Remove hardcoded AWS bucket Name from a Few Files #5799
Conversation
pr-triage
bot
added
PR: reviewed-approved
There was a problem hiding this comment.
What a tedious tasks, thanks for doing this @mstruve! I know you said you're going to do views in a later PR, but the static public/offline.html could potentially also go in this one.
There was a problem hiding this comment.
Hi @mstruve, thanks a lot for surfacing this but I don't think this is the right approach to fix the issue.
I believe we should download these images and add them to app/assets/images with the others, so that we don't have hardcoded paths pointing to dying buckets or non harcoded paths pointing to potentially non existing images
| @@ -5,7 +5,7 @@ class ProfileImage | |||
| def initialize(resource) | |||
| @resource = resource | |||
| @image_link = resource.profile_image_url | |||
| @backup_link = "https://thepracticaldev.s3.amazonaws.com/i/99mvlsfu5tfj9m7ku25d.png" | |||
| @backup_link = "https://#{ApplicationConfig['AWS_BUCKET_NAME']}.s3.amazonaws.com/i/99mvlsfu5tfj9m7ku25d.png" | |||
There was a problem hiding this comment.
wouldn't this break local installations when merged?
For example:
- the local
AWS_BUCKET_NAMEis notthepracticaldevby default, it'sOptional - mine is called
devto-testsbecause I configured it to point to my AWS installation - this would mean that I'd have to manually upload a file named
99mvlsfu5tfj9m7ku25d.pngfor this to work
I think it'd have an impact on the other communities as well.
Maybe we could keep this hardcoded until we find a better way to handle this:
- either we download these images and put them in our assets (my fav option)
- or we find a way to upload them when the app is setup, through a script or something
Basically I don't think these images should be in AWS at all :)
| @@ -2,7 +2,7 @@ module HtmlCssToImage | |||
| AUTH = { username: ApplicationConfig["HCTI_API_USER_ID"], | |||
| password: ApplicationConfig["HCTI_API_KEY"] }.freeze | |||
|
|
|||
| FALLBACK_IMAGE = "https://thepracticaldev.s3.amazonaws.com/i/g355ol6qsrg0j2mhngz9.png".freeze | |||
| FALLBACK_IMAGE = "https://#{ApplicationConfig['AWS_BUCKET_NAME']}.s3.amazonaws.com/i/g355ol6qsrg0j2mhngz9.png".freeze | |||
There was a problem hiding this comment.
same as the previous observation
| @@ -14,7 +14,7 @@ class SiteConfig < RailsSettings::Base | |||
| field :social_networks_handle, type: :string, default: "thepracticaldev" | |||
|
|
|||
| # images | |||
| field :main_social_image, type: :string, default: "https://thepracticaldev.s3.amazonaws.com/i/6hqmcjaxbgbon8ydw93z.png" | |||
| field :main_social_image, type: :string, default: "https://#{ApplicationConfig['AWS_BUCKET_NAME']}.s3.amazonaws.com/i/6hqmcjaxbgbon8ydw93z.png" | |||
There was a problem hiding this comment.
he're we're definitely breaking the main social image for a pristine installation
| @@ -46,7 +46,7 @@ def remove_profile_info | |||
| twitch_url: nil, feed_url: nil | |||
| ) | |||
|
|
|||
| user.update_columns(profile_image: "https://thepracticaldev.s3.amazonaws.com/i/99mvlsfu5tfj9m7ku25d.png") | |||
| user.update_columns(profile_image: "https://#{ApplicationConfig['AWS_BUCKET_NAME']}.s3.amazonaws.com/i/99mvlsfu5tfj9m7ku25d.png") | |||
There was a problem hiding this comment.
this is further evidence we might want to rethink this approach, we are storing a broken path in the DB:
- it's either a broken URL for new installers
- or a URL that points to non existent images if the bucket name is not
thepracticaldev
pr-triage
bot
added
PR: reviewed-changes-requested
|
@rhymes you are right, a lot of these images should be downloaded and put in the assets folder. Since there is a bit more scope to this than I anticipated I am going to close and make an issue to get it sorted out. |
|
@mstruve went ahead and put them all haahaahha |
What type of PR is this? (check all applicable)
Description
This is only a small number of the places we have hardcoded our old AWS bucket name. I figured I would do these and then make an Issue to handle all of the ones that are in the view and emails which are A LOT and will need some closer attention as some are cloudinary links and some point straight to S3.
Related Ticket
Followup for #5771 and #5555
Added to documentation?