Optimize rendering of big lists

MobSF / Mobile-Security-Framework-MobSF

Mobile Security Framework (MobSF) is an automated, all-in-one mobile application (Android/iOS/Windows) pen-testing, malware analysis and security assessment framework capable of performing static and dynamic analysis.

https://opensecurity.in

GNU General Public License v3.0

16.68k stars 3.16k forks source link

Optimize rendering of big lists #2351

Closed JPSxzy8 closed 3 months ago

JPSxzy8 commented 4 months ago

Describe the Pull Request

Most browsers doesn't like rendering thousands/millions of lines if not needed, so wrap (potentially) big lists in <details> and auto-open only if they are small (currently below 100 items)

This makes the "static report" page much more responsive for big apps:

Checklist for PR

[x] Run MobSF unit tests and lint tox -e lint,test
[x] Tested Working on Linux, ~Mac, Windows, and Docker~ (this PR only adds cosmetic changes to HTML)
[x] Add unit test for any new Web API (Refer: StaticAnalyzer/tests.py) (this PR only adds cosmetic changes to HTML, so no new tests)
[x] Make sure tests are passing on your PR

ajinabraham commented 4 months ago

Nice! I will test this and get back.

ajinabraham commented 3 months ago

This is such an impactful PR in improving the performance of rendering. Thanks for sending it across @JPSxzy8

ohyeah521 commented 3 months ago

I think the response mechanism for static analysis should be further optimized. I’ve noticed that when requesting the static analysis page, sometimes the server returns more than 10MB of data, which is mostly strings and such. When the frontend makes requests, the server dumps all the data at once to the frontend, but users may not be concerned with all the data, which can cause excessive server load.

ajinabraham commented 3 months ago

We hope to split up reporting and use AJAX to load them in async. But we are not there yet.

ohyeah521 commented 3 months ago

this is a long-term goal, but unfortunately, my front-end skills are too weak to contribute code. I hope this feature can be implemented soon.