Skip to content

Ohm 2.2.0#4

Open
cyx wants to merge 164 commits intosinefunc:masterfrom
cyx:ohm-2.2.0
Open

Ohm 2.2.0#4
cyx wants to merge 164 commits intosinefunc:masterfrom
cyx:ohm-2.2.0

Conversation

@cyx
Copy link
Collaborator

@cyx cyx commented Feb 12, 2015

No description provided.

frodsan and others added 29 commits December 25, 2013 14:30
From Ohm 2, if an attribute is set to false, it's not saved.
Related commit: soveran/ohm@717080f.
Fix Ohm::SoftDelete#restore. Remove key instead of set it to nil.
Remove Ohm::Datatypes#bool method.
Cast `deleted` attribute in Ohm::SoftDelete.
Add iconv to the gem manifest for Ruby versions >1.9.
No changes, just a matching release for ohm 2.0.0.rc2
- Rename Ohm::Locking#mutex to Ohm::Locking#spinlock
  - Yes I admit the naming has some coincidence with the HN
    thread involving linus here:

    http://yarchive.net/comp/linux/semaphores.html

  - More importantly, the rename has something to do with the
    fact that `Ohm::Model.mutex` now exists, so trying to
    extend `Ohm::Locking` into a model will result in a
    conflict.

- Fix the bug
  - So apparently, no one cares about Ohm::Locking anymore,
    and I sort of fell off the grid so I never bothered to look.
  - The bug involves not checking the value of SETNX explicitly
    as 1 / 0, but just doing the truthy check (compliant with the
    redis-rb protocol)
  - Dunno why this is using HSET instead of SET as we're dealing
    with a direct key here.

- Don't return self, as I don't remember why we even did that
  in the past. It's surprising and more often is not the
  behavior we're looking for.
- Rename Ohm::Locking#mutex to Ohm::Locking#spinlock
  - Yes I admit the naming has some coincidence with the HN
    thread involving linus here:

    http://yarchive.net/comp/linux/semaphores.html

  - More importantly, the rename has something to do with the
    fact that `Ohm::Model.mutex` now exists, so trying to
    extend `Ohm::Locking` into a model will result in a
    conflict.

- Fix the bug
  - So apparently, no one cares about Ohm::Locking anymore,
    and I sort of fell off the grid so I never bothered to look.
  - The bug involves not checking the value of SETNX explicitly
    as 1 / 0, but just doing the truthy check (compliant with the
    redis-rb protocol)
  - Dunno why this is using HSET instead of SET as we're dealing
    with a direct key here.

- Don't return self, as I don't remember why we even did that
  in the past. It's surprising and more often is not the
  behavior we're looking for.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

8 participants