FARR 1.13.01 launches SLOWER than 1.09.05

[keyboardy]

First let me say that FARR is one of the best tools I've encountered in a long time.

Like many, I have found the startup delay to be a bit of a problem, particularly since the first few keys I type might be lost between pressing the hotkey and the window appearing.  (Yes, I know this will mostly be resolved if I turn off history...but history is useful!)  So I intentionally wait a second after hitting the hotkey before I start typing.

Anyway, I'd been using 1.09.05 all this time.  I just upgraded to 1.13.01 (noticing that the changelog mentions that an attempt has been made to improve upon this issue), and it's DEFINITELY slower for me, meaning even more keys will be lost unless I wait even longer.

I reinstalled 1.09.05...faster again.  Reinstalled 1.13.01...slower again.  I repeated a few times, and it was always the same.

So...I'll go back to 1.09.05 and stay there for now!  Just wanted to pass this information along.  FYI, I'm using Windows XP SP2 on a 1.7GHz machine with 1.5GB of RAM.

Do you keep a link available to older versions?  I want to recommend this tool to friends, but I'm less inclined to recommend 1.13.01 than I am the 1.09.05 that I know is faster.

Thanks again for your great tool!

[lanux128]

Do you keep a link available to older versions?  I want to recommend this tool to friends, but I'm less inclined to recommend 1.13.01 than I am the 1.09.05 that I know is faster.

this is definitely mouser's call but imo, this will cause confusion as users will be posting bug reports and feature requests based on an older versions..

so keyboardy, why don't help out mouser by giving him more input to work on the improvement of the current release of Farr, escpecially in the two experimental areas..

[keyboardy]

Well, people are STILL using older versions regardless (as I was all this time).  Clearly labelling each version is usually the way to go.  Most sites tend to keep legacy versions around for just these kinds of reasons.

Anyway, I'd be happy to help!  How do I help in those areas?

[lanux128]

yes agreed, i myself test-drive new programs several times before deciding to keep them and sometimes do revert back to an older version..

and by posting here, you have taken the 1st step. just hang on until this thread appears on mouser's radar..

[mouser]

1.13.01 should definitely NOT run slower then earlier versions at all, if it does it means something's going wrong, either in code or differences in settings - i'd be very interested in tracking down what it might be.  maybe you expand a little about what you mean by "slower to startup".  one question i always ask is is there any noticable delay in if you toggle it opened and closed with break key several times in a row, beyond the first time becoming visible after a period of disuse.  And definitely check out the options lanux posted about.

[keyboardy]

Firstly, I am using the identical settings file each time.  I am installing 1.09.05 on top of 1.13.01, then 1.13.01 on top of 1.09.05, then back again, etc., and I'm making no configuration changes at any point.

Secondly, I am doing REPEATED opens and closes (I'm actually using Ctrl-space).  I'm not waiting and doing other stuff.  I ctrl-space, then ctrl-space to close it, then a second later I try again with ctrl-space to open it, etc.  So there is no period of disuse at all in my test.  And the slowness I'm reporting is ignoring the time it takes to launch the very first time, but rather I'm only talking about the time it takes for subsequent launches.

I'll say more about my observation:

With 1.09.05, when I hit ctrl-space, the top portion of the FaRR window comes up almost immediately - the part with the title bar, and the text field.  Then, about a second later, the area below that appears, and is filled in with my 9 most recently launched applications.

With 1.13.01, when I hit ctrl-space, nothing appears immediately.  About a second later, the entire window (title bar, text field, and recent applications) appears all at the same instant.

I hope that observation helps point things in the right direction.

Regarding the options lanux posted about...aha, I hadn't understood that those are options now present in 1.13.01.  I will reinstall 1.13.01 and see if that helps!

[keyboardy]

And definitely check out the options lanux posted about.

-mouser (December 11, 2006, 01:57 AM)

OK, the options were:
1. Try to keep search directories in system cache for faster search
2. Try to keep program in memory for faster wakupe

#2 was already checked.  I checked #1.  No change.  Reverted back to 1.09.05, and startup is faster again.

[lanux128]

keyboardy, take a look here & maybe you can get a few pointers.. Startup Speed

[mouser]

keyboardy, can you send me your 1.09.05 ?  just the main program exe zipped up would be fine ([email protected]).

i believe you that there is a noticable delay in 1.13.01, but i think it must be related to something weird going on.

perhaps you can help me by trying this.  backup any settings you want to preserve first, and then try experimenting with telling it NOT to show the previously launched file history at start up.  see if that is where the problem is.  if so, try turning it back on but clearing that list of previously launched files and see if it makes a difference..  see if you have any files in that list that are on a network drive or something.

are you testing these on a clean install of the program or are your old settings files with old history and search paths being used.  if you are using your old history you might try backing up your settings and then deleteing your settings files and starting a completely clean install and comparing the two versions.

what operating system are you using by the way?

ps. if you want to email me we can try to figure out the problem over email which might be a bit faster ([email protected]).

[nontroppo]

To test if there is some difference in resource usage, I highly recommend Process monitor by SysInternals - you can compare the registry / process/ file demands when you start both to compare them "objectively". I haven't measured the differences and would be interested, though I do not find the latest version of FARR slower for me.

Oh, an make sure your settings are identical - if you have slightly different search paths that could substantially cange your performance...