No more rpi engine downloads
Moderators: FourthWorld, heatherlaine, Klaus, kevinmiller, robinmiller
-
- VIP Livecode Opensource Backer
- Posts: 3581
- Joined: Mon Jan 22, 2007 7:36 am
- Location: Berkeley, CA, US
- Contact:
No more rpi engine downloads
I'm sad to see that the Raspbery Pi builds have disappeared from the download site.
PowerDebug http://powerdebug.ahsoftware.net
PowerTools http://www.ahsoftware.net/PowerTools/PowerTools.irev
PowerTools http://www.ahsoftware.net/PowerTools/PowerTools.irev
-
- VIP Livecode Opensource Backer
- Posts: 9857
- Joined: Sat Apr 08, 2006 7:05 am
- Location: Los Angeles
- Contact:
Re: No more rpi engine downloads
The RPi builds were released under GPL v3, and were removed along with the other open source offerings.
If there's an upside it's that the build never really worked, so the loss is minor.
If there's an upside it's that the build never really worked, so the loss is minor.
Richard Gaskin
LiveCode development, training, and consulting services: Fourth World Systems
LiveCode Group on Facebook
LiveCode Group on LinkedIn
LiveCode development, training, and consulting services: Fourth World Systems
LiveCode Group on Facebook
LiveCode Group on LinkedIn
-
- Posts: 722
- Joined: Thu Sep 11, 2014 1:49 pm
- Location: The Netherlands
Re: No more rpi engine downloads
It is a bummer that LiveCode does not support Raspberry-Pi platform 100%.
I personally think LC is the perfect development Language for R-Pi . Especially now that R-Pi has made a big boost in performance.
Ik would love to have the app I am building now to be able to run on a R-Pi.
I personally think LC is the perfect development Language for R-Pi . Especially now that R-Pi has made a big boost in performance.
Ik would love to have the app I am building now to be able to run on a R-Pi.
-
- Livecode Opensource Backer
- Posts: 9454
- Joined: Fri Feb 19, 2010 10:17 am
- Location: Bulgaria
Re: No more rpi engine downloads
Well, you can continue to use the Open Source version.Ik would love to have the app I am building now to be able to run on a R-Pi.
The fact that LiveCode have stopped supporting the Open Source LiveCode offerings does not
mean they have suddenly and mysteriously stopped working: only, that when they go "kabumf"
LiveCode are not going to accept responsibility for that. But, as I have never had any of the Open
Source versions of LiveCode go "kabumf" on me, I don't entirely understand the fuss re support.
In fact, considering the vast amount of work the people at LiveCode central put into the Open Source
versions over an 8 year period I cannot quite understand why people should suddenly stop using them.
I understand that with the upcoming LiveCode 10 there will be all sorts of new stuff; and if one wants
that one has to pay and/or go cap in hand. And that there will, probably, be no further Raspberry-Pi versions of LiveCode.
But the Open Source versions of LiveCode are still available for download (due, largely, to the generosity of
LiveCode central) and capable of being used.
After all I still use ClarisWorks/AppleWorks on a PPC iMac running MacOS 10.5 on a regular basis:
the fact that Steve Jobs is dead, and if I sent a message to Apple that my PPC iMac had gone "kabumf"
because of ClarisWorks they probably wouldn't bother to reply. doesn't stop it being a good, not-overly cluttered
office suite.
-
- VIP Livecode Opensource Backer
- Posts: 3581
- Joined: Mon Jan 22, 2007 7:36 am
- Location: Berkeley, CA, US
- Contact:
Re: No more rpi engine downloads
Must be my faulty memory then. I thought the LC6.x builds were pre-open-source.
PowerDebug http://powerdebug.ahsoftware.net
PowerTools http://www.ahsoftware.net/PowerTools/PowerTools.irev
PowerTools http://www.ahsoftware.net/PowerTools/PowerTools.irev
-
- Livecode Opensource Backer
- Posts: 9454
- Joined: Fri Feb 19, 2010 10:17 am
- Location: Bulgaria
Re: No more rpi engine downloads
Here: https://community.livecode.com/
There are Open Source downloads available back to 6.0.1, but no RPI downloads
as far as I can ascertain.
