undefined method `[]' for nil:NilClass about delayed_paperclip HOT 23 CLOSED

I have the same problem after updating to Paperclip 3.5.2

Stack :
ruby 2.0.0p247
rails 3.2.16
paperclip 3.5.2
delayed_job_active_record 4.0.0

The problem is in delayed_paperclip/attachment.rb:16

@instance.class.paperclip_definitions[@name] now returns nil

I submitted a quick fix for this issue : #67

Can we have a comment from the author on this ?

from delayed_paperclip.

I feel like just nil checking it would be covering up a deeper issue. This method https://github.com/jrgifford/delayed_paperclip/blob/master/lib/delayed_paperclip.rb#L74 may have been broken with the latest Paperclip update.

from delayed_paperclip.

It might have been. I'll need to look, that method might be the root of some other problems people are experiencing...

from delayed_paperclip.

any word?

from delayed_paperclip.

I just merged in #67. I'll be aiming for a release in the next week or so with it.

from delayed_paperclip.

Does this code need to be kept in delayed_paperclip? It looks like it's covering up a problem with STI (as fixed in thoughtbot/paperclip#1331, not yet released in gem form).

from delayed_paperclip.

One more followup to this: using the latest code from paperclip master fixes the problem on 2.6.1 (and allows STI classes to work properly), so I think this patch should be reverted.

from delayed_paperclip.

@aripollak Thanks for getting to the bottom of this! @jrgifford agree that we should revert it?

from delayed_paperclip.

Sure. I might get to this today, but if someone does first, I'd be open to
merging it.

from delayed_paperclip.

I have a pull request that fixes this issue. See here:
#73

from delayed_paperclip.

Fixed by #73.

from delayed_paperclip.

Did you have plan to release a version on rubygems with this fix?

from delayed_paperclip.

I'm looking to release one shortly.

from delayed_paperclip.

Version 2.7 released.

from delayed_paperclip.

Thanks a lot

from delayed_paperclip.

It looks like the code from #67 is still in master, which is not necessary (and masks the problem) since the bug was fixed in paperclip.

from delayed_paperclip.

@aripollak your right. Can you make a PR to remove it please. Only the code that's necessary should exist in the repo

from delayed_paperclip.

Could you just run git revert 8c1c12b?

from delayed_paperclip.

We don't rewrite git history once it's in a published gem.

On Tuesday, March 4, 2014, Ari Pollak [email protected] wrote:

Could you just run git revert 8c1c12b?

Reply to this email directly or view it on GitHubhttps://github.com//issues/66#issuecomment-36702647
.

Scott Carleton
Artsicle, Co-Founder / CTO
[email protected]
C. 914 400 5085

from delayed_paperclip.

git revert doesn't rewrite history; it's not reset or rebase. https://www.kernel.org/pub/software/scm/git/docs/git-revert.html

from delayed_paperclip.

I'm still getting this error with last master... but I don't understand which revert is needed

git revert 8c1c12b

won't cleanly merge into master, and that check for nil is not done anymore.

from delayed_paperclip.

Oh whoops, you're right @deivid-rodriguez, I was looking at the wrong repository, so nothing further needs to happen on this issue. If you're still seeing this error, you might need to upgrade paperclip or delayed_paperclip (the latter due to #73).

from delayed_paperclip.

I'm running latest versions of both. I'll try to give more information if I find it again.

from delayed_paperclip.