Leftovers from 2.0

[bajtos]

List of leftovers from #276.

• trackChanges should be well documented (remove from Model)
• remove global modelBuilder
• remove loopback.createModel() in favor of Model.extend()
• rename DataModel to PersistedModel
• remove unused explorer - see Remove app.docs() #392
• remove DataModel mixins in favor of simple delegation
• DataModel methods should delegate to the dataSource directly
• refactor DataModel#replicate() into Replication class.
• refactor DataModel#bulkUpdate() into BulkUpdate class.
• refactor DataModel#diff() into Diff class.
• remove Model autoAttach
• password should be validatable
• docs for any class should include any available method (eg. inherited / mixed-in / etc)
• tests should not share any objects (eg. create a new app / modelBuilder for all tests)
• karma tests should run in ci
• beef up CHANGES.md
• create upgrade guide 1.x => 2.x
• remove app.remoteObjects()
• move memory connector filtering into its own module
• rename ACL => AccessControl and options.acls to options.acl.
• remove sharedCtor and remoted prototype methods (probably not - see sharedCtor not working - examples borked strong-remoting#73)
• document model.settings.remoting / remoting options
• document base URL setting for models
• all URLs should be lowercase / dasherized (both configurable)
• groundwork for allowing ACL to provide route names
• groundwork for allowing non-model remote classes (eg. Resource and/or Controller)
• groundwork for supporting promises via remoting
• groundwork for supporting promises for DataModel methods
• migrate app.boot() into its own separate module
• migrate ACL into its own separate module
• groundwork for removing remoting metadata from JS
• groundwork for removing juggler as a dependency (Get rid of peer dependency on loopback-datasource-juggler #275)
• remove or properly impelment "property" from ACL

List of issues assigned to 2.0 milestone but not implemented in 2.0:

• HTTP REST inconsistency: create should return HTTP 201 #360
• Related model shared = false vs. REST exposure #346
• [2.0] Programmatically defined modes vs. public/shared #345
• Permissions vs. REST find/list - not applied? #343
• [2.0] Changes not tracked in browserified models #305
• support HEAD requests to the rest api #278
• Get rid of peer dependency on loopback-datasource-juggler #275
• Remote methods mapping in JSON #209
• Bad middleware makes all requests hang #194
• app.model should report unknown options #169
• LoopBack models should inherit config.options from the parent #157
• Ability to define a bulk ALLOW once a wildcard DENY is in place. #148
• Custom/composite identifiers in client SDKs #126
• ViewModels #114
• Transaction Support #109
• Data Importing #80

[bajtos]

autoAttach cannot be removed while supporting 1.x project layout - see the comment in strongloop/loopback-boot#25.

[fabien]

Re: remove sharedCtor and remoted prototype methods

@bajtos I urge you to not drop remoted prototype methods, as discussed before. It is in heavy use whenever you want to add functionality in a sensible manner, I can assure you. Why even consider dropping this?

@altsang I thought we arrived at 2.0 yesterday, yet there's still some breaking changes listed here? What's the final ETA?

[bajtos]

@fabien I apologize for the confusion. 2.0 has definitely arrived and there won't be any breaking changes until 3.0.

This GH issue lists items there were mentioned in #276 (the pull request for 2.0 branch). My intention was to move them to a place where they are easy to find, then review them and decide which of them are still relevant. I have renamed the issue from "2.0 TODO" to "Leftovers from 2.0" to make this more clear.

@bajtos I urge you to not drop remoted prototype methods, as discussed before. It is in heavy use whenever you want to add functionality in a sensible manner, I can assure you. Why even consider dropping this?

If I remember it correctly, this item was added by @ritch. I don't recall what was the reasoning behind that. We are thinking about providing a more generic approach for creating scopes/hierarchies, prototype methods then may become a specific application of this generic approach.

[bajtos]

create upgrade guide 1.x => 2.x

The migration guide is available here: http:https://docs.strongloop.com/display/LB/Migrating+existing+apps+to+version+2.0

[fabien]

@bajtos thanks for the clarification, leftovers makes a lot more sense.

Regarding the prototype methods, @ritch chimed in on a discussion that he appreciated the heads-up concerning this functionality, and that it shouldn't be dropped just like that. The functionality it provides is something that should make it into the refactored approach to scoping/hierarchies.

[ritch]

We aren't dropping remote prototype methods.

On Jul 23, 2014, at 6:30 AM, Miroslav Bajtoš [email protected] wrote:

@fabien I apologize for the confusion. 2.0 has definitely arrived and there won't be any breaking changes until 3.0.

This GH issue lists items there were mentioned in #276 (the pull request for 2.0 branch). My intention was to move them to a place where they are easy to find, then review them and decide which of them are still relevant. I have renamed the issue from "2.0 TODO" to "Leftovers from 2.0" to make this more clear.

@bajtos I urge you to not drop remoted prototype methods, as discussed before. It is in heavy use whenever you want to add functionality in a sensible manner, I can assure you. Why even consider dropping this?

If I remember it correctly, this item was added by @ritch. I don't recall what was the reasoning behind that. We are thinking about providing a more generic approach for creating scopes/hierarchies, prototype methods then may become a specific application of this generic approach.

—
Reply to this email directly or view it on GitHub.

[seckin]

@ritch hey Ritchie, isn't this line trying to create a new object even if there is one that just needs to be updated accordingly?
2582c3f#diff-6a78f5046ecce36859baaf85c9b2b86aR475

I am trying the full-stack example with mongodb connection and bulk-update POSTs are returning HTTP500 because of this. I haven't dug into the replication system enough to really make sense out of all the moving parts but just wanted to ask here to see if there is something trivial I am missing (for instance; do I need to use the rectify changes remote function instead of bulk update to resolve this?)

[bajtos]

@seckin please open a new github issue to discuss that.

[crandmck]

I don't believe this should have the "doc" label; although some of the items are doc items, many are not. If and when we create specific doc issues from the items on the list, then those should have the "doc" label.

[bajtos]

I don't believe this should have the "doc" label; although some of the items are doc items, many are not. If and when we create specific doc issues from the items on the list, then those should have the "doc" label.

Agreed 👍

[bajtos]

Considering that this issue is over 18 months old and the issues in the list were not important enough to fix them so far, I doubt we will ever work on them. I am closing this issue.