format idAttribute on save/delete (#1676)

[osher]

format idAttribute on save/delete (#1676)

• Related Issues : .format(attr) is not applied to where clause on update?  #1676

Introduction

Working with a DB and Web that do not agree on field-names conventions - trying to leverage Model#format() and Model#parse() worked for selects and inserts - but fails for udpates and deletes.

Motivation

This PR fixes that. Closes #1676, Closes #1338.

Proposed solution

use Model#format to transform the id attribute name passed to .where in delete and update operations.

Current PR Issues

none

Alternatives considered

none

[rapzo]

I'll give a round on this, although, we could move the format up a bit instead of running it twice, even for a single key, it looks harder to read than it'd be with a single call.

[osher]

Lost you there, you mean just to do it in two lines instead of one, or create a more reusable method and call it from there?

LMK - I'll comply

[osher]

P.S - I met more bugs around the same problem which appears in different places around delete and updates, although I'm yet sure if the root cause is in Bookshelf itself, or in bookshelf-modelbase plugin.

I tend to go for creating a method that provides a selector/filter/where clause that should be used wherever a key to a table is needed.

Not just a private function, but a part of the API.
If all implementations will pull table key from this API - it will also make supporting compound keys in the future easy peasy

[osher]

anything?

[ricardograca]

I'm not that sure about increasing our dependency on .format() and .parse(), but I'm not contributing much code to the project anyway, so maybe someone more active can comment on this issue.

[ricardograca]

This is very similar to #1338. Can you explain why your solution is better? Also, can you provide test cases for the issues this fixes?

[osher]

Yup. looks very similar. I think mine uses less operations, but it's so negligible that it doesn't worth the discussion.
I don't care which of them you take, as long as the issue is fixed 👍 😉

[ricardograca]

Great. This just needs a test case then and I'll merge it.

[osher]

I hear you. I'll find time for it soon...

…

On Tue, Feb 6, 2018 at 8:06 PM, Ricardo Graça ***@***.***> wrote: ***@***.**** requested changes on this pull request. Great. This just needs a test case then and I'll merge it. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#1680 (review)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAxBHTJYDWiob8l30RHVrObLl5hlbK2dks5tSJSfgaJpZM4QPVXz> .

[osher]

ok, done. There you go 😃
@ricardograca - I think the ball is in your court again

Updated: Sorry it took so long, I misread the picture...

[osher]

@ricardograca
Uh...
Is there something I have to do to handle the change-request except the commits I pushed?

[ricardograca]

@osher Sorry, I'll take a look later today. Currently busy with something else.

[osher]

+1 👍