Use a different user-agent header for debug/test runs (#fix 1269)

[jrichter1]

so that the http requests for testing purposes can be filtered out in the statistics. Feel free to edit the agent string itself if you don't like it :)

[codecov-io]

Codecov Report

Merging #1271 into master will decrease coverage by 0.51%. The diff coverage is 80%.

@@            Coverage Diff             @@
##           master    #1271      +/-   ##
==========================================
- Coverage   93.42%   92.91%   -0.52%     
==========================================
  Files          43       43              
  Lines        4624     4697      +73     
  Branches      511      804     +293     
==========================================
+ Hits         4320     4364      +44     
- Misses        304      310       +6     
- Partials        0       23      +23

Impacted Files	Coverage Δ
browser/model/helpers/downloader.js	95.53% <ø> (-2.26%)	:arrow_down:
browser/services/request.js	97.61% <0%> (-2.39%)	:arrow_down:
browser/model/installable-item.js	92.8% <100%> (-0.6%)	:arrow_down:
browser/model/devstudio-autoinstall.js	97.29% <0%> (-2.71%)	:arrow_down:
browser/model/che.js	95% <0%> (-1.62%)	:arrow_down:
browser/pages/account/controller.js	98.61% <0%> (-1.39%)	:arrow_down:
browser/model/cdk.js	98.71% <0%> (-1.29%)	:arrow_down:
browser/model/devstudio-p2-zip.js	64.4% <0%> (-1.12%)	:arrow_down:
browser/model/devstudio-p2.js	81.66% <0%> (-1.1%)	:arrow_down:
... and 23 more

---

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 24daa00...578b35e. Read the comment docs.

[jrichter1]

Updated the agent string. The build is red though, need to merge #1276 to fix that.

[jrichter1]

test this