Skip to content
This repository has been archived by the owner on Jan 13, 2022. It is now read-only.

I tested it with my PC on wifi network, the CSS font download time takes... #4

Closed
wants to merge 1 commit into from

Conversation

zflairz
Copy link

@zflairz zflairz commented Jun 28, 2012

... more than 100ms to complete however the timeout here is 100ms which is not reasonable, esp. considering mobile and 2G network. Let's set to 1s timeout for now and revisit the test methodology later.

-> for i in seq 1 5; do curl -k -s -o /dev/null http://rng.io/tests/cssfont/LeagueGothic.eot -w "@curl-format.txt"; done

        time_namelookup:  0.001
           time_connect:  0.059
        time_appconnect:  0.000
       time_pretransfer:  0.059
          time_redirect:  0.000
     time_starttransfer:  0.112
                        ----------
             time_total:  0.277


        time_namelookup:  0.001
           time_connect:  0.056
        time_appconnect:  0.000
       time_pretransfer:  0.056
          time_redirect:  0.000
     time_starttransfer:  0.110
                        ----------
             time_total:  0.282


        time_namelookup:  0.002
           time_connect:  0.054
        time_appconnect:  0.000
       time_pretransfer:  0.054
          time_redirect:  0.000
     time_starttransfer:  0.113
                        ----------
             time_total:  0.281


        time_namelookup:  0.001
           time_connect:  0.057
        time_appconnect:  0.000
       time_pretransfer:  0.057
          time_redirect:  0.000
     time_starttransfer:  0.123
                        ----------
             time_total:  0.358


        time_namelookup:  0.002
           time_connect:  0.055
        time_appconnect:  0.000
       time_pretransfer:  0.055
          time_redirect:  0.000
     time_starttransfer:  0.109
                        ----------
             time_total:  0.285

…kes more than 100ms to complete however the timeout here is 100ms which is not reasonable, esp. considering mobile and 2G network. Let's set to 1s timeout for now and revisit the test methodology later.


-> for i in `seq 1 5`; do curl -k -s -o /dev/null http://rng.io/tests/cssfont/LeagueGothic.eot -w "@curl-format.txt"; done

            time_namelookup:  0.001
               time_connect:  0.059
            time_appconnect:  0.000
           time_pretransfer:  0.059
              time_redirect:  0.000
         time_starttransfer:  0.112
                            ----------
                 time_total:  0.277


            time_namelookup:  0.001
               time_connect:  0.056
            time_appconnect:  0.000
           time_pretransfer:  0.056
              time_redirect:  0.000
         time_starttransfer:  0.110
                            ----------
                 time_total:  0.282


            time_namelookup:  0.002
               time_connect:  0.054
            time_appconnect:  0.000
           time_pretransfer:  0.054
              time_redirect:  0.000
         time_starttransfer:  0.113
                            ----------
                 time_total:  0.281


            time_namelookup:  0.001
               time_connect:  0.057
            time_appconnect:  0.000
           time_pretransfer:  0.057
              time_redirect:  0.000
         time_starttransfer:  0.123
                            ----------
                 time_total:  0.358


            time_namelookup:  0.002
               time_connect:  0.055
            time_appconnect:  0.000
           time_pretransfer:  0.055
              time_redirect:  0.000
         time_starttransfer:  0.109
                            ----------
                 time_total:  0.285
@rwaldron
Copy link
Contributor

Thanks for doing this research!

Before I can accept another PR, I'm obligated to have you sign the Facebook CLA, which is available here: https://developers.facebook.com/opensource/cla

w/r to the patch - ideally, I'd like to construct a system that allowed a registered callback to be executed when loading of fonts had completed. Instead of modifying the timeout duration - would you be interested in tackling this?

@rwaldron rwaldron closed this in 9cfe8ee Jul 13, 2012
@wonsuk73
Copy link

For font downloading time, how about testing it two times? First test would be same with current existing one for high bandwidth connected devices. But if it is failed, this test case tries to test again with long downloading time for low-bandwidth connected devices. What do you think?

@zflairz
Copy link
Author

zflairz commented Jul 27, 2012

@wonsuk73 you suggested a feasible solution, but the hard part is to define "high bandwidth" and "low bandwidth", and even more the definition probably would need to "upgrade" along the time. Similar to bandwidth, there are other factors which would largely impact the pass rate in ringmark, for example, there are perf tests which require the device to finish all scenarios within a predefined time, I did these tests on some low end mobile devices, most of them would fail.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants