As mentioned in #619, #1840, and #2064, the automated tests are missing a handful of data points. From ticket:1840#comment:6, most of the useful data points such as batch delay, quality, et al aren't included in the CSV. As far as I can tell, there's no error or traceback printed when the tests run xpra info
, so it's likely that the Xpra Info page has changed, as is mentioned. The test code will likely need to be updated to reflect these changes.
I'm pretty sure the data collection breakage is fairly recent, whereas these tickets are not, so you should be able to test the version they were completed for:
Wow, this whole thing is in terrible shape. Minor cleanup in r21958. Bug fix in r21959. Now at least some of the tests do run.
Now the real problem with the missing data is the same thing that required fixes to the session info dialog (r21340).
As of r21960, we capture the batch delay again, more fixes needed.
As of r21962, the actual batch delay is back (but now this is a normalized value... and maybe this should not be compared?), as well as the latency values.
Still need to restore:
r21998 reverts some of the "actual batch delay" changes from r21960 + r21962 so that the values will be more in line with older versions.
Should now be usable again and the data should make sense.
tests data from browser test
I attached some data from some browser tests. Still seems to be some missing data points.
download limit (KB), upload limit (KB), latency (ms) all read 0 for example
Can you take a look and let me know?
Still seems to be some missing data points. download limit (KB), upload limit (KB), latency (ms) all read 0 for example
Those are test configuration options for trickle. They should be showing up if you enable it.
Confirmed the missing data does show up when you turn on the trickle options.
this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/2112