Re: Android Speed Woes
Ok... so here's an update... The card that contains all this is a calendar. On this card is a header, a datagrid that lists the events of the selected day, and another group with 84 controls in it (one button and field for each possible day) that makes up the calendar's grid, and more. When you click "New Entry" or "Edit Entry", we hide the calendar and show the group "newEventG" which, as you know, contains several other groups for setting the event's properties (title, notes, dates, repeat, and more). It is in the group "newEventG" that we are seeing all the sluggish behavior. Last night, I painstakingly moved group "newEventG" to its own card. Now, the speed is no longer an issue. It's snappy and responsive. I have to wonder if having a hidden group with so many controls was messing things up. Mark, to answer your question, there isn't one image on the entire card. ShowFocusBorder was on. All these were correct: - opaque must be false - showName must be false - hScrollbar must be false - vScrollbar must be false - showBorder must be false - all bitmap effects must be empty I am in a good place now and ok to continue with my project (although the tablet version is now funky as it moves to another card for event editing rather than the group overlay). I hope all this helps LC resolve whatever this issue is. -Dan On 8/29/19, 11:56 PM, "use-livecode on behalf of Mark Waddingham via use-livecode" wrote: On 2019-08-29 16:24, Dan Friedman via use-livecode wrote: > Hello again! I am have a tough time getting things to run snappy on > an Android device. It's blazing fast on my computer, and more than > acceptable on my iPhone, but on any android device, it's slow... > noticeably sluggish. The issue is mainly hiding/showing groups. For > example: > > put the milliseconds into m > lock screen --for visual effect in rect (the rect of grc "ne_BG") > //removed the animation for speed testing > hide grp "ne_mainControls" > show grp "ne_repeatG" > unlock screen --with visual effect push left > answer (the milliseconds - m) > > The result of m is: > iMac = 10 > iOS = 1 (iPhoneX) > Android = 516 (Moto G5 Plus Android 7.0) > > The groups contain no images or complex graphics. Just some buttons > and fields. > > I have tried it with and without acceleratedRendering. Actually, with > acceleratedRendering enabled it took longer, m=558. > > I have it setup like this: > Card "calendar" >Group "newEventG" --> layerMode set to "Container" > Group "ne_mainControls" --> layerMode set to "Dynamic" > Group " ne_repeatG" --> layerMode set to "Dynamic" > > Any thoughts or advice? Most other functions are working within > acceptable ranges/speeds. So the first thing to do is (on Desktop) run with acceleratedRendering and ensure that: 1) 'the effective layerMode of group "newEventG"' is container 2) 'the effective layerMode of group "ne_mainControls" is dynamic 3) 'the effective layerMode of group "ne_repeatG" is dynamic If (1) does not turn out to be the case then acceleratedRendering will have a slightly negative effect (as the engine will be doing what it would do without acceleratedRendering on, but also paying the overhead of caching the result!). To fix this you'll need to tweak the properties of the newEventG group: - opaque must be false - showName must be false - hScrollbar must be false - vScrollbar must be false - showBorder must be false - all bitmap effects must be empty Essentially the rule here is that for a group to be a container layer it must not have to render anything itself (as it is just meant to 'contain' other layers). If you get the groups effective layerModes as above I'd imagine you *might* see a slight increase in performance with acceleratedRendering on - the showing and hiding of the dynamic layers is essentially 'free' in this scenario - but their content does need to be cached first. Caching of dynamic layers happens incrementally as parts of them are visible. However, given the difference in speed between the platforms with acceleratedRendering off I'm inclined to think something else is going on here... The CPU on a Moto G5 Plus is substantially less powerful than even an iPhone6S - but definitely not 50x! You mentioned in another thread that you were using custom fonts... Do you get better performance if you remove the use of custom fonts and just go with the default? Warmest Regards, Mark. -- Mark Waddingham ~ m...@livecode.com ~ http://www.livecode.com/ LiveCode: Everyone can create apps
Re: Android Speed Woes
