nnn@po.CWRU.Edu (Nik N. Nik Mahdi) (01/31/91)
Hello, You know, I have been wondering about my Amiga, which sometimes gurus when I run certain softwares. I wonder what's wrong with it. Is it the computer system, or is it the software that causes it to Guru? Why doesn't it happen to other computer systems, like IBM or Macintosh (sometimes I got "System Error" on Mac, is it the same as the "Guru Meditation" on Amiga?). Thus, can someone explain this to me. Why the software developer (if it is their fault!) didn't do anything to avoid this? Thanks in advance. NIK -- kjfyrwhgbvcxxbmiltioyrazsgwqashhfdhyjhmnmygikuykjiyttqsgffhipkkur hgthtywqegzbvxmmjkjltiourwagagfgzrqtrtwythgshgiuitpukjhfrewqvcxzh rewqsafdyteurhgmnblkjpvcxzgfauytjhgiuytjhnhgsfiytkjhoifdsgfastrew vcxvfghgnnbjhrrtewuiyudsaseqhuyiliupoligvjfew niknorrosdinikmahdi
t22918@ursa (Matt Ranney ) (02/02/91)
david@dogmelb.dog.oz.au (David Le Blanc) writes: >Unix boxes 'segmentation fault' or 'core dump' or sometimes have >to reboot!, Macinsloshes give 'System Error'. IBM pc's simply lock, Why can't Amiga's handle failed tasks like the Unix OS can? I'd be perfectly happy to have my computer tell me that a task failed, and all it's allocated resources had been returned. I'm not an OS programmer, so I don't know why this would cause a problem, or how hard it would be, so could someone answer this? I had my Amiga for 45 minutes, and had GURU'ed it just by opening messing with things on the Workbench 1.3 and Extras disk. These are commercial programs, and I wasn't really doing anything out of the ordinary. -- Matt Ranney t22918@ursa.calvin.edu
djohnson@beowulf.ucsd.edu (Darin Johnson) (02/02/91)
In article <t22918.665454268@ursa> t22918@ursa (Matt Ranney ) writes: >Why can't Amiga's handle failed tasks like the Unix OS can? I'd be perfectly >happy to have my computer tell me that a task failed, and all it's allocated >resources had been returned. Herein lies the problem. The OS does not keep track of allocated resources. (sigh) However! the GOMF program (commercial, but old PD version on a Fish disk) will attempt to recover resources from failed tasks, remove the task, and let you continue on your way. It doesn't do a perfect job, but works most of the time, with only a little loss of memory. -- Darin Johnson djohnson@ucsd.edu - Political correctness is Turing undecidable.
dave@cs.arizona.edu (Dave P. Schaumann) (02/02/91)
In article <16264@sdcc6.ucsd.edu> djohnson@beowulf.ucsd.edu (Darin Johnson) writes: |In article <t22918.665454268@ursa> t22918@ursa (Matt Ranney ) writes: ||Why can't Amiga's handle failed tasks like the Unix OS can? I'd be perfectly ||happy to have my computer tell me that a task failed, and all it's allocated ||resources had been returned. | |Herein lies the problem. The OS does not keep track of allocated resources. |(sigh) However! the GOMF program (commercial, but old PD version on |a Fish disk) will attempt to recover resources from failed tasks, remove |the task, and let you continue on your way. It doesn't do a perfect |job, but works most of the time, with only a little loss of memory. Resource tracking is not the only problem. A buggy program can render vital operating system (or other running processes) structures hopelessly corrupt. Weren't any of you around the last time this thread made the rounds (finally died about two weeks ago)? We really need a FAQ list, if only for this one topic. |Darin Johnson |djohnson@ucsd.edu Dave Schaumann | And then -- what then? Then, future... dave@cs.arizona.edu | -Weather Report
ccplumb@rose.uwaterloo.ca (Colin Plumb) (02/02/91)
t22918@ursa (Matt Ranney ) wrote: >david@dogmelb.dog.oz.au (David Le Blanc) writes: > >>Unix boxes 'segmentation fault' or 'core dump' or sometimes have >>to reboot!, Macinsloshes give 'System Error'. IBM pc's simply lock, > >Why can't Amiga's handle failed tasks like the Unix OS can? I'd be perfectly >happy to have my computer tell me that a task failed, and all it's allocated >resources had been returned. I'm not an OS programmer, so I don't know why >this would cause a problem, or how hard it would be, so could someone answer >this? For cost reasons, the Amiga wasn't built with an MMU, so there is nothing the OS can do to stop one program from scribbling all over another. There is an upside to this, as Amiga message-passing and I/O is *much* faster than Unix. But it means there's no sure-fire enforcement of the rules. As for resource tracking, it's sorely missed. Apparently there were plans to include it, but the Amiga's gestation was rushed by an urgent need to sell them for money, and it got, unfortunately, lost when Tripos was invited in to do the file system. -- -Colin
