ZapRedraw and ARMx6/mini.m

There appears to be an issue regarding ZapRedraw and big screen modes as happen on some newer machines.
Fred Graute identified it as a problem around the use of the software pointer on screen modes over 2048 pixels whilst fixing it in StrongEd.
He has been kind enough to provide a patch for ZapRedraw.

As I’m sitting outside writing something for my blog and trying not to get a sugar high from fizzy drinks ¹, I’ve dropped the archive be sent me here: https://heyrick.eu/blog/files/ZapFix.zip

Can anybody with big screen modes that experiences screen corruption when using Zap please try this and let me know if it fixes the problem? I can’t test it myself, my screen is only 1280×1024.

Thanks.

And extra thanks to Fred for, once again, fixing the competition. Somebody buy that man a beer!

¹ It’s 28°C in the shade, and the sensor on my window to tell me the in-the-sun temperature is reading 39°C. Ugh. The only reason I’m outside under a tree is that, for the moment, there’s a (small) breeze. Plus birdsong. 🐦

After a small oops where it locked the mini.m on loading (quit Zap an double clicked on), I rebooted and quit Zap and loaded it ok. On running Zap again it seems to be working ok at 2560×1440×32, and has seemingly fixed the corruption of Zap selections issue. I haven’t managed to leave a close box behind on a Zap Window either. So thanks to Fred, any yourself.

There are still issues on mini.m which is probably due to the accelaration, rather than Zap. I did a Ctrl+F12 to bring up a TaskWindow and increased its width from 120 to 200 characters, it redrew but a rectangle consisting of small part of the work area and the scroll bar was left showing the pinboard background.

Somebody buy that man a beer!

Oh, now you’re being mean, make that some wine please. ;-)

So thanks to Fred, any yourself.

No problem, but as I wrote to Rick I’ve little knowledge of Zap’s redraw system so it may not be a perfect fix. Use at your own risk etc, etc.

There are still issues on mini.m which is probably due to the accelaration,

Yes, I see those regularly. However there doesn’t seem to be any discernible pattern to it so it’ll hard for Adrian to track it down.

Unforunately I seem to be seeing a lot more Zap (1.49 rick-05) errors than usual.

19 Jul 14:46:41 000 00000185: Error from Zap: Not a heap block
19 Jul 15:15:45 002 80000002: Error from Zap: Internal error: abort on data transfer at &FC0267E8
19 Jul 15:16:17 002 80000002: Error from Zap: Internal error: abort on data transfer at &FC0267E8
19 Jul 15:16:26 002 80000002: Error from Zap: Internal error: abort on data transfer at &FC0267E8
19 Jul 15:16:31 000 80000002: Error from (unknown): Internal error: abort on data transfer at &FC0267E8
19 Jul 15:21:52 000 00000185: Error from Zap: Not a heap block
19 Jul 15:22:04 000 00000185: Error from Zap: Not a heap block
19 Jul 15:23:55 002 80000002: Error from Zap: Internal error: abort on data transfer at &FC0267E8
19 Jul 15:26:35 002 80000002: Error from Zap: Internal error: abort on data transfer at &FC0267E8
19 Jul 15:26:44 002 80000002: Error from Zap: Internal error: abort on data transfer at &FC0267E8
19 Jul 15:26:47 002 80000002: Error from Zap: Internal error: abort on data transfer at &FC0267E8
19 Jul 15:26:50 000 80000002: Error from (unknown): Internal error: abort on data transfer at &FC0267E8

*where
Address &FC0267F0 is at offset &000267F0 in the Kernel

Oh, now you’re being mean, make that some wine please. ;-)

I live in France, wine is possible. ;-)

[just a tip when eating out in France: avoid the house wine left in an open bottle on the table; it’s a cardinal sin known as “temptation” as will become clear when the bill comes (especially if “foreigner prices” are applied), and it is so awful that even the flies avoid it (and flies are happy slurping excrement)!]

Unforunately I seem to be seeing a lot more Zap (1.49 rick-05) errors than usual.

I think (but haven’t looked at the source) that the ‘fix’ is simply turning the cursor off and back on again around the actual redraw. This really shouldn’t be having weird effects.

I’ll need to copy out this ZapRedraw into my setup and use it a while to see if it’s a problem with Zap generally, or the big screen modes.

However… rick-05? You know that’s two versions old, right? Pop over to !Store for an update! ;-)

Arrh! I only ran !Store a few days ago, the new version of Zap is there, and although its in the My Software list as I downloaded a previous version, it makes no attempt to let me know it’s been updated. That and the greif from PacMan hanging when tring to update anything now, it’s making upgrading harder rather than easier.