On 2019-08-30 02:56, Mark Waddingham via use-livecode wrote: If (1) does not turn out to be the case then acceleratedRendering will have a slightly negative effect (as the engine will be doing what it would do without acceleratedRendering on, but also paying the overhead of caching the result!). To fix this you'll need to tweak the properties of the newEventG group: - opaque must be false - showName must be false - hScrollbar must be false - vScrollbar must be false - showBorder must be false - all bitmap effects must be empty Additionally and entirely unrelated to performance but might be causing the visual artefacts you mentioned... For now you also need to ensure that your three groups also all have 'showFocusBorder' set to false - this is due to bug 22322, a fix for which is pending (https://github.com/livecode/livecode/pull/7144). [ ShowFocusBorder has absolutely no effect on groups, however it causes a miscalculation to occur when rendering in certain circumstances when it is enabled and acceleratedRendering is on ]. Warmest Regards, Mark. -- Mark Waddingham ~ m...@livecode.com ~ http://www.livecode.com/ LiveCode: Everyone can create apps ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
On 2019-08-29 16:24, Dan Friedman via use-livecode wrote: Hello again! I am have a tough time getting things to run snappy on an Android device. It's blazing fast on my computer, and more than acceptable on my iPhone, but on any android device, it's slow... noticeably sluggish. The issue is mainly hiding/showing groups. For example: put the milliseconds into m lock screen --for visual effect in rect (the rect of grc "ne_BG") //removed the animation for speed testing hide grp "ne_mainControls" show grp "ne_repeatG" unlock screen --with visual effect push left answer (the milliseconds - m) The result of m is: iMac = 10 iOS = 1 (iPhoneX) Android = 516 (Moto G5 Plus Android 7.0) The groups contain no images or complex graphics. Just some buttons and fields. I have tried it with and without acceleratedRendering. Actually, with acceleratedRendering enabled it took longer, m=558. I have it setup like this: Card "calendar" Group "newEventG" --> layerMode set to "Container" Group "ne_mainControls" --> layerMode set to "Dynamic" Group " ne_repeatG" --> layerMode set to "Dynamic" Any thoughts or advice? Most other functions are working within acceptable ranges/speeds. So the first thing to do is (on Desktop) run with acceleratedRendering and ensure that: 1) 'the effective layerMode of group "newEventG"' is container 2) 'the effective layerMode of group "ne_mainControls" is dynamic 3) 'the effective layerMode of group "ne_repeatG" is dynamic If (1) does not turn out to be the case then acceleratedRendering will have a slightly negative effect (as the engine will be doing what it would do without acceleratedRendering on, but also paying the overhead of caching the result!). To fix this you'll need to tweak the properties of the newEventG group: - opaque must be false - showName must be false - hScrollbar must be false - vScrollbar must be false - showBorder must be false - all bitmap effects must be empty Essentially the rule here is that for a group to be a container layer it must not have to render anything itself (as it is just meant to 'contain' other layers). If you get the groups effective layerModes as above I'd imagine you *might* see a slight increase in performance with acceleratedRendering on - the showing and hiding of the dynamic layers is essentially 'free' in this scenario - but their content does need to be cached first. Caching of dynamic layers happens incrementally as parts of them are visible. However, given the difference in speed between the platforms with acceleratedRendering off I'm inclined to think something else is going on here... The CPU on a Moto G5 Plus is substantially less powerful than even an iPhone6S - but definitely not 50x! You mentioned in another thread that you were using custom fonts... Do you get better performance if you remove the use of custom fonts and just go with the default? Warmest Regards, Mark. -- Mark Waddingham ~ m...@livecode.com ~ http://www.livecode.com/ LiveCode: Everyone can create apps ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
That’s perfect and very near the results we’d expect to see. So it narrows it down to display. What’s inside the groups and what settings do they have? If there are images what quality setting do they have, normal better or best? Any scaling factors anywhere? Rotations? Anti-aliasing? Shapes, widgets or paths? Borders, effects (shadow, glow, overlay). Theoretically these settings would affect both iOS and Android but as we are seeing it only on Android we need to narrow it down to see what specifically is causing the issue. To start, try setting any images to mid quality (not normal or best). I’ve occasionally seen this performs best (especially recently on HTML5 deployment). Sean > On 30 Aug 2019, at 01:15, Dan Friedman via use-livecode > wrote: > > Mark, > > put the milliseconds into m > put "" into jones > repeat 1 >put "a" after jones > end repeat > put jones into pFart > put the milliseconds - m into m > answer m > > iMac = 1 > Moto G5 = 16 > Moto G6 = 10 > iPhone X = 7 > > -Dan > ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
