Open jbcr opened 5 years ago
You are right.
In fact, looking in the code, I see that this feature was enabled in the past, then commented out - but forgotten in the docs as 'available'.
The reason is found in a comment in the code: disallowing matches by remote_id allows us to implement KeyMatcherInterface without the risk of users getting confused as to what they are matching...
In other words: there are places where a match for a ContentType is done using a single value which can be either an integer (id) or a string (ct identifier). This is slightly different from, say, matching Contents or Locations, where matching on a single value will either look for id (for integers) or for remote_id (for strings). The current thinking is that, if we allow to match CTs by remote_id in some places, developers might expect to be matching on remote_id by default when matching a string value, instead of matching on identifier.
Since there now seems to be at least a user who is asking for matching on remote_id, I am open to revisit this decision. A possible compromise would be to keep matching on identifier if a single string value is passed, and match on remote_id only when the user expressly asks for it - and make sure that this is well documented in the docs...
Another answer is update the documentation because the identifier
field is like remote_id
(the same on each environment).
As you wish... do you need yo use the remote_id
field or is identifier
enough?
In my experience, usage of remote_id
is only necessary when sync with a remote system is at play...
@gggeek In my case, the identifier
is the same in all environment and is not used on remote system.
Well, I have solved my problem with the field identifier
.
But, I like use the remote_id
because is generaly it's unique.
In the DSL doc, I can use the content class
remote_id
in the match section into my migration.But, if I want execute the migration, I have this error: