AmigaOS 3.1.4 withdrawn from market after Amiga

[kolla]

Originally Posted by McTrinsic

I beg to differ. There was a huge significant effort of testing going into the release. The situation has vastly improved compared to 3.1 ROMs.

Beg all you like, factum is that the shell that ships with KS 3.1.4 has a serious bug that will forever haunt the users and developers.

3.1 is well known, its flaws and bugs are well understood, patches and fixes are plentyful. It will take years to reach this level of understanding for the current 3.1.4 kickstarts, and only a fraction of users are upgrading anyways. I already have scripts that checks for shell v46 and bail out with EndCLI if such is the case (something weird even with internal Quit it seems, in v46).

[malko]

Picture from Cloanto's site... funny to see how they promote their emulator... with a nice and fully coloured WB... from Hyperion
OK, easy joke... nevertheless
(I have lowered the resolution a bit not to kill my attachment space)

[idrougge]

Originally Posted by kolla

Beg all you like, factum is that the shell that ships with KS 3.1.4 has a serious bug that will forever haunt the users and developers.

Please expand.

[kolla]

Originally Posted by idrougge

Please expand.

Shell v46 that ships in the 3.1.4 kickstart does not respect failat levels, so scripts keep running even when commands fail.

[McTrinsic]

Originally Posted by kolla

Shell v46 that ships in the 3.1.4 kickstart does not respect failat levels, so scripts keep running even when commands fail.

And there is no possible way around that?

[kolla]

Originally Posted by McTrinsic

And there is no possible way around that?

Feel free to think out a way around that also works with existing scripts and when people just insert a random old floppy.

The work-around is to fix and ship new kickstarts to all customers. Instead it will probably be tried to fix in a new SetPatch, which then all users would need to stuff on all their boot filesystems, floppies etc. and what to do if Setpatch fails? Or hopefully through a new ROM module, and we are back to rebooting and loading modules again...

[wawa]

@kolla
sry 4 ot, but can you look into your pms?
thx

[idrougge]

Originally Posted by kolla

Shell v46 that ships in the 3.1.4 kickstart does not respect failat levels, so scripts keep running even when commands fail.

Wow, how could that pass?

I can just imagine wasting several hours making a correct script behave, then giving up thinking you're an idiot.

[jurassicman]

Originally Posted by idrougge

Wow, how could that pass?

I can just imagine wasting several hours making a correct script behave, then giving up thinking you're an idiot.

I think that I've found this problem a few days ago using pipe
I give a command like the following:

cd SYS: || list

or:

 
cd SYS: && list

(is it possible to add set _mchar "&&" in S:shell-startup file instead of set _mchar "||" )


What I get is the list of the current directory.

[malko]

Originally Posted by kolla

[...] factum is that the shell that ships with KS 3.1.4 has a serious bug that will forever haunt the users and developers. [...]

Without all these court "things" they could already be in a position to propose corrective fix.

Regarding the linked document, I have to say that since UK is not my mother-tongue it's more than probable that I understand the documents in a slightly different way as some of you here in EAB (that are native UK speaking people). Anyway, after a second reading I can easily understand the agreement in the same way as Hyperion did. The way it is written make me think that the OS number was set to 4 "like this" (as an example). They just said 4 like they could have said 3.1.2 or 3.2 or 3.3 or etc.
Again it's the understanding from a non native UK speaker. I don't want to rewrite the story, just to share my understanding.

[kolla]

Originally Posted by jurassicman

I think that I've found this problem a few days ago using pipe
I give a command like the following:

cd SYS: || list

or:

 
cd SYS: && list

(is it possible to add set _mchar "&&" in S:shell-startup file instead of set _mchar "||" )


What I get is the list of the current directory.

First of all, in v46, _pchar and _mchar variables are no longer used, they are hard coded to | and ||.

Secondly, external "pipe" command is no longer used, instead piping is done internally, (apparently using "jobs", as you can see with "set echo on", it uses &, and this also requires CONSOLE: to be available, so you can no longer run scripts in the background that uses pipes... meh.)

Thirdly, I think what you describe is correct - you run two commands in parallel, one cd sys:, which is sileny, and one list, that just lists out current directly. I haven't really used this feature much, did it behave differently in earlier shell versions?

[Gorf]

Originally Posted by kolla

First of all, in v46, _pchar and _mchar variables are no longer used, they are hard coded to | and ||.

not really an improvident

Secondly, external "pipe" command is no longer used, instead piping is done internally, (apparently using "jobs", as you can see with "set echo on", it uses &, and this also requires CONSOLE: to be available, so you can no longer run scripts in the background that uses pipes... meh.)

wtf ... who made that decision and WHY???

[Mr.Flibble]

Originally Posted by malko

Without all these court "things" they could already be in a position to propose corrective fix.

