Lucas_Werkmeister_WMDE added a comment.

  > So after gzip, we almost cut the init module in half (☺) but bloat the app 
module by more than that amount (☹).
  
  It occurred to me that a lot of this might be due to the fact that the init 
module is more minified than the app module, so here’s another table with even 
more numbers.
  
  | file                 | size before | size after |
  | -------------------- | ----------- | ---------- |
  | init                 | 92 kB       | 58 kB      |
  | init, gzip           | 29 kB       | 17 kB      |
  | init, uglifyjs       | 91 kB       | 58 kB      |
  | init, uglifyjs, gzip | 29 kB       | 17 kB      |
  | app                  | 323 kB      | 429 kB     |
  | app, gzip            | 72 kB       | 97 kB      |
  | app, uglifyjs        | 104 kB      | 141 kB     |
  | app, uglifyjs, gzip  | 32 kB       | 43 kB      |
  |
  
  (Numbers obtained by running `for file in dist/data-bridge.{init,common}.js; 
do wc -c < "$file"; gzip < "$file" | wc -c; uglifyjs -c -m < "$file" | wc -c; 
uglifyjs -c -m < "$file" | gzip | wc -c; done | numfmt --to=iec` on master and 
my branch. kB means 1000 bytes, gzip is v1.9, uglifyjs is uglify-es 3.0.19.) So 
after uglification, init shrinks by 33 kB and app grows by 37 kB, a 4 kB 
difference; after uglification and gzip, init shrinks by 12 kB and app grows by 
11 kB, a 1 kB difference in the other direction.
  
  While there are probably some rounding errors in there and this might vary by 
gzip/uglify settings (you could also use a different uglifier), in general this 
suggests to me that my change does not, in fact, bloat the app module with code 
conjured out of thin air. A roughly equivalent size of code seems to have moved 
from the init to the app module – it just looked otherwise because it moved 
from a more-compressed module to a less-compressed one.
  
  (I don’t quite remember if the app module being less optimized/compressed was 
a conscious decision we made or just a consequence of how we set things up. In 
theory, it’s nice to be able to see the original code in the browser debugger 
when using ResourceLoader debug mode, but I admit I haven’t actually made use 
of that yet. Perhaps we should compress the app module after all?)

TASK DETAIL
  https://phabricator.wikimedia.org/T235771

EMAIL PREFERENCES
  https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: Lucas_Werkmeister_WMDE
Cc: Lucas_Werkmeister_WMDE, Aklapper, Pablo-WMDE, Hook696, Daryl-TTMG, 
RomaAmorRoma, 0010318400, E.S.A-Sheild, darthmon_wmde, Meekrab2012, 
joker88john, Michael, DannyS712, CucyNoiD, Nandana, NebulousIris, Gaboe420, 
Versusxo, Majesticalreaper22, Giuliamocci, Adrian1985, Cpaulf30, Lahi, Gq86, 
Af420, Darkminds3113, Bsandipan, Lordiis, GoranSMilovanovic, Adik2382, 
Th3d3v1ls, Ramalepe, Liugev6, QZanden, LawExplorer, WSH1906, Lewizho99, 
Maathavan, _jensen, rosalieper, Scott_WUaS, Wikidata-bugs, aude, 
Lydia_Pintscher, Mbch331
_______________________________________________
Wikidata-bugs mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs

Reply via email to