actually I think that it doesn't really fit
into the model either, but it's used in AR and it's
a main part of integrating into the rails router.
This is different from #dom_id which is only used
in view helpers and can be implemented on top of a
semantically more meaningful method like #key.
2) The #to_param method
Since the rails router relies on #to_param to be
present, AR::Base implements it and returns the
id by default, allowing the user to overwrite the
method if desired.
Now with different ORMs integrating into rails,
every ORM railtie needs to implement it's own
#to_param implementation while already providing
code to be AMo compliant. Since the whole point of
AMo compliance seems to be to integrate any ORM
seamlessly into rails, it seems fair that all we
really need to do as another ORM, is to be AMo
compliant. By including #to_param into the official
interface, we can make sure that this code can be
centralized in the various AMo compliance layers,
and not be added separately by every ORM railtie.
3) All specs pass
