dchuk dchuk

## concurrency-in-go.md

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                dchuk
                / concurrency-in-go.md
            
            
              Created
              March 12, 2016 01:52
                — forked from kachayev/concurrency-in-go.md
            
              
                Channels Are Not Enough or Why Pipelining Is Not That Easy
              
          
    Channels Are Not Enough

... or Why Pipelining Is Not That Easy
Golang Concurrency Patterns for brave and smart.
By @kachayev
Intro


## gist:beca213533a5289ac5eab891d139fc1a
Spree API Server Needs
- Easy Setup & Deployment (Dokku)
- New Relic
- Error Monitoring (Raven)
- Performance Monitoring (Skylight)
- Logging (Loggly or Logentries?)
- SSL (Dokku handles this)
- Database Backups (http://donpottinger.net/blog/2014/11/22/bye-bye-heroku-hello-dokku-part-2.html (See Clockwork section))
- Caching (Memcached + Dalli)
- Worker (Sidekiq)

## prefetch.js
// This code is to be used with https://turbo.hotwire.dev. By default Turbo keeps visited pages in its cache
// so that when you visit one of those pages again, Turbo will fetch the copy from cache first and present that to the user, then
// it will fetch the updated page from the server and replace the preview. This makes for a much more responsive navigation
// between pages. We can improve this further with the code in this file. It enables automatic prefetching of a page when you
// hover with the mouse on a link or touch it on a mobile device. There is a delay between the mouseover event and the click
// event, so with this trick the page is already being fetched before the click happens, speeding up also the first
// view of a page not yet in cache. When the page has been prefetched it is then added to Turbo's cache so it's available for
// the next visit during the same session. Turbo's default behavior plus this trick make for much more responsive UIs (non SPA).
	Spree API Server Needs
	- Easy Setup & Deployment (Dokku)
	- New Relic
	- Error Monitoring (Raven)
	- Performance Monitoring (Skylight)
	- Logging (Loggly or Logentries?)
	- SSL (Dokku handles this)
	- Database Backups (http://donpottinger.net/blog/2014/11/22/bye-bye-heroku-hello-dokku-part-2.html (See Clockwork section))
	- Caching (Memcached + Dalli)
	- Worker (Sidekiq)
	// This code is to be used with https://turbo.hotwire.dev. By default Turbo keeps visited pages in its cache
	// so that when you visit one of those pages again, Turbo will fetch the copy from cache first and present that to the user, then
	// it will fetch the updated page from the server and replace the preview. This makes for a much more responsive navigation
	// between pages. We can improve this further with the code in this file. It enables automatic prefetching of a page when you
	// hover with the mouse on a link or touch it on a mobile device. There is a delay between the mouseover event and the click
	// event, so with this trick the page is already being fetched before the click happens, speeding up also the first
	// view of a page not yet in cache. When the page has been prefetched it is then added to Turbo's cache so it's available for
	// the next visit during the same session. Turbo's default behavior plus this trick make for much more responsive UIs (non SPA).