ATTENTION: You are viewing a page formatted for mobile devices; to view the full web page, click HERE.

Main Area and Open Discussion > General Software Discussion

Make Firefox startup faster.

<< < (6/9) > >>

Bamse:
I was thinking the same, just about a 3rd party program trying to jump on the wagon build by this blog entry http://blog.mozilla.com/oremj/2009/08/20/speedup-firefox-with-vacuum/ Actually one before that but this shows that Placescleaner and few others do work within browser, and automatically. Also have an eye for other parameters like visited sites. Dumping those not really useful to keep. Number of days to save history should be lowered inside options of course. 90 is deault? May be 10 is ok too ;) I have tried Placescleaner, work as advertised - but I think you will see this in most "cleaner/tweakers" not just CCleaner. There is another one I can't remember name of.

May be dreaming but I think I have heard something about this getting into Firefox 4, also checking of "corruption". Could be 3.7 if that even exist. I am confused about their release schedule. Pretty sure they are aware of this though. Blog entry is from a Mozilla dude... Until then there is much traffic to gain by using magical keywords, Firefox, boost, tweak ;)

Tuxman:
Number of days to save history should be lowered inside options of course. 90 is deault? May be 10 is ok too ;)-Bamse (December 30, 2009, 11:32 AM)
--- End quote ---
These are the minimum days IIRC, so there is no notable increase...

Bamse:
Yes, 90 is default http://kb.mozillazine.org/Browser.history_expire_days_min

MilesAhead:
btw- Fodder just posted a link to the same article I already posted.-MilesAhead (December 29, 2009, 08:39 PM)
--- End quote ---
Yeah, I wrote my post before seeing new replys to the thread.

Btw, I'm one of those running my firefox profile from a ramdisk - extensions as well as internet cache etc. It does help a fair amount wrt. speed :)
-f0dder (December 30, 2009, 05:48 AM)
--- End quote ---

I thought that might be you. :)

I did reread that blog.  If you plow through the thick language what they are really saying is, it was a gimmick for WMP to hog the prefetch.  IOW, if I write an editor and handle file types such as .txt .doc .srt whatever, instead of just putting "%1" as param, if I put a bogus pararm that is different for each file type, chances are I could end up with multiple entries in the prefetch.  People would marvel how fast my editor loads compared to SuperEditWhatever brand.  One more way to give the home team an edge.

I have WMP in my kill list, just because I really don't want it indexing stuff.  Now everyone will look through their file type associations to see which apps are trying to hog the prefetch using the same trick. Another popular one is the "helper" or "agent" that sits in the tray. Just in case you may run that app someday.



f0dder:
IOW, if I write an editor and handle file types such as .txt .doc .srt whatever, instead of just putting "%1" as param, if I put a bogus pararm that is different for each file type, chances are I could end up with multiple entries in the prefetch.  People would marvel how fast my editor loads compared to SuperEditWhatever brand.  One more way to give the home team an edge.-MilesAhead (December 30, 2009, 12:41 PM)
--- End quote ---
You risk pushing other people out of the limited amount of prefix cache entries, yes (unless multiple /prefetch:n shortcuts for one app are counted as single entries, I don't know) - but I really do think this was implemented because it was believed to be a genuinely good idea, not for nefarious purposes.

That was XP, anyway Vista/Win7 SuperFetch has had a lot of changes (and works much better), so I dunno how relevant it is today :)

PS: reason I asked if SpeedyFox (and friends, for that matter) do anything but sqlite VACUUM was that it could make sense to defrag the files as well - after VACUUM handling crud and internal fragmentation, the sqlite databases can have a fair amount of filesystem fragmentation as well.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version