rlcollins@miavx1.acs.muohio.edu (Ryan 'Gozar' Collins) (02/05/91)
In article <156@dogmelb.dog.oz.au>, david@dogmelb.dog.oz.au (David Le Blanc) writes: > In article <1991Jan31.035105.14277@usenet.ins.cwru.edu>, nnn@po.CWRU.Edu (Nik N. Nik Mahdi) writes: >> gurus when I run certain softwares. I wonder what's wrong with > It is the software ** OR ** the interaction between different pieces > of software. If the software is public domain, its probably bug > ridden! (though I feel that alot of PD programs recently are generally > more robust than what we got in 1986..) If it is commercial, you [Stuff Deleted] > decided I wanted to. ( I feel safe, because one of the first things > I do to PD programs is force them to crash by doing stupid things > to them. If it crashes violently, or locks the system, I usually dont > want to know about it, after all, why should I have to be gentle to > a bug ridden program!:-).) [Stuff Deleted] > It can be your fault for running a PD program in the background. > At a software house I worked at, PD programs were *BANNED* because > bugs introduced by them were considered harmful, and unneccesary. > We had to use commercial editors/compilers/assemblers and utilities > in general. But that's fine, because what you pay for is usually > better than what you get for free.. Are Programs so hard to write for the Amiga that every PD prg is bug ridden? Or is a bad experience with a minute few. I've used many other computer system, with tons of PD/Shareware sw, and I've only had problems on very RARE occasions. Case in point, on my ST, I really only use one or two pieces of commercial software, the rest is PD/Shareware, and have never really had many problems. (I know the chance of crashing is greater in a multi-tasking Environment, especially without hardware protection, but wouldn't the programmers know how to deal with this?) How hard is it really to prg for the Amiga? (Just in case I get a new computer someday!) ----------------------------------------------------------------------------- Ryan 'Gozar' Collins rlcollins@miavx1.BITNET ("UUdecode and Type with vt100 emulation") begin 666 SIG.LZH M'.<M;&@Q+5D```#E````CXX_%B``!E-)1RY65&V*C``$:S$Q0%EI4$[N~"[Va M'=]&%?!$>"HT&:?S\'FU@3A<R;`D8&*@,5LK49'<_8'\?]=\?U`/I4@,_'!&a ><*>"BIY*JFI/?'ULXCLX_MBQ(~?%_/\H9<R3_ZP`a a end -----------------------------------------------------------------------------
rjc@geech.ai.mit.edu (Ray Cromwell) (02/05/91)
In article <3680.27ad65e7@miavx1.acs.muohio.edu> rlcollins@miavx1.acs.muohio.edu (Ryan 'Gozar' Collins) writes: >Are Programs so hard to write for the Amiga that every PD prg is bug >ridden? Or is a bad experience with a minute few. I've used many other >computer system, with tons of PD/Shareware sw, and I've only had problems >on very RARE occasions. Case in point, on my ST, I really only use one or >two pieces of commercial software, the rest is PD/Shareware, and have never >really had many problems. (I know the chance of crashing is greater in a >multi-tasking Environment, especially without hardware protection, but >wouldn't the programmers know how to deal with this?) > >How hard is it really to prg for the Amiga? (Just in case I get a new >computer someday!) All I can say is, I have over 200 floppies filled with PD stuff from various sources (fred fish for one) and 90% of that software works flawlessly. I think PD software is the Amiga's strong point. Mostra:Best IFF viewer around, better than any commercial viewer DME/Qed/UEdit: Some of the best editors around VLT/Handshake/Jrcomm:All fine term prgs which IMHO beat commercial crap like Online, Access, Diga, etc. There's alot of nonsense going around concerning FreeWare!