Improving QQBrowser detection, separating platforms some to reduce pattern count

[jaydiablo]

Fixes #1495.

Adding platforms/versions/patterns for QQ Browser to improve detection and eliminate some mis-detections I was seeing.

As mentioned in the issue, I did split the files up since the version ranges seem to differ from platform to platform. For the low end of the versions, I tried to only inlucde what I’ve seen with UAs.

I was able to reduce the pattern count some by doing this, here’s line count and pattern count for master today:

$ wc -l ../browscap/build/browscap-ua-test-1496937604/build/full_php_browscap.ini
 2367990 ../browscap/build/browscap-ua-test-1496937604/build/full_php_browscap.ini

$ ack -c "\[.+\]" ../browscap/build/browscap-ua-test-1496937604/build/full_php_browscap.ini
214169

And here’s this branch (based from master):

$ wc -l ../browscap/build/browscap-ua-test-1496937991/build/full_php_browscap.ini
 2349461 ../browscap/build/browscap-ua-test-1496937991/build/full_php_browscap.ini

$ ack -c "\[.+\]" ../browscap/build/browscap-ua-test-1496937991/build/full_php_browscap.ini
212690

Also, here’s a browser name comparison between master and this branch for about 1,600 UAs I collected that should be QQBrowser:

+------------------------+---------------------+----------------------------+
|  Expected Browser Name | browscap-php-3-full | browscap-php-3-full-master |
+------------------------+---------------------+----------------------------+
| qqbrowser (1612)       | qqbrowser (1612)    | qqbrowser (1582)           |
|                        |                     | chrome (18)                |
|                        |                     | android (6)                |
|                        |                     | safari (6)                 |
+------------------------+---------------------+----------------------------+

[codecov]

Codecov Report

Merging #1496 into master will decrease coverage by 0.02%.
The diff coverage is 32.52%.

@@             Coverage Diff              @@
##             master    #1496      +/-   ##
============================================
- Coverage     36.99%   36.97%   -0.03%     
  Complexity      816      816              
============================================
  Files          1507     1513       +6     
  Lines         56801    56878      +77     
  Branches      23637    23679      +42     
============================================
+ Hits          21015    21030      +15     
- Misses        35786    35848      +62

Impacted Files	Coverage Δ	Complexity Δ
...ents/browsers/mqq-browser/mqq-browser-android.json	62.24% <0%> (ø)	0 <0> (?)
...ents/browsers/mqq-browser/mqq-browser-generic.json	3.7% <0%> (-2.65%)	0 <0> (ø)
...gents/browsers/mqq-browser/mqq-browser-hd-ios.json	50% <50%> (ø)	0 <0> (?)
...r-agents/browsers/mqq-browser/mqq-browser-ios.json	59.25% <59.25%> (ø)	0 <0> (?)
...nts/browsers/mqq-browser/mqq-browser-mac-2-on.json	80% <80%> (ø)	0 <0> (?)
...ents/browsers/mqq-browser/mqq-browser-mac-1-x.json	80% <80%> (ø)	0 <0> (?)
...ents/browsers/mqq-browser/mqq-browser-windows.json	90.9% <90.9%> (ø)	0 <0> (?)
...ces/user-agents/browsers/edge-browser/edge-12.json	75% <0%> (-5%)	0% <0%> (ø)
...ser-agents/browsers/edge-browser/edge-generic.json	20% <0%> (ø)	0% <0%> (?)
... and 5 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 c6c6cf4...9763aea. Read the comment docs.

[mimmi20]

Was it intended to leave the version token Version/6.0 here?

[jaydiablo]

Ah probably not, it was copied from the iPhone pattern, which did contain that at the time.

I'll remove.