Uploading new files post migration to own s3 bucket not saving on parse.com

[kvreem]

• Server: parse server, heroku for deployment
• Database: mLab shared plan
• iOS application Parse 1.13 written in swift

Problem Occurs When:

• User who is on new infrastructure changes their profile picture
• A new user signs up on new infrastructure and uploads there own profile picture

Problem:
On my new infrastructure all new files are being stored in my s3 bucket. All users who haven't updated will not be able to see these changes. I looked at the parse.com dashboard to see what was happening and this was what I found.

The file that says "ilePicture.png" is the newly changed file. So it is updating that column in my database. However when I go to download it I get this:

For users on the new infrastructure they can get the new and old files. But for those who have not updated they cannot see the new s3 files.

[drew-gross]

Probably related to parse-community/parse-dashboard#351

[kvreem]

@drew-gross can you elaborate what they mean by use indexOf? My app is an iOS app on Parse SDK 1.13. Would the way I save the files changes in my client code? Or is this a change in parse-server itself?

[hramos]

This appears to be a duplicate of issue #1582

[kvreem]

@hramos any idea on the time frame or if a fix can be posted here on the comments? @flovilmart said he was looking at it 25 days ago

from parse-community/parse-dashboard#351

"Possible solution is to avoid the length based calculation, and use indexOf(""), and this will be a good solution for it, but needs to check first if tfss is found, since there, theres no "" separating the name from the uuid."

[flovilmart]

@kkhattab did't get a chance to have a look

[kvreem]

@flovilmart no problem, i am currently blocked from pushing the migrated version until then, my team and i found that reverting back to the previous version and trying to change the image immediately crashes the app. i can follow up again here, any estimated time frame?

[hramos]

I'll go ahead and close this, feel free to continue the discussion in #1582. I don't think the issue is being actively worked on at this time, but we'd be happy to take a look at any pull request you may send our way to address it.

[kvreem]

@hramos forgive me if i am mistaken but looking at the discussions and related issues this seems to be a big deal. Could a time range be estimated for this issue? It can be very insightful for other people who come across this issue to know. Thanks!

[kvreem]

@hramos just wanted to check up on this issue, any status updates? any alternative solutions in the mean time? thanks!