There are Open Source downloads available back to 6.0.1, but no RPI downloads
as far as I can ascertain.
-
- VIP Livecode Opensource Backer
- Posts: 3581
- Joined: Mon Jan 22, 2007 7:36 am
- Location: Berkeley, CA, US
- Contact:
Re: No more rpi engine downloads
I do see the LC7.0.0->7.0.4 rpi builds there, but (from memory) I think the LC6 build was more stable albeit with fewer bells and whistles.
PowerDebug http://powerdebug.ahsoftware.net
PowerTools http://www.ahsoftware.net/PowerTools/PowerTools.irev
PowerTools http://www.ahsoftware.net/PowerTools/PowerTools.irev
-
- VIP Livecode Opensource Backer
- Posts: 9857
- Joined: Sat Apr 08, 2006 7:05 am
- Location: Los Angeles
- Contact:
Re: No more rpi engine downloads
I don't recall the first build date, but IIRC all the RPi downloads they made available were licensed under GPL v3.
If you find an exception it would be good to know, but likely inconsequential: given their uniqueness proprietary licenses are difficult to digest, and given their role they're usually subject to on-the-fly reinterpretation (often through blanket clauses like "at the sole discretion of Company").
Richard Gaskin
LiveCode development, training, and consulting services: Fourth World Systems
LiveCode Group on Facebook
LiveCode Group on LinkedIn
LiveCode development, training, and consulting services: Fourth World Systems
LiveCode Group on Facebook
LiveCode Group on LinkedIn
Re: No more rpi engine downloads
Indeed. For people building human machine interface control systems - that's me - it's too bad there's no rpi engine that actually works, because LC + rpi would make a great industrial control station. Writing to a field repeatedly leaks memory, and if you're writing clock or positional data at high frame rate to lots of fields, you'll eat memory and crash within a few hours to days.
I would buy 9.6.6, but it's busted for OS X with a similar field memory leak, just like 9.6.3 OS X, although it's a slower leak! Just write the time or positional data to a dozen or so fields at high rate, like you would in a telescope control interface, and you'll see memory consumed. You can speed this up to make it really visible by writing in a repeat forever loop, but it happens regardless, and has nothing to do with using repeat. In 9.6.3 OS X it's plainly severe - after a couple of hours memory went from 30M to 445M writing the seconds to 9 fields. In 9.6.6 OS X it's still a killer, going from 30M to 74M in the same time. All this happens on both Yosemite and Big Sur. I see 9.6.3 OS X leaking memory writing to 9 fields at 5 Hz in my production app with data acquired by URL, so I can't use it.
Maybe I'll try this when LC releases 10, and in the meantime I'm stuck in 6.1.1 for OS X. It's not about the money: I'd pay for a working rpi or OS X engine!
This OS X memory leak on 9.6.3 does not happen in Windows! For some reason the Windows standalone has always been more reliable than Linux or OS X.
I would buy 9.6.6, but it's busted for OS X with a similar field memory leak, just like 9.6.3 OS X, although it's a slower leak! Just write the time or positional data to a dozen or so fields at high rate, like you would in a telescope control interface, and you'll see memory consumed. You can speed this up to make it really visible by writing in a repeat forever loop, but it happens regardless, and has nothing to do with using repeat. In 9.6.3 OS X it's plainly severe - after a couple of hours memory went from 30M to 445M writing the seconds to 9 fields. In 9.6.6 OS X it's still a killer, going from 30M to 74M in the same time. All this happens on both Yosemite and Big Sur. I see 9.6.3 OS X leaking memory writing to 9 fields at 5 Hz in my production app with data acquired by URL, so I can't use it.
Maybe I'll try this when LC releases 10, and in the meantime I'm stuck in 6.1.1 for OS X. It's not about the money: I'd pay for a working rpi or OS X engine!
This OS X memory leak on 9.6.3 does not happen in Windows! For some reason the Windows standalone has always been more reliable than Linux or OS X.
