search

sanchahar / speedcrunch

Automatically exported from code.google.com/p/speedcrunch
1 stars 0 forks source link

Regression, SpeedCrunch nightly takes 2-3 times longer to start up than v0.11 #568

Closed GoogleCodeExporter closed 8 years ago

GoogleCodeExporter commented 8 years ago 
Steps to reproduce:
Install SpeedCrunch v0.11
Obtain SpeedCrunch nightly from 
https://github.com/Tey/speedcrunch-nightlies/tree/master/win32
I obtained version SpeedCrunch master-09d874ee223c0045556b2c7a307c36803dcea2aa
Open SpeedCrunch multiple times so that the EXE gets cached into RAM.
Time how long it takes to open. both the stable and nightly versions.

Expected behavior:
both versions take same amount of time to open.

Experienced behavior:
v0.11 takes less than 0.5sec on an Intel Core I5-750 CPU when using a stopwatch.
v0.11 takes 0.17sec when using ProcessExplorer to see total CPU time.

latest nightly takes just over 1.0sec on the same computer when using a 
stopwatch.
latest nightly takes 0.70sec when using ProcessExplorer to see total CPU time.

Product version:
v0.11 and master-09d874ee223c0045556b2c7a307c36803dcea2aa

Operating system:
Windows 7 64-bit

Additional Information:
Ideally, launch time should be non-noticeable to a human. v0.11 launch time is 
fast but noticeable on my PC. Other small utilities such as notepad2 open 
faster but v0.11 launch time is good enough.

Latest nightly is painstakingly slow to open, to the point I sometimes start 
typing my math and the calculator misses the beginning if i don't move my head 
up to look at the screen and make sure it finished loading.

To compound this issue, using a shortcut to open latest nightly opens up a new 
instance, it doesn't bring current instance to front. This really slows input 
speed when pinging back and fourth between screens. I have just opened a 
different bug for the new instance regression.

Original issue reported on code.google.com by alibab...@gmail.com on 9 Aug 2015 at 6:36

GoogleCodeExporter commented 8 years ago 
The Windows nightlies are absolutely experimental, unofficial and unsupported. 
We have actually not even run them yet, just built. However, it's great to have 
feedback like this, appreciate it. Please don't forget about the new issue 
tracker that I mentioned on the other issue. Thank you

Original comment by helder.p...@gmail.com on 9 Aug 2015 at 6:41

GoogleCodeExporter commented 8 years ago 
Wow, thanks very much for the incredibly fast feedback! Given the feedback I 
won't add the same items to Bitbucket as they will just be extra garbage for 
you to deal with.

I will add a note that the "Help->Send Feedback" should now point to Bitbucket 
in the latest beta. That's where I went to enter the bugs and it brought me to 
the google page.

Original comment by alibab...@gmail.com on 9 Aug 2015 at 6:46

GoogleCodeExporter commented 8 years ago 
Please add this issue on Bitbucket too, it's important that we don't forget it.
Good point on the Send Feedback link, thanks for creating an issue already.

Original comment by helder.p...@gmail.com on 9 Aug 2015 at 6:56

GoogleCodeExporter commented 8 years ago 
Sounds good. Will do.

Original comment by alibab...@gmail.com on 9 Aug 2015 at 6:57

GoogleCodeExporter commented 8 years ago 
Should be fixed for next builds. More explanations on the Bitbucket ticket:
<https://bitbucket.org/heldercorreia/speedcrunch/issues/568/regression-speedcrun
ch-nightly-takes-2-3#comment-20792667>

Original comment by te...@free.fr on 9 Aug 2015 at 11:44