[MRG+1] A different S3 Endpoint URL is now possible when uploading images

[Ordepsousa]

AWS_ENDPOINT_URL in Settings was read by files.py but not by images.py. I don't know if it was intentionally but i couldn't connect to my local minio S3 Server when uploading images with ImagesPipeline

[codecov]

Codecov Report

Merging #3625 into master will increase coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #3625      +/-   ##
==========================================
+ Coverage    84.5%   84.52%   +0.01%     
==========================================
  Files         167      167              
  Lines        9406     9410       +4     
  Branches     1397     1397              
==========================================
+ Hits         7949     7954       +5     
  Misses       1199     1199              
+ Partials      258      257       -1

Impacted Files	Coverage Δ
scrapy/pipelines/images.py	90.71% <100%> (+0.27%)	⬆️
scrapy/utils/trackref.py	86.48% <0%> (+2.7%)	⬆️

[kmike]

A great catch, thanks @Ordepsousa!

I wonder if you can add other missing AWS options here as well, as you're on it - it looks like AWS_REGION_NAME, AWS_USE_SSL and AWS_VERIFY are also missing.

By the way, the existing code doesn't look good at all: it mutates a global object, so one can't start two spiders via CrawlerRunner and have different AWS options for them. Also, the code is duplicated between files and images pipelines. But that're separate issues.

[Ordepsousa]

No problem @kmike :)

Added the missing AWS settings as you asked for
Yeah, it doesn't look so good, probably will need some cleaning in the future