Status of Zap

Thanks Peter and Chris. I’ll have a read whilst I’m holiday this coming 7 days.

Making use of Procedure (space) I’ve added some patches to Zap module.

I was going to use Rick’s BreakAid – but instead used Reporter.

It looks like swi OS_HeapSort doesn’t use flags – and have given
swi OS_HeapSort32 r7 =#0 (anyone know different).

As a note -:is there a easy way of converting -
&3F,6 into &FC000000
BASIC doen’t execept the former.
ie.
BIC   R6,R6,#&3F,6; =&FC000000

As a note -:is there a easy way of converting -
&3F,6 into &FC000000

a%=(&3F << 26)

What about (29,30)
these values seem to be given by the Debugger.
On RO 4.02 the output is different – and the RO5.23 output is not understood by BASIC – via Zap.

The early posts in this thread are amongst those I’ve never read before due to a computer transition – so I’ve only now seen the comment about the mailing lists…

Perhaps Vince could take over the mailing lists, since RISCOSitory uses mailman to provide the lists it hosts, so it may be possible to import the mailing list archives with no lost history?

I’ve no idea about importing the archives – I’ve just had a look at the options available to me via the hosting I use, and can see no obvious way. That doesn’t mean it’s not possible – just that I don’t know how.

However, I note the suggestion in the OP:

it might be an idea for someone to set up a single new mailing list so there is someone for people to discuss issues going forward.

Along with the comment that the current lists are quiet I’m inclined to think multiple lists would probably be wasteful – so even if I could see how to import the archives, importing different messages from multiple lists might be ‘interesting’. :)

That said, if the list archives are available, I should be able to devise some way to make them readable online separatly. (A quick look on the old mailing list page shows the lists themselves are now shut down).

So, with all that in mind zap-list at riscository.co.uk is now set up.

(Edit: deliberately split up the email address because the forum was automatically turning it into a mailto rather than use the link I’d used for the mailing list page.)

Ref strange numbers generated via :-
********
SYS"Debugger_Disassemble",&E28AAF29 ¸ ,ANS$
PRINT ANS$
END
********
Gets – ADD R10,r10,#&29,30
On RO4 it’s ADD R10,R10,#&A4
*******
To translate to Human Maths.
********

INPUT " ""Example &29,30 =&A4 "" "A$,B%
A%=EVAL
DIM c% 16
P%=c%
[OPT 2
MOV R0,R0,ROR #B%
MOV PC,R14
]
PRINT “&”;~USR
**********
Can BASIC do ‘ROR’?

Could BASIC Assembler be make to handle variables like -
A$= “mov r1,r1” ?

Can BASIC do ‘ROR’?

Not in a single operation, no. It does have Logical Shift Left and Logical Shift Right (<< and >>> operators) and Bitwise OR, so you can construct it fairly simply:

R1% = (R2% << (32-n)) OR (R2% >>> n)

I have updated Tank’s source up to Zap 1.48 tnk-11 and put them in an archive http://www.riscos-digitalcd.net/apps/ZapSource148tnk11.zip

The built module differs in its binary due to the way of encoding constants in instruction because I still use an old DDE version (the newer seem to encode the value without shift if possible when the old one nearly systematically encode it as a shifted constant) but it disassemble to the same code.

I am using Zap v1.48 (12 Jul 2015) tnk-11, and fairly frequently I get swamped by one particular ZeroPain problem. Today I had over 300 in about 2 minutes. Since I noticed them I have been trying to identify what I did to provoke them, but as yet I have found no specific correlation – but Basic mode is likely to be involved.

I will post an extract from one of the logs in the hope that someone can help with a fix…

Location: Offset 00018788 in module Zap
Current Wimp task: Zap

R0 = 00000003 R1 = 000019fa R2 = 000019f8 R3 = 0000004d
R4 = 09090909 R5 = 0a0a0a0a R6 = 00000072 R7 = 00000072
R8 = 0000000e R9 = 000b5750 R10 = 000b4a6c R11 = 000e9070
R12 = 20dba62c R13 = 000086d0 R14 = 00000003 R15 = 20dba764
DFAR = 000019f8 Mode USR32 Flags nzCv if PSR = 20000010

R15 = 20dba764 = Zap +18790 = cln_forward +138
The offsets are always the same, but the DFAR (r2) varies from 0×0088 to 0×138c.

