Tiny Core Linux

Tiny Core Extensions => TCE Bugs => Topic started by: vinnie on March 28, 2011, 07:33:49 AM

Title: minefield4
Post by: vinnie on March 28, 2011, 07:33:49 AM
This is not very bug but, with stable version was lost symlink /usr/local/bin/minefield4
Title: Re: minefield4
Post by: Jason W on March 28, 2011, 04:41:31 PM
I was not aware that previous builds had that symlink.
Title: Re: minefield4
Post by: vinnie on March 29, 2011, 04:51:07 AM
I thought the symlink was yours, but instead was mozilla, there was no doubt because I used it in the settings of pidgin.
Title: Re: minefield4
Post by: vinnie on April 23, 2011, 09:10:31 AM
communication service, from last update minefield4 no longer loads the HTML5 video, however firefox.tcz does it, and also inherits the settings.
Title: Re: minefield4
Post by: Jason W on April 23, 2011, 09:53:50 AM
Hey, can you give a link to the video that won't play.
Title: Re: minefield4
Post by: jls on April 23, 2011, 10:07:08 AM
any youtube video, first go to http://www.youtube.com/html5
Title: Re: minefield4
Post by: Jason W on April 23, 2011, 10:25:44 AM
Thanks

I am in the process of passing on the Firefox extensions to Matt Causey, who did the latest minefield4.tcz build, and I have alerted him.     
Title: Re: minefield4
Post by: curaga on April 23, 2011, 11:09:29 AM
Heh, I see the new minefield4 was done in the spirit of my old build :) Even with the same text.

Since it seems that a lightweight build has made a comeback, perhaps there's need for a custom "full" build, with all the video and other bells and whistles.
Title: Re: minefield4
Post by: Jason W on April 23, 2011, 12:42:57 PM
Actually, my builds of minefield were done in a similar spirit, disabling updater, installer, dbus/gnome.  As well as also using Swiftfox patches as a guide to set default cache size and other things.  And in addition, mine also disabled the urlclassifier which creates that ever growing sqlite file.  But mine did depend on libasound, where the original minefield.tcz didn't. 

So mine was not exactly a full build though it was pretty much fully functional in audio/video.  I remember discussing whether do disable or leave in features with Matt, and he seems to prefer a more stripped down build than what I produced.  At the time I advised to leave the features in as most folks would want them, but perhaps there is room for 2 minefields, one like I did or maybe adding in some stuff, and one stripped down.  Or we may just want the custom build to be fully stripped as the official Firefox provides all the bells and whisltes, we will see.

I will get with Matt and see what he thinks about it.
Title: Re: minefield4
Post by: curaga on April 23, 2011, 01:11:18 PM
Oh, sorry for not getting that right. I only tend to use FF when forced to (when Opera doesn't work ;)).

Seeing that Matt's build depends on libasound, I wonder what broke video then.
Title: Re: minefield4
Post by: Jason W on April 23, 2011, 01:22:04 PM
It's all good.  I had used your original minefiled.tcz as a guide on building the later versions, though I am sure some things have crept in along the way.  And as Firefox is such a memory hog, using 320MB just to start and go to Facebook one time, I would be very interested in any reduction in resource usage in a sparser build.

 
Title: Re: minefield4
Post by: vinnie on April 23, 2011, 02:28:40 PM
Thank you for the support  :D
Title: Re: minefield4
Post by: Jason W on April 23, 2011, 10:54:44 PM
Just for memory footprint comparison, I tested the Firefox extensions by opening them in a terminal to www.google.com.  I was astonished to see how each release of Firefox is fattening up in terms of RAM use, as we see Opera in that same time has hardly increased at all.  Here are the numbers, I am posting them to help in deciding whether it would be worth maintaining 2 custom Firefox builds, or would there really even be enough of a difference in RAM use by the further stripping of features.  Even launched in safe mode, Firefox uses the same amount of RAM as in normal mode.  Matt can decide the feature set of minefield4, and we can then determine if we need another build.

Code: [Select]
granparadiso (firefox 3.0.x)              97mb
shiretoko    (firefox 3.5.x)             136mb
namoroka     (firefox 3.6.x)             160mb
minefield4   (firefox 4.0)               217mb
firefox.tcz  (official firefox 4.0) 238mb
swiftfox     (version 3.6.13)            177mb
opera    (version 9.64, browser only)     74mb
opera    (version 11.10, browser only)    92mb
midori       (version 0.3.3)             109mb
Title: Re: minefield4
Post by: curaga on April 24, 2011, 02:30:56 AM
How did you measure?
Title: Re: minefield4
Post by: Jason W on April 24, 2011, 07:41:59 AM
Quick and dirty with top.  Perhaps not the best way, though I don't have time today to delve into it further.  If top 'VSZ' does not give a good idea, I can later use another method.
Title: Re: minefield4
Post by: curaga on April 24, 2011, 11:33:39 AM
VSZ also lists all libs, which while still useful doesn't take into account the sharing possible. If that's the only gtk2 app you run, then they are not shared, but otherwise they shouldn't count against the app.

What I use is "pmap -d $pid | tail -n1" (the private column). Others have also recommended the improved top apps.


For example, top VSZ lists 21mb for bash for me, while pmap says it's only 850kb. pmap -d without tail lists where the other parts are used, how much which lib takes etc.
Title: Re: minefield4
Post by: tinypoodle on April 24, 2011, 01:10:05 PM
Always wondered why top would show VSZ only and not RSS as I remember it from older version on other systems (busybox top as well).
Checking on this occasion here, I found that 'top -m' would spit out a lot more of detailed info related to memory usage.   :)

One thing I can not understand why with 'top -m' the X server PID seems to be absent...
Title: Re: minefield4
Post by: tinypoodle on April 24, 2011, 01:11:59 PM
Just for memory footprint comparison, I tested the Firefox extensions by opening them in a terminal to www.google.com.  I was astonished to see how each release of Firefox is fattening up in terms of RAM use, as we see Opera in that same time has hardly increased at all.

Perhaps it might be of interest also to include seamonkey in the comparison.
Title: Re: minefield4
Post by: Jason W on April 24, 2011, 01:19:03 PM
Thanks for the info, since I posted some numbers I will use a more accurate method so the numbers are not misleading.
Title: Re: minefield4
Post by: Jason W on April 27, 2011, 02:48:32 PM
I have tried several times without success to contact the new maintainer of minefield4.  I will let the standard week or a little more pass since first attempting to communicate before I reinstate the previous build.  I know that life happens and I want to give the new maintainer ample opportunity to respond to attempts at communication before I take back over maintenance.    
Title: Re: minefield4
Post by: Jason W on April 28, 2011, 05:52:57 PM
I have been in contact with Matt, and he is out of town right now.  He will fix the html5 issue when he returns.
Title: Re: minefield4
Post by: vinnie on April 28, 2011, 07:59:03 PM
thanks jason ;)
Title: Re: minefield4
Post by: Jason W on April 29, 2011, 05:26:58 PM
Since 4.0.1 is released and there is an issue with the current one, I will provide an updated build until the maintainer is available to take over.
Title: Re: minefield4
Post by: Jason W on April 29, 2011, 09:25:42 PM
I am uploading minefield4.tcz version 4.0.1, but changing to 24 depth color seemed to fix the html5 vidoe issue with the current 4.0 extension.

EDIT:  24 bit with the current minefield4.tcz still seems to pose html5 video problems that the update to 4.0.1 does not exhibit.