Regarding the linked document, I have to say that since UK is not my mother-tongue it's more than probable that I understand the documents in a slightly different way as some of you here in EAB (that are native UK speaking people). Anyway, after a second reading I can easily understand the agreement in the same way as Hyperion did. The way it is written make me think that the OS number was set to 4 "like this" (as an example). They just said 4 like they could have said 3.1.2 or 3.2 or 3.3 or etc.
Again it's the understanding from a non native UK speaker. I don't want to rewrite the story, just to share my understanding.

The contract will be taken as a whole rather than a single passage, so taking some wording in the contract and twisting it into something which the other parties didn't agree to, and which runs contradictory to the rest of the contract could render the whole agreement invalid (The "Meeting of minds" quote in one of the documents).

It'll be interesting to see how Hyperion respond.

[jurassicman]

Originally Posted by kolla

First of all, in v46, _pchar and _mchar variables are no longer used, they are hard coded to | and ||.

Secondly, external "pipe" command is no longer used, instead piping is done internally, (apparently using "jobs", as you can see with "set echo on", it uses &, and this also requires CONSOLE: to be available, so you can no longer run scripts in the background that uses pipes... meh.)

Thirdly, I think what you describe is correct - you run two commands in parallel, one cd sys:, which is sileny, and one list, that just lists out current directly. I haven't really used this feature much, did it behave differently in earlier shell versions?

I'm not aware of the technical aspects inside the Amiga Shell but yes, in the previous version I was able to execute two commands on the same line using "&&" and in fact I wrote a few scripts that did so and that now don't work anymore.


I've also faced the "CONSOLE" request running one of them and I "fixed" that with a silly assign on my startup-sequence

Assign >NIL: CONSOLE: DH0:

I know it's just a workaround but I really don't know a better solution.

[malko]

Originally Posted by Mr.Flibble

The contract will be taken as a whole rather than a single passage, so taking some wording in the contract and twisting it into something which the other parties didn't agree to, and which runs contradictory to the rest of the contract could render the whole agreement invalid (The "Meeting of minds" quote in one of the documents). [...]

You are right. We are missing important informations that doens't allow us to have a detailed view of the story. Would be interesting if these contract were available for reading.

Originally Posted by Mr.Flibble

[...] It'll be interesting to see how Hyperion respond.

Yep.

[Mr.Flibble]

Originally Posted by malko

You are right. We are missing important informations that doens't allow us to have a detailed view of the story. Would be interesting if these contract were available for reading.

Hmm, it should be I think as it's a submitted document to the court...

[malko]

Found this but it's not free

PACER Fees and Case Currency

Access to case information costs $0.10 per page. The cost to access a single document is capped at $3.00, the equivalent of 30 pages. The cap does not apply to name searches, reports that are not case-specific, or transcripts of federal court proceedings.

By Judicial Conference policy, if usage does not exceed $15 in a quarter, fees are waived.

Newly filed cases will typically appear on this system within 24 hours. Check the Court Information page for data that is currently available on the PCL. The most recent data is available directly from the court.

Following this link (from another post), I can see that there are a lot of documents to read and for all those I clicked on, I was asked to register (with a payment plan). Unless another solution to access the documents free of charge (as unlike a lawyer, I don't get paid for it ), think I will wait for official news.

[McTrinsic]

Well sounds like the situation we had since 3.1 release. Like adding a big Harddrive without a patched SCSI.device and / or FFS and / or HDToolBox.

Nothing new to me.

E.g. take the old shell and add it via LoadModule.

Furthermore the new shell does respect most failats - not all, obviously.

Originally Posted by kolla

Feel free to think out a way around that also works with existing scripts and when people just insert a random old floppy.

The work-around is to fix and ship new kickstarts to all customers. Instead it will probably be tried to fix in a new SetPatch, which then all users would need to stuff on all their boot filesystems, floppies etc. and what to do if Setpatch fails? Or hopefully through a new ROM module, and we are back to rebooting and loading modules again...

[Mr.Flibble]

Originally Posted by malko

Found this but it's not free


Following this link from another post, I can see that there are a lot of documents to read and for all those I clicked on, I was asked to register (with a payment plan). Unless another solution to access the documents free of charge (as unlike a lawyer, I don't get paid for it ), think I will wait for official news.

I think it's referring to this: https://docs.google.com/file/d/0BxlY...0M2I4MzQ1/edit

The language is rather confusing and we probably haven't a hope of properly understanding it without a law degree...

[kolla]

Originally Posted by McTrinsic

E.g. take the old shell and add it via LoadModule.

Right, which is what I do. Now that Hyperion, with Thor's blessing, is in total ownership, I doubt there will be more aminet updates though. So we have yet another incarnation of the shell to worry about in scripts.

Furthermore the new shell does respect most failats - not all, obviously.

Really, I have not seen it respect any failar level yet - can you provide details? How high must one go to make it abort?