Suggestions: Set tile from website screenshot when .png not specified

[bennypowers]

You can use http://grabz.it/api/javascript/ to generate website screenshots based on folder name or domain from .ftp-config and then set the result as project tile.

[GuilhemN]

Not sure this is a good idea from a performance point of view...

[bennypowers]

Maybe have it default to icon while loading screenshot asynchronously

On Sun, Dec 6, 2015, 12:18 PM Ener-Getick notifications@github.com wrote:

Not sure this is a good idea from a performance point of view...

—
Reply to this email directly or view it on GitHub
#40 (comment)
.

Benny Powers
Licensed Israel Tour Guide

In Israel 054-482-5726
In Canada 705-300-0072
In the USA 601-255-3532

bennypowers.com

[GuilhemN]

What do you think @Exter-N?

[Exter-N]

What do I think ?

• It would require an API key for the webservice : should one be generated and bundled with the package (might hit rate limits, might be against ToS) or should each user generate his own (initial configuration might be complicated) ?
• From a performance PoV :
  • Loading the screenshot must be done asynchronously (as the package's primary goal is to allow the user to open their project quickly, synchronously loading a remote resource which, additionally, is computationally expensive to generate would render the package quite useless) but it would make the tile ugly (and the screenshot may not be loaded in time, because, see above) ;
  • The screenshots should be cached to limit the impact of the "screenshot not loaded in time" problem, but it would create another one, namely : how long the cached screenshot should be retained ? On the beginning of a project, the front page may change quite often, but it might only seldom change after a couple months ;
• Not every server hosts one and only one website : for example I have many sites on one of my HTTP virtual machines, just one on at least two of my other HTTP VMs, but none on my Minecraft one and on my HTTP reverse proxy one ;
• Even if a server hosts one and only one website, automatically getting its HTTP(S) address is a hard (when feasible) problem : there are lots of different HTTP server packages and libraries, and the FTP(S)/SFTP address may be an internal address, different from the public one (the server might be behind a firewall or reverse proxy).