Research Engineer
McDonald Observatory, Texas
McDonald Observatory, Texas
-
- VIP Livecode Opensource Backer
- Posts: 7258
- Joined: Sat Apr 08, 2006 8:31 pm
- Location: Minneapolis MN
- Contact:
Re: No more rpi engine downloads
Are you sure it's a leak? LC holds everything in memory and the field contents are rapidly increasing. Every character uses 4 bytes.
Jacqueline Landman Gay | jacque at hyperactivesw dot com
HyperActive Software | http://www.hyperactivesw.com
HyperActive Software | http://www.hyperactivesw.com
Re: No more rpi engine downloads
So @jwkuehne's example here is simply (when reduced to the minimum required):
Now the actual example we got had 9 fields, and each one was having `the seconds` put into them on each iteration of the loop... However, the seconds, nor there being nine fields, or indeed text being put into them had anything to do with the problem.
Turns out there was the kind of memory leak which Java apps can suffer from rather than your traditional 'someone forgot to call delete/free/release' - i.e. ones where you have lingering references to objects from other live objects, and those references are never removed so the memory can be reclaimed (typically in Java this can occur due to caches, or working lists and such).
Simply put, every time a screen update was being requested (i.e. each time the field was prodded in the loop), Cocoa was allocating a 32-byte temporary object - which was being put on the list to free when control returned to the event loop... The problem? In that example, control *never* returns to the root event loop (well, until you do Cmd+C to stop the loop!).
The reason this hadn't been noticed before is because almost all apps will return the runloop very frequently - and oftentimes (for performance reasons) will only update the screen once or twice before they do. Presumably, the use-case which gave rise to this looping without ever returning control to the system directly is some sort of polling use-case which never really has need to return to the system event loop.
Anyway, after a fair bit of head scratching when the normal native 'memory leak' detection tools showed nothing, this problem has been fixed for 9.6.8-rc-1.
Code: Select all
on mouseUp
repeat forever
put "foo" into field 1
end repeat
end mouseUp
Turns out there was the kind of memory leak which Java apps can suffer from rather than your traditional 'someone forgot to call delete/free/release' - i.e. ones where you have lingering references to objects from other live objects, and those references are never removed so the memory can be reclaimed (typically in Java this can occur due to caches, or working lists and such).
Simply put, every time a screen update was being requested (i.e. each time the field was prodded in the loop), Cocoa was allocating a 32-byte temporary object - which was being put on the list to free when control returned to the event loop... The problem? In that example, control *never* returns to the root event loop (well, until you do Cmd+C to stop the loop!).
The reason this hadn't been noticed before is because almost all apps will return the runloop very frequently - and oftentimes (for performance reasons) will only update the screen once or twice before they do. Presumably, the use-case which gave rise to this looping without ever returning control to the system directly is some sort of polling use-case which never really has need to return to the system event loop.
Anyway, after a fair bit of head scratching when the normal native 'memory leak' detection tools showed nothing, this problem has been fixed for 9.6.8-rc-1.
Re: No more rpi engine downloads
Very interesting!
I have a nagging itchy allergic feeling that running a repeat forever loop is an unhealthy way to conduct yourself
But it sounds like it was felt necessary for this use-case and no better approach was feasible?
Mark's explanation kinda makes sense, but does it explain why this occurs on Mac and not Windows?
I have a nagging itchy allergic feeling that running a repeat forever loop is an unhealthy way to conduct yourself
But it sounds like it was felt necessary for this use-case and no better approach was feasible?
Mark's explanation kinda makes sense, but does it explain why this occurs on Mac and not Windows?
Last edited by stam on Fri Apr 01, 2022 10:07 pm, edited 1 time in total.
-
- VIP Livecode Opensource Backer
- Posts: 7258
- Joined: Sat Apr 08, 2006 8:31 pm
- Location: Minneapolis MN
- Contact:
Re: No more rpi engine downloads
Would a "wait 0" fix it for now?
Jacqueline Landman Gay | jacque at hyperactivesw dot com
HyperActive Software | http://www.hyperactivesw.com
HyperActive Software | http://www.hyperactivesw.com
Re: No more rpi engine downloads
"with messages" maybe?