Task/3068 ngsiv2 update forwarding #3480
Conversation
…v2_update_forwarding
| @@ -1,4 +1,5 @@ | |||
| - Fix: idPattern '.*' is now allowed in NGSIv2 registrations (#3458) | |||
| - Add: Forwarded queries work for registrations using idPattern == '.*', but only for simple cases (#3458) | |||
| - Add: Forwarded updates work for registrations using idPattern == '.*', but only for simple cases (#3458) | |||
There was a problem hiding this comment.
This line seems to be duplicated.
There was a problem hiding this comment.
One line is about queries, the other about updates.
They are very similar but also completely different
There was a problem hiding this comment.
I misses the "queries" vs "udpates" difference. I's ok. Sorry the noise.
NTC
| @@ -16,7 +16,7 @@ In NGSIv1 (deprecated), the request `POST /v1/updateContext` has a field called | |||
| * APPEND_STRICT | |||
| * REPLACE | |||
|
|
|||
| > Side-node: The first three are "standard NGSIv1" while the second two were added for NGSIv2. | |||
There was a problem hiding this comment.
This file modificaiton seems to be related with fixes in the query forward documentation, but the update forwared documetnation about (FW-02 if I'm remembering correctly) has not be modified. Probably an update of the png for FW-02 is needed also, but I haven't seen yet the changes in the code so I'm not fully sure.
There was a problem hiding this comment.
As you probably can see, the modifications in this file are just corrections.
About documenting Update Forward, part of it is added, but most probably not all that is necessary
| forwarded updates do not work for registrations with idPattern (the plan is to fix this asap). | ||
| And, right now (2019-04-24), forwarding for registrations with idPatterns works for queries using `GET /v2/entities` | ||
| as well as updates using `PATCH /v2/entities/{Entity-ID}/attrs`. | ||
| More requests will be added to this list. |
There was a problem hiding this comment.
I think is better the old wording, without providing the detail of the particular operations. At least until the translation of cases/0787 test so we can have more insight in which operations are supported and which operations doesn't.
There was a problem hiding this comment.
Until we have a clear path forward, let's wait with these kind of changes. No need to spend time before we are sure what we want exactly
There was a problem hiding this comment.
... and if there is no need of spending time, why this file was modified? ;)
I'd suggest to go back for its previous version.
There was a problem hiding this comment.
I was following a perhaps unclear spec and this was my interpretation, a logical one by the way.
I did not see this as spending time for nothing.
| @@ -2703,17 +2703,24 @@ static void searchContextProviders | |||
| std::string err; | |||
There was a problem hiding this comment.
Is there any change in this file appart from LM_T LmtForward adddition? It seems so but I would like to have a confirmation from your side, pls.
There was a problem hiding this comment.
Seems to be only LM_T in this file, yes
| TIMED_RENDER(payload = eP->toJson(NGSI_V2_NORMALIZED, nullFilter, false, nullFilter)); | ||
|
|
||
| resource = prefix + "/entities/" + eP->id + "/attrs"; | ||
| verb = "PATCH"; |
There was a problem hiding this comment.
I'm afraid this is contrary to specification #3068 (comment):
POST /v2/op/updatewill use"update"asactionType(to be consistent with NGSIv1)
That's the operation which has to be used to implemente NGSIv2 forwarding.
There was a problem hiding this comment.
No really contrary to the "specification" though.
You are talking about batch ops (POST /v2/op/update). This PR only deals with the PATCH operation.
As before, once we have agreed on a way to attack the forwarding, this will either stay the same or be modified.
Before we have that agreement, nothing can be done really.
| @@ -25,6 +25,9 @@ | |||
| #include <string> | |||
| #include <vector> | |||
|
|
|||
| #include "logMsg/logMsg.h" | |||
There was a problem hiding this comment.
Really needed in this file? I don't see any change in this file using the things added by these includes...
There was a problem hiding this comment.
I probably had an LM_T and then removed it.
These things happen during PRs.
The inclusion of logMsg.h doesn't hurt (except for a millisecond extra during compilation)
There was a problem hiding this comment.
I agree it doesn't hurt, but let's try to keep files tidy. If an include isn't used by a file, then it should be removed (if I'm remembering correctly, I think we have some rules about that in style guide).
| @@ -715,6 +715,14 @@ function accumulatorStart() | |||
| shift | |||
| fi | |||
|
|
|||
| url="/notify" | |||
There was a problem hiding this comment.
This is a fix in harness suite? I mean, until this fix the --url parameter was being ignored?
There was a problem hiding this comment.
Yeah, it seems so ...
I needed to change the accumulator to accumulate the request for update forwards and as this was already implemented I decided to use it. Didn't really work though, not without this modification
Implementation of NGSIv2 forwarding of updates for the request