Please can everyone go back to posting to comp.sys.acorn.announce – it just works!

I’ve done a clean install of the application this time, as comparing I found some old 2001/2002 stuff from the 26 bit days:-

Code.Extensions.ExtAsm
Code.Extensions.ExtAsm.ExtBasAsm
Code.Extensions.FlashCar
Modules.!ZapBasAsm
Modules.!ZapDS
Modules.!ZapInform
Modules.!ZapMan
Modules.!ZapSAsm

But unfortunately just had a another Not a Heap Block error, followed by a Zap must exit :(

it makes no attempt to let me know it’s been updated.

Yeah. Apparently even though Store knows what you’ve downloaded and what version it was and what version is available now, it doesn’t make any attempt to notify you of updates without talking to the mothership because it seems this stuff is tracked server side, which must be rather interesting if you have Store on multiple machines which may not have the same things installed on each.

Sounds like catastrophically broken behaviour to me, but I’m not the guy that wrote it, I’m just the guy that bitches about it on a forum…

Please can everyone go back to posting to comp.sys.acorn.announce

Not been near Usenet in several years. Can’t say that I miss it.

as comparing I found some old 2001/2002 stuff from the 26 bit days

Is that in my ultimate edition? I thought I got rid of the 26 bit only stuff.

Modules.!ZapInform

I got in touch with Mr. Nelson. Sources lost to time, sadly enough. I don’t have the time to reverse engineer it.

followed by a Zap must exit :(

With the “fixed” ZapRedraw or the one supplied with Zap?

Yeah. Apparently even though Store knows what you’ve downloaded and what version it was and what version is available now, it doesn’t make any attempt to notify you of updates without talking to the mothership because it seems this stuff is tracked server side, which must be rather interesting if you have Store on multiple machines which may not have the same things installed on each.
bq.
Sounds like catastrophically broken behaviour to me, but I’m not the guy that wrote it, I’m just the guy that bitches about it on a forum

Rick not sure how many times you have been told this. If you logon to !Store and there are updates to download then the “My Software” icon flashes green on the text. If you open the “My Software” by clicking on it it then highlights any updates in Green for free updates or Red if Chargeable.

True is do not send you an email to notify you of updates, but i don’t see why it should.

It also works on multiple computers

If you logon to !Store and there are updates to download

My emphasis, highlighting the entire problem.

True is do not send you an email to notify you of updates,

Please read what was written – nobody mentioned sending emails. Even elsewhere (Google app store, etc) you don’t get emails!

Just to clarify !Store and updates – it does offer/notify updates via My Software provided it can parse version numbers. It knows the version you last downloaded, and will offer you new ones if higher. The problem, I suspect, is that this is stored numerically. I think the parser is designed for x.y.z version numbers. I believe it can cope with things like 2.22a etc, but I suspect it can’t handle “rick-05”.

In hindsight, it probably makes more sense to notify you if the upload date on the content is newer than your last download date. I’m not sure whether that’s stored though.

I wrote:

as comparing I found some old 2001/2002 stuff from the 26 bit days

Rick wrote:

Is that in my ultimate edition? I thought I got rid of the 26 bit only stuff.

No it’s not in your ultimate edition, it must have been left from an older version.

I’m now using DirSync to apply updates to applications, as just dropping the new one over the top wont spot files which should be removed, and datestamps are becoming increasingly unreliable given the number of different ways which updates may have been obtained.

Rick, there are two non-32-bit modules in your distribution, ZapJRF and ZapDWExt.

It’s kind of a relief…two less modules to worry about the license on!

it does offer/notify updates via My Software

Oh, right. You mean the part that’s greyed out?

GOTO 10

;-)

Oh, right. You mean the part that’s greyed out?

Its only greyed out because you have not logged on what is so hard about that. Really you are making a mountain out of nothing, just register and logon. Its not hard you know.

I was logged in and the My Software just showed different coloured alternative lines. I was expecting newer software to be highlighted with a different text colour or a symbol or even good old text. I’m pretty sure it was once.

I was logged in and the My Software just showed different coloured alternative lines. I was expecting newer software to be highlighted with a different text colour or a symbol or even good old text. I’m pretty sure it was once

Druck, when logged in on My software, as it says in the top section. text in green mean a new FREE download available, while text in Red means a new chargeable upgrade usually, like I have a red entry on mine for messenger Pro 7.08, but just above this in my case I also have an entry for the latest messenger pro 8.04, which I upgraded too, and since that is the latest the text is in black.
Also if you double click on any line it opens the info window on the software and at the bottom left is a more info button it provides details of what downloads versions and dates your updated, and if their, is a newer version it also sometimes will list whats new in the latest version.