20dba71c : 1a000005 : BNE &20DBA738
20dba720 : e1580003 : CMP R8,R3
20dba724 : 8affffe0 : BHI &20DBA6AC
20dba728 : e2899001 : ADD R9,R9,#1
20dba72c : e2877001 : ADD R7,R7,#1
20dba730 : e2866001 : ADD R6,R6,#1
20dba734 : e8bd8002 : LDMIA R13!,{R1,PC}
20dba738 : e1580003 : CMP R8,R3
20dba73c : 2affffda : BCS &20DBA6AC
20dba740 : e2899001 : ADD R9,R9,#1
20dba744 : e59d1000 : LDR R1,[R13,#0]
20dba748 : e3510b01 : CMP R1,#&0400 ; =1024
20dba74c : b8bd8002 : LDMLTIA R13!,{R1,PC}
20dba750 : e92d000c : STMDB R13!,{R2,R3}
20dba754 : e0811088 : ADD R1,R1,R8,LSL #1
20dba758 : e3c12003 : BIC R2,R1,#3
20dba75c * e5923000 * LDR R3,[R2,#0] <<<<<<<<<<<<<<
20dba760 : e2012003 : AND R2,R1,#3
20dba764 : e1a02182 : MOV R2,R2,LSL #3
20dba768 : e1a01273 : MOV R1,R3,ROR R2
20dba76c : e8bd000c : LDMIA R13!,{R2,R3}
20dba770 : e0888821 : ADD R8,R8,R1,LSR #16
20dba774 : e1580003 : CMP R8,R3
20dba778 : 21a08003 : MOVCS R8,R3
20dba77c : eaffffab : B &20DBA630

Other possibly relevant versions are:
RISC OS 5.23 (26 May 2016)
ZapRedraw 0.48 (07 Apr 2010)
ZeroPain 0.06 (05 Apr 2016)

I have updated Tank’s 1.48 tnk11 sources to fix all the ZeroPains that occurred on my machine and put them all will a compiled Zap module in an archive http://www.riscos-digitalcd.net/apps/ZapSource148tim01.zip.

Well, fix is big word since I merely added checks in the Mode0 entry points to avoid try to process documents marked as “deleted” (i.e. buffer pointer set to -1). I haven’t figured why such calls where made on “deleted” documents though.

I have used this version for a month now without noticeable side effect but then again I probably don’t use much of Zap’s functionality so don’t forget to backup your copy of of of the module (!Zap.Code.Zap) before trying my version.

Edit: just updated the archive as I had not replaced “tnk-11” by “tim-01” to mark my version.

Now I’m puzzled. How do I find the compiled code to copy over my version of !Zap tnk-11? The ‘o’ directory seems to have files of type ‘Data’.

Root dir of !ZapSource.

!ZapSource.Zap.

Took me a while. It’s just the module.

I like to thank those individuals (Rick, Julie,…) who have put their valuable time and energy into keeping !Zap alive. It’s amazing that its almost 2021 and its still running on RISC OS!

But I’m reluctant to put more time in myself without knowing that Zap has a legal future, e.g. could be included in ROOL’s distrubtion, and can be hosted on GitHub by us to manage contributions.

That’s understandable. Just a thought. Would putting !Zap into Package Manager be a viable alternative solution?

Would putting !Zap into Package Manager be a viable alternative solution?

For what? Package Manager isn’t a source control system.

Zap is already on GitHub.com, and the ReadMe is clear that should anyone wish to take it on, they should talk to the former(?) developers about handing over the code. The stumbling block last time was that those who have done some of the work up to now were vocal in their opposition to keeping the code in Git to enable collaborative devlopment; there is a strong hint in the notes above, however, that this could well be a (very valid and reasonable) prerequisite.

I’d suggest that someone still needs to talk to the Zap developers about this, and find a way forward to reunite all of the different patched versions that have been produced.

PS. Sprow said all of this, and more, back in 2016.

I like to thank those individuals (Rick, Julie,…)

Thank you – but thanks mostly due to Tank, Julie, and Fred for finding the complicated problems.
Me? I’m just fiddling with it so it keeps working on my machine.

For what? Package Manager isn’t a source control system.

And that’s not even what the quoted poster was saying – " without knowing that Zap has a legal future ".
Which implies it’s not whether or not the source is available so much as having the licence terms actually defined.

should anyone wish to take it on, they should talk to the former(?) developers about handing over the code

My versions, probably like Tank’s, are not any sort of official acceptance so much as “tweaking it to work on my system” that has been beneficial to others.

were vocal in their opposition to keeping the code in Git

I’m not vocal about this issue ¹. I just don’t Git. I haven’t Git, ever, and it’s only recently that some sort of Git appeared on RISC OS. Plus, my two attempts at Git on the PC were abject failures, a lot of online information is contradictory gibberish, and even Randall points out that Git can be a complicated pain in the arse.
Zip archives may have no revision comparison and be so last decade, but they’re a hell of a lot less complicated to deal with.