Mark, put the milliseconds into m put "" into jones repeat 1 put "a" after jones end repeat put jones into pFart put the milliseconds - m into m answer m iMac = 1 Moto G5 = 16 Moto G6 = 10 iPhone X = 7 -Dan On 8/29/19, 5:09 PM, "use-livecode on behalf of Mark Wieder via use-livecode" wrote: On 8/29/19 4:40 PM, Dan Friedman via use-livecode wrote: > Sean, > >>> Another test would be to see how fast it creates a variable with 10,000 characters using a repeat loop ... > > put "" into jones > repeat 1 > put "a" after jones > end repeat > put the milliseconds into m > put jones into pFart > put the milliseconds - m into m > answer m You really want to store the milliseconds *before* the repeat loop to have any chance of seeing a difference. -- Mark Wieder ahsoftw...@gmail.com ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
On 8/29/19 4:40 PM, Dan Friedman via use-livecode wrote: Sean, Another test would be to see how fast it creates a variable with 10,000 characters using a repeat loop ... put "" into jones repeat 1 put "a" after jones end repeat put the milliseconds into m put jones into pFart put the milliseconds - m into m answer m You really want to store the milliseconds *before* the repeat loop to have any chance of seeing a difference. -- Mark Wieder ahsoftw...@gmail.com ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
Looks like we should be comparing GPUs instead of CPUs. -- Jacqueline Landman Gay | jac...@hyperactivesw.com HyperActive Software | http://www.hyperactivesw.com On August 29, 2019 6:42:09 PM Dan Friedman via use-livecode wrote: Sean, Another test would be to see how fast it creates a variable with 10,000 characters using a repeat loop ... put "" into jones repeat 1 put "a" after jones end repeat put the milliseconds into m put jones into pFart put the milliseconds - m into m answer m iMac = 1 Moto G5 = 0 Moto G6 = 0 iPhone X = 0 -Dan On 8/29/19, 4:10 PM, "use-livecode on behalf of Pi Digital via use-livecode" use-livecode@lists.runrev.com> wrote: Dan is talking 500 times difference in response tho. That’s a bit more substantial. Plus speed degradation to a device designed to run Android 7.0 shouldn’t be to that degree either, for display or under the hood. The G5 has an octo 1.4ghz, the iPhoneX is 6core 2.39, only 4 times as fast (8-10 times offscreen), not 200-500. \?/ Dan, Another test would be to see how fast it creates a variable with 10,000 characters using a repeat loop then how long to copy that to another variable. If the numbers are wildly different to the iPhone’s and desktop’s then maybe Jacque is on to something and it is a device specific lag. Otherwise it’s likely a display issue, which could be device specific but could equally be something deeper worth looking into. Sean > On 29 Aug 2019, at 23:23, J. Landman Gay via use-livecode wrote: > > I'd say it's the phone. The Moto G5 has a budget processor and an old version of Android OS. An iPhoneX is newer and much faster so it isn't really a fair comparison. ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
Sean, >>Another test would be to see how fast it creates a variable with 10,000 >>characters using a repeat loop ... put "" into jones repeat 1 put "a" after jones end repeat put the milliseconds into m put jones into pFart put the milliseconds - m into m answer m iMac = 1 Moto G5 = 0 Moto G6 = 0 iPhone X = 0 -Dan On 8/29/19, 4:10 PM, "use-livecode on behalf of Pi Digital via use-livecode" wrote: Dan is talking 500 times difference in response tho. That’s a bit more substantial. Plus speed degradation to a device designed to run Android 7.0 shouldn’t be to that degree either, for display or under the hood. The G5 has an octo 1.4ghz, the iPhoneX is 6core 2.39, only 4 times as fast (8-10 times offscreen), not 200-500. \?/ Dan, Another test would be to see how fast it creates a variable with 10,000 characters using a repeat loop then how long to copy that to another variable. If the numbers are wildly different to the iPhone’s and desktop’s then maybe Jacque is on to something and it is a device specific lag. Otherwise it’s likely a display issue, which could be device specific but could equally be something deeper worth looking into. Sean > On 29 Aug 2019, at 23:23, J. Landman Gay via use-livecode wrote: > > I'd say it's the phone. The Moto G5 has a budget processor and an old version of Android OS. An iPhoneX is newer and much faster so it isn't really a fair comparison. ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