Chris, I’ll have another look when someone releases more software.

My biggest frustration at the moment is the unreliability of !Zap 1.49rick-07 (sorry Rick) on my mini.m (patched ZapRedraw module), on a Pi 3B (normal ZapRedraw), and an Iyonix (normal ZapRedraw), all running 5.27 RC16. I’m getting frequirent No a heap block errors, various exceptions and must exits. I used to get the occasional fault with RISC OS 5.24, but it is almost constant now.

just register and logon. Its not hard you know.

<sigh>

I give up.

I used to get the occasional fault with RISC OS 5.24, but it is almost constant now.

Ah. That’ll perhaps be why I’ve never seen a “Not a heap block” error. I’m running 5.23.

So… What in the OS has changed that would generate that error?

I give up.

I don’t use !Store (I’ve been watching the updates and I’m waiting for an obvious stable status, but…

I believe your point (not registering in certain brains) was that the “what has been downloaded/updated” record is on the server and if that is indexed against a person rather than a computer then updating one system leaves the updater confused about whether system 2,3 or 10 are updated correctly.
A tag/cookie/token on the individual device marking its unique update status is the actual requirement and forcing a full system-wide update would be a simple as deleting said token(s)

My biggest frustration at the moment is the unreliability of !Zap 1.49rick-07 (sorry Rick) on my mini.m (patched ZapRedraw module)

A bit of diagnostic chit chat between you should home in on the issue and Rick will fix that.

I don’t use !Store (I’ve been watching the updates and I’m waiting for an obvious stable status,

Whats not stable about it, it gets updated with new software or updates to it each time you run the application.

I believe your point (not registering in certain brains) was that the “what has been downloaded/updated” record is on the server and if that is indexed against a person rather than a computer then updating one system leaves the updater confused about whether system 2,3 or 10 are updated correctly

The information about what you have purchased or downloaded is stored on your computer! Here it is located in Scrap.ScrapDirs.ScrapDir.PlingStore in a file called mysoft this lists all the software downloaded and updates etc. The newest downloads (which are not updates) are at top of file.

You can save this file anywhere you want by changing the option for the “Data Store” in Choices from the !Store icon bar menu.

For other machines you could simply copy the appropriate files to another computer.

Druck, when logged in on My software, as it says in the top section. text in green mean a new FREE download available,

I find that that is true for all the software I’ve obtained from PlingStore except for Zap. Right now ‘My software’ shows Zap Ultimate as version 1.49 rick-05. It’s not highlighted despite 07 being available in the catalogue. Been the same for previous versions of Zap, but only Zap.

Rick, there are two non-32-bit modules in your distribution, ZapJRF and ZapDWExt.

Thanks. Just taken a photo of the screen so hopefully I’ll see that later and remember to remove those modules.

and I’m waiting for an obvious stable status

Sorry, we’re now talking about Store, not Zap. :-)

I believe your point (not registering in certain brains) was

My point is that the record of what’s been downloaded is more concerned with what has been purchased. I can understand that, and I fully understand the need to log in to access purchases.

However, by conflating that with “versions of stuff downloaded”, rather than treating the two separately, there is a notification of updates that is only available to some people, but not others. Specifically, there is no automatic notification in the list of available software for people who are using Store without having logged in.

There doesn’t seem to be a lot of point in my registering myself with the app because not only is logging in an extra hoop to jump through prior to looking at what’s available but also because I will never purchases anything from Store. That’s not aimed at R-Comp, by the way. I have the exact same policy with Apple’s App Store (not that I can even find my iPad these days, it became mom’s toy…) and Google Play. The simplest way to keep my bank card information safe is to not give it out to an online service. Indeed, I think the only two providers that have it are Amazon and ErDF (I still pay my electricity bill manually, the other stuff is by direct debit).

The information about what you have purchased or downloaded is stored on your computer!

If that’s the case, there’s no real reason not to have updates be tracked client side (and independant of purchases).

Oh, and some nerdy speak should somebody think of adding this: just treat the version as a string. “0.1.2” is later than “0.1.1” and “1.23beta47c” is later than “1.23beta47a”. There’s no need whatsoever to try to parse the version identifier to figure out what the hell it is, just toss it at strcmp(). It’ll be less than (result is -ve), equals (result is 0), or greater than (result is +ve). That’s all you need to know.

It’s not highlighted despite 07 being available in the catalogue.

It’s probably trying too hard to work out what the version number is, and gets stuck when somebody does something weird like use an incrementing suffix. ;-)

Refer to the last paragraph of the previous post for a suggestion of an alternative. Sometimes it’s better not to overthink things. ;-)