However… I agree, it would be good to have somebody overseeing any development on Zap (even if they’re only the custodian for patches submitted by others) because this and that and the other version on this or that or the other website is not ideal. It’s about as much of a mess as trying to find the right (working) build of Samba ²…

At least, to aid in this respect, I have put my version on !Store so that people who want it can get it from a centralised location instead of hunting for it.

¹ You want me to be vocal? How about the stupidity of having an incarnation of GitLab displaying the RISC OS sources in a nice webified format that cannot be viewed on the most commonly used RISC OS browser. That’s worth pointing out loudly. ;-)

² Which, weirdly, appears to have failed as of November. My phone can access RISC OS Samba shares without a problem. However my XP box can connect, but attempts to look at the files within causes an error reported on XP as the server doesn’t support this operation.
I fired up the EeePC (which hasn’t been used in many months) and it did the same thing, so… uh…?

And that’s not even what the quoted poster was saying – " without knowing that Zap has a legal future ".

You forgot a bit, Rick: “e.g. could be included in ROOL’s distrubtion, and can be hosted on GitHub by us to manage contributions”.

but they’re a hell of a lot less complicated to deal with.

Not for an Open, collaborative project, they’re not.

I agree, it would be good to have somebody overseeing any development on Zap (even if they’re only the custodian for patches submitted by others) because this and that and the other version on this or that or the other website is not ideal.

So, a serious question: would anyone else be interested in collaboratively doing this? I’m not going to do it alone, but if there are a few others willing to use GitHub and share the handling of any submissions and admin, I might be prepared to look into what would be required in the new year. I feel that any approach to the Zap folk would need to be a team effort, however, as I don’t think that it would be useful to saddle any one person as “Zap Maintainer”.

Rick? Julie? Anyone else?

I am comfortable with Git ⁰ and GitHub, and insofar as I can be of use I am happy to help, but I am not a Zap user, and I am emphatically not a competent Risc OS developer, so I cannot in good conscience take command of code review and the like.

⁰ I mean, reasonably comfortable. I still need to look things up regularly, because (g)it is a beast. I am more comfortable with it than any other VCS paradigm though. I have worked with mercurial quite a bit as well, but I once had a (very memorable) nightmare after a particularly nasty three-way merge…

So, a serious question: would anyone else be interested in collaboratively doing this? I’m not going to do it alone, but if there are a few others willing to use GitHub and share the handling of any submissions and admin, I might be prepared to look into what would be required in the new year. I feel that any approach to the Zap folk would need to be a team effort, however, as I don’t think that it would be useful to saddle any one person as “Zap Maintainer”.

Collaboratively: yes, though I really don’t take well to Git. It would mean that I have to get at least a minimal level of competence with it, though, and the responsibility should give me a firm push in that direction.

So, a serious question: would anyone else be interested in collaboratively doing this?

Yes, let’s do it.

Which implies it’s not whether or not the source is available so much as having the licence terms actually defined.

Rick’s on the money here. The licensing is key, everything else comes after.

When I did the original changes I asked the developers if they wanted the new version. At that time they said that they no longer developed !Zap, but allowed me to put my version on my web site and linked to it from their tartarus.org web site. My mods were very belt and braces, just to get it working on the Beagleboard (at the time).

@ Steve Fryatt

Rick? Julie? Anyone else?

Sure, on github we can also create a community repo, for instance RISC OS Community and create teams, so no limitations on that regards.

As always I am happy to work in teams and support which ever project is considered useful. My limitation is time (due my “day job” which is already very demanding), but if I can help with my limited time I am in :)

Also does anyone still have a copy of tEEDit?

Can’t recall ever having seen that editor at all.

Maybe he means this: https://ksquiggle.neocities.org/asmutil/tedit.htm

tEEDit is the tedious version of tedit.

Maybe he means this: https://ksquiggle.neocities.org/asmutil/tedit.htm

Nah, couldn’t be:
Wildly non-standard (any standard) key shortcuts with non-intuitive allocation like ^C for “Erase text in memory” – What!? ^X for directory, right, that’s because you threw away ^D on a cursor down…
30 minutes¹ of using that, and I’d be saving up to hire a hit man.

Info says the author is a “Computer Science major at the University of Wisconsin” so that narrows the search area.

¹ Probably not make it that far