=Good software. (you don't get something for nothing!) This is just untrue. Alot of Commercial software starts out as FreeWare. Sometimes I get the impression that commercial software is buggier than my freeware. Reason? Some companies have deadlines, so products are rushed. I don't think the aMiga is hard to program, again, thats me. As for bugs, I think bad-software gets publicized more than 'good' software, hence it gives people the impression that a large portion of Amiga software is bad. I've had IBMs, Macs, and C64's lock up on me plenty of times, the reason the Amiga is infamous for crashing, is that the GURU stands out, its actually kinda funny. The Amiga crashes more specular then the IBM(Lockup), Mac/ST (bombs), so people take notice. Often I see IBM users quoting the Amiga's GURU when they don't even realize what it means. The GURU is informative to a certain extent, and helps debug software.
djohnson@beowulf.ucsd.edu (Darin Johnson) (02/05/91)
In article <3680.27ad65e7@miavx1.acs.muohio.edu> rlcollins@miavx1.acs.muohio.edu (Ryan 'Gozar' Collins) writes: >Are Programs so hard to write for the Amiga that every PD prg is bug >ridden? Or is a bad experience with a minute few. Probably, just those few programs... My current environment has lots of PD/Shareware programs running, with rarely any problem. Some of the things I run are inherently 'asking for trouble' - like input handlers - but work fine. I don't push things to their limits, so I don't find as many problems as other people. If there's a problem with PD software, I can usually upgrade to a new version. Commercial packages I am stuck with (like Sculpt 3D which fails to work correctly with my 68020, but which I can't afford to upgrade). I think part of the problem - which exists on most computers - is that there isn't a whole testing division shaking these utilities out before the user gets them. Also, the author may not be able to test the program on all the various combos of Amigas out there (such as machines with 1M chip ram, accelerators, etc). The early Macintosh computers solved part of this problem by trying to make all the machines look identical. Nowdays, I see them crash more often than my Amiga (even though I rarely use a Mac). Clones aren't a big problem, since recovering from a crash is usually very quick (branch to command.com and continue), but PD software on Clones still has a similar crash rate from what I've seen. As an example, I wrote a PD 'background' utility. I did this in my spare time, like most PD stuff is done. Within a week after giving it to some sample people, bugs were found that I wouldn't have found after a year of testing. Partly because I didn't have the particular setups that people had, and partly because I didn't run the program the same way. I also wasn't running all the various other utilities out there to find out incompatabilities (ie, I assumed they all followed the rules, so things would be safe). Even after fixing these problems, as soon as it became public, other problems popped up - enough menu items to overflow the screen, etc. People had crashes that I couldn't duplicate - making finding the bug hard. I also didn't expect that the program would be even remotely popular, or more time would have been spent verifying everything (it worked for ME, and if a few other people wanted it great, but midnight calls from Australia wanting upgrades?) I think these sorts of problems would have occured on most computers (crashes on UNIX systems would be benign though) - well.. on some computers, that type of utility wouldn't have even been attempted. >How hard is it really to prg for the Amiga? (Just in case I get a new >computer someday!) Well, harder then a UNIX filter, and easier than a UNIX device driver :-) Definately easier than early Macintosh versions, and definately less frustrating than current Mac versions. Depends upon what you want to do also. -- Darin Johnson djohnson@ucsd.edu - Political correctness is Turing undecidable.
kent@swrinde.nde.swri.edu (Kent D. Polk) (02/05/91)
In article <3680.27ad65e7@miavx1.acs.muohio.edu> rlcollins@miavx1.acs.muohio.edu (Ryan 'Gozar' Collins) writes