[SERVER] Support gzip encoding instead of deflate.

[mgautierfr]

The compress function is copied from httplib

Fix #753

[codecov]

Codecov Report

Merging #757 (c52a903) into master (dc42f83) will increase coverage by 0.15%.
The diff coverage is 100.00%.

❗ Current head c52a903 differs from pull request most recent head fba0f09. Consider uploading reports for the commit fba0f09 to get more accurate results

@@            Coverage Diff             @@
##           master     #757      +/-   ##
==========================================
+ Coverage   61.09%   61.25%   +0.15%     
==========================================
  Files          58       58              
  Lines        3789     3802      +13     
  Branches     2080     2084       +4     
==========================================
+ Hits         2315     2329      +14     
+ Misses       1473     1472       -1     
  Partials        1        1              

Impacted Files	Coverage Δ
src/server/internalServer.cpp	82.54% <ø> (ø)
src/server/request_context.cpp	86.73% <100.00%> (ø)
src/server/request_context.h	100.00% <100.00%> (ø)
src/server/response.cpp	92.11% <100.00%> (+0.88%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update dc42f83...fba0f09. Read the comment docs.

[veloman-yunkan]

Since our unit tests don't check the correctness of compressed content, did you verify that the code actually works with the browsers?

[veloman-yunkan]

I wonder if this threshold is too low - why trying to compress payload that is smaller than the header section?

[mgautierfr]

Probably yes. It is a pretty old magic value and it seems it has been introduce as some kind of workaround for a segfault kiwix/kiwix-tools@4c79cfc

The question is what would be the good value ? I've seen different values but it seems that MTU is a good limit (https://www.computerworld.com/article/2693941/why-it-doesn-t-make-sense-to-gzip-all-content-from-your-web-server.html). So we can move this to 1400 Bytes.

[veloman-yunkan]

I am fine with the code but I didn't build and manually test it (since the unit tests don't exhaustively validate this change). Before merging, this functionality must be tested via kiwix-serve and the browser development console (in order to verify that kiwix-serve sends gzip-compressed content and the browser is able to decompress it).

[mgautierfr]

Before merging, this functionality must be tested via kiwix-serve and the browser development console (in order to verify that kiwix-serve sends gzip-compressed content and the browser is able to decompress it).

Done

Everything works correctly, as far as I can see on my browser

[veloman-yunkan]

Before merging, this functionality must be tested via kiwix-serve and the browser development console (in order to verify that kiwix-serve sends gzip-compressed content and the browser is able to decompress it).

Done

Everything works correctly, as far as I can see on my browser

Perfect! Why don't you merge this PR then?

[mgautierfr]

Perfect! Why don't you merge this PR then?

Because I was waiting for CI to pass after my rebase-fixup and switch to something else.