Sean, I wouldn't say AR is non-functional. When I enable it I do see a difference. Although, in almost all cases it makes it worse. I see jerky scrolling and screen artifacts. I am using 9.5.0 Business. I compiled the app in 9.0.4 Indy, and m=562. Interesting that I just tried it on a Moto G6 and m=585. Slower than the G5! >> Is this timing consistent if triggered by a button? That's how I am firing this. >> What timing do you get with lock screen muted? If I comment out the lock screens, I get m=756 (Moto G5) -Dan On 8/29/19, 2:40 PM, "use-livecode on behalf of Pi Digital via use-livecode" wrote: Hi Dan I don’t think this is fully related to your issue based on the script you show but... I’ve been doing extensive testing this week and have come close to the conclusion now that AcceleratedRendering is completely non functional for all platforms. However, I’ve got a feeling that none of my recent emails to LC on the matter have got through during their outage. Is no one else seeing this loss in performance? So far all my tests have shown it to be affecting every platform standalone and the IDE. Back to your issue, Dan. What version LC are you using? Is this timing consistent if triggered by a button? What timing do you get with lock screen muted? Also, put the timer calculation to a variable to ensure it is not the answer dialogue itself producing erroneous timings. NB. Additionally, if your groups are not moving up down left right then they are best set to static rather than dynamic or scrolling. This won’t be affecting your performance in this instance but worth noting. It’s only useful setting it to ‘dynamic’ if the values, text or images are changing frequently/rapidly or moving about the display. Sean Cole Pi Digital Prod Ltd > On 29 Aug 2019, at 21:24, Dan Friedman via use-livecode wrote: > > Hello again! I am have a tough time getting things to run snappy on an Android device. It's blazing fast on my computer, and more than acceptable on my iPhone, but on any android device, it's slow... noticeably sluggish. The issue is mainly hiding/showing groups. For example: > > put the milliseconds into m > lock screen --for visual effect in rect (the rect of grc "ne_BG") //removed the animation for speed testing > hide grp "ne_mainControls" > show grp "ne_repeatG" > unlock screen --with visual effect push left > answer (the milliseconds - m) > > The result of m is: > iMac = 10 > iOS = 1 (iPhoneX) > Android = 516 (Moto G5 Plus Android 7.0) > > The groups contain no images or complex graphics. Just some buttons and fields. > > I have tried it with and without acceleratedRendering. Actually, with acceleratedRendering enabled it took longer, m=558. > > I have it setup like this: > Card "calendar" > Group "newEventG" --> layerMode set to "Container" > Group "ne_mainControls" --> layerMode set to "Dynamic" > Group " ne_repeatG" --> layerMode set to "Dynamic" > > Any thoughts or advice? Most other functions are working within acceptable ranges/speeds. > > -Dan > > ___ > use-livecode mailing list > use-livecode@lists.runrev.com > Please visit this url to subscribe, unsubscribe and manage your subscription preferences: > http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
Dan is talking 500 times difference in response tho. That’s a bit more substantial. Plus speed degradation to a device designed to run Android 7.0 shouldn’t be to that degree either, for display or under the hood. The G5 has an octo 1.4ghz, the iPhoneX is 6core 2.39, only 4 times as fast (8-10 times offscreen), not 200-500. \?/ Dan, Another test would be to see how fast it creates a variable with 10,000 characters using a repeat loop then how long to copy that to another variable. If the numbers are wildly different to the iPhone’s and desktop’s then maybe Jacque is on to something and it is a device specific lag. Otherwise it’s likely a display issue, which could be device specific but could equally be something deeper worth looking into. Sean > On 29 Aug 2019, at 23:23, J. Landman Gay via use-livecode > wrote: > > I'd say it's the phone. The Moto G5 has a budget processor and an old version > of Android OS. An iPhoneX is newer and much faster so it isn't really a fair > comparison. ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
Just for your future reference, MarkWs excellent description about AccRend: http://lists.runrev.com/pipermail/use-livecode/2017-May/236782.html Sean > > On 29 Aug 2019, at 21:24, Dan Friedman via use-livecode < > use-livecode@lists.runrev.com> wrote: > > > > I have tried it with and without acceleratedRendering. Actually, with > acceleratedRendering enabled it took longer, m=558. > > > > I have it setup like this: > > Card "calendar" > > Group "newEventG" --> layerMode set to "Container" > > Group "ne_mainControls" --> layerMode set to "Dynamic" > > Group " ne_repeatG" --> layerMode set to "Dynamic" > > > > Any thoughts or advice? Most other functions are working within > acceptable ranges/speeds. > > > > -Dan > ___ > use-livecode mailing list > use-livecode@lists.runrev.com > Please visit this url to subscribe, unsubscribe and manage your > subscription preferences: > http://lists.runrev.com/mailman/listinfo/use-livecode > ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
