Over the years, I've rewritten the object editor numerous times but it's pretty stable now. If we didn't have to deal with arrays, things would be much simpler! Here's the tree of components. #angular
Just had to fill in a form to request a free tier for Azure Devops parallel jobs. Apparently I need to wait 2 or 3 days for an answer. That felt odd. I guess the platform was being abused and they had to quickly restrict access 🤷♂️
Just had to fill in a form to request a free tier for Azure Devops parallel jobs. Apparently I need to wait 2 or 3 days for an answer. That felt odd. I guess the platform was being abused and they had to quickly restrict access 🤷♂️
I could build in optimizations: batch a bunch of single endpoint purges, ensure the same endpoint isn't purged twice, if a purgeall is added the the queue, clear everything first. A purge takes about 2mins so that could be factored in as well.
I need to come up with a more efficient way of invalidating the #azure cdn. If there are too many requests at the same time, it can reject the request. Maybe need some kind of queue and a single worker that processes it. 🤔 #dotnet
@traversymedia An in-house management system which I'm now using for all of my applications: @be_managed. Focus is on simplicity, rapid development and low cost hosting (serverless). Possibly a bit crazy re-inventing the wheel but it seems to have worked out. #dotnet#angular
In an effort to try and avoid content editors accidentally destroying an app by accidentally deleting a required node (think homepage) all node types are now locked down by default and you have to opt in for deletes, copies, moves, re-orders and pastes.
I want to remove NextJS as a dependency in front end apps, I think it means I need to take care of:
* Serverside rendering (inc hydration of react on client)
* Endpoints to get JSON props for each page (used after app is hydrated)
* Routing
* Serverside redirects
* Other...
Soooo, NextJS writes the QueryString values to the HTML to use with the router. Now, imagine you're caching pages using a CDN, now also imagine that you pass sensitive data in the QueryString...a password reset token for example. Guess what's cached on the CDN. 😬
Setting up another Managed system, in AWS this time. Using #Terraform to deploy Lambdas, RDS, IAM, VPC, S3 etc. It's so nice having your infrastructure as code.
Still need to get my build and deploy game on though!
A wise person once said: "Sometimes you have to go backwards in order to move forwards". It's true. But it's still frustrating. I really want to move forwards!😬
@milaw Basically, I'm trying to keep Managed as simple and as lean as possible. It was starting to feel like I was adding loads of code and libs and not getting much return.
Hard decision: after spending weeks developing a GraphQL api and an entire app that uses it, I've decided to rip it out.
Better to cut my losses early. For me, the pros just don't outweigh the cons. The major cons being the unnecessary complexity and continual hoop jumping.
@milaw Mainly problems with sending dynamic data whose shape isn't known at build time and working with interfaces. REST and JSON just seems to work better for my use cases.
Next up. Getting Azure DevOps setup for builds and deployments. I've not had a huge amount of experience with this (I use it on another project but it was originally setup by someone else). So, I'm prepped for some pain.
It feels cleaner having the crops under a separate path. Easier to backup only the originals or show just the originals when browsing storage. Need to decide today.
I have the first website going live next week with image crops. Currently, I'm storing the crops in the same folder as the originals but I'm thinking of moving all crops to their own path. That's how I originally had it and I can't remember why I changed it! 😬