Task/3068 ngsiv2 update forwarding by kzangeli · Pull Request #3480 · telefonicaid/fiware-orion

kzangeli · 2019-04-24T14:23:46Z

Implementation of NGSIv2 forwarding of updates for the request

PATCH /v2/entities/{EntityID}/attrs

…v2_update_forwarding

fgalan · 2019-04-29T08:36:56Z

@@ -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)


This line seems to be duplicated.

One line is about queries, the other about updates.
They are very similar but also completely different

I misses the "queries" vs "udpates" difference. I's ok. Sorry the noise.

NTC

fgalan · 2019-04-29T08:46:06Z

 * APPEND_STRICT
 * REPLACE

-> Side-node: The first three are "standard NGSIv1" while the second two were added for NGSIv2.


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.

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

fgalan · 2019-04-29T08:49:31Z

-  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.


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.

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

... and if there is no need of spending time, why this file was modified? ;)

I'd suggest to go back for its previous version.

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.

fgalan · 2019-04-29T08:50:33Z

@@ -2703,17 +2703,24 @@ static void searchContextProviders
  std::string                        err;


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.

Seems to be only LM_T in this file, yes

fgalan · 2019-04-29T09:04:51Z

+    TIMED_RENDER(payload = eP->toJson(NGSI_V2_NORMALIZED, nullFilter, false, nullFilter));
+
+    resource = prefix + "/entities/" + eP->id + "/attrs";
+    verb     = "PATCH";


I'm afraid this is contrary to specification #3068 (comment):

POST /v2/op/update will use "update" as actionType (to be consistent with NGSIv1)

That's the operation which has to be used to implemente NGSIv2 forwarding.

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.

fgalan · 2019-04-29T09:07:20Z

 #include <string>
 #include <vector>

+#include "logMsg/logMsg.h"


Really needed in this file? I don't see any change in this file using the things added by these includes...

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)

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).

fgalan · 2019-04-29T09:09:25Z

    shift
  fi

+  url="/notify"


This is a fix in harness suite? I mean, until this fix the --url parameter was being ignored?

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

kzangeli added 7 commits April 10, 2019 22:10

Implemented forwarded updates in NGSIv2

2a9c04c

Merge branch 'task/3068_non_legacy_notifications' into task/3068_ngsi…

4e9f7de

…v2_update_forwarding

Merge branch 'master' into task/3068_ngsiv2_update_forwarding

5496aaf

merged with master

53551dc

Pre-PR fixes

6502454

merged with master

fc0f97d

documentation

5dfea53

fgalan reviewed Apr 29, 2019

View reviewed changes

fgalan added a commit that referenced this pull request May 10, 2019

FIX partial porting of Ken's PR #3480

2b7798d

fgalan mentioned this pull request May 20, 2019

CPr NGSIv2 forwarding based in POST /v2/op/update #3489

Merged

2 tasks

mapedraza mentioned this pull request Jun 9, 2022

Request to implement pagination for request forwarding. #4149

Closed

		@@ -2703,17 +2703,24 @@ static void searchContextProviders
		std::string err;

Conversation

kzangeli commented Apr 24, 2019

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

fgalan Apr 29, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

fgalan Apr 29, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

fgalan Apr 29, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

fgalan Apr 29, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

fgalan Apr 29, 2019 •

edited

Loading

fgalan Apr 29, 2019 •

edited

Loading

fgalan Apr 29, 2019 •

edited

Loading

fgalan Apr 29, 2019 •

edited

Loading