I'd say it's the phone. The Moto G5 has a budget processor and an old version of Android OS. An iPhoneX is newer and much faster so it isn't really a fair comparison. My Pixel has a much faster processor running the latest OS and is quite snappy; my Samsung S5 is old and slow and responds as such. I develop with my Pixel and test with the Samsung to see how bad it gets. The up side is that users of older or budget phones are used to the response times and may not notice. But if you can find a higher-end test device, it's useful to compare. On 8/29/19 3:24 PM, Dan Friedman via use-livecode wrote: Hello again! I am have a tough time getting things to run snappy on an Android device. It's blazing fast on my computer, and more than acceptable on my iPhone, but on any android device, it's slow... noticeably sluggish. The issue is mainly hiding/showing groups. For example: put the milliseconds into m lock screen --for visual effect in rect (the rect of grc "ne_BG") //removed the animation for speed testing hide grp "ne_mainControls" show grp "ne_repeatG" unlock screen --with visual effect push left answer (the milliseconds - m) The result of m is: iMac = 10 iOS = 1 (iPhoneX) Android = 516 (Moto G5 Plus Android 7.0) The groups contain no images or complex graphics. Just some buttons and fields. I have tried it with and without acceleratedRendering. Actually, with acceleratedRendering enabled it took longer, m=558. I have it setup like this: Card "calendar" Group "newEventG" --> layerMode set to "Container" Group "ne_mainControls" --> layerMode set to "Dynamic" Group " ne_repeatG" --> layerMode set to "Dynamic" Any thoughts or advice? Most other functions are working within acceptable ranges/speeds. -Dan ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode -- Jacqueline Landman Gay | jac...@hyperactivesw.com HyperActive Software | http://www.hyperactivesw.com ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Android Speed Woes
Hi Dan I don’t think this is fully related to your issue based on the script you show but... I’ve been doing extensive testing this week and have come close to the conclusion now that AcceleratedRendering is completely non functional for all platforms. However, I’ve got a feeling that none of my recent emails to LC on the matter have got through during their outage. Is no one else seeing this loss in performance? So far all my tests have shown it to be affecting every platform standalone and the IDE. Back to your issue, Dan. What version LC are you using? Is this timing consistent if triggered by a button? What timing do you get with lock screen muted? Also, put the timer calculation to a variable to ensure it is not the answer dialogue itself producing erroneous timings. NB. Additionally, if your groups are not moving up down left right then they are best set to static rather than dynamic or scrolling. This won’t be affecting your performance in this instance but worth noting. It’s only useful setting it to ‘dynamic’ if the values, text or images are changing frequently/rapidly or moving about the display. Sean Cole Pi Digital Prod Ltd > On 29 Aug 2019, at 21:24, Dan Friedman via use-livecode > wrote: > > Hello again! I am have a tough time getting things to run snappy on an > Android device. It's blazing fast on my computer, and more than acceptable > on my iPhone, but on any android device, it's slow... noticeably sluggish. > The issue is mainly hiding/showing groups. For example: > > put the milliseconds into m > lock screen --for visual effect in rect (the rect of grc "ne_BG") //removed > the animation for speed testing > hide grp "ne_mainControls" > show grp "ne_repeatG" > unlock screen --with visual effect push left > answer (the milliseconds - m) > > The result of m is: > iMac = 10 > iOS = 1 (iPhoneX) > Android = 516 (Moto G5 Plus Android 7.0) > > The groups contain no images or complex graphics. Just some buttons and > fields. > > I have tried it with and without acceleratedRendering. Actually, with > acceleratedRendering enabled it took longer, m=558. > > I have it setup like this: > Card "calendar" > Group "newEventG" --> layerMode set to "Container" > Group "ne_mainControls" --> layerMode set to "Dynamic" > Group " ne_repeatG" --> layerMode set to "Dynamic" > > Any thoughts or advice? Most other functions are working within acceptable > ranges/speeds. > > -Dan > > ___ > use-livecode mailing list > use-livecode@lists.runrev.com > Please visit this url to subscribe, unsubscribe and manage your subscription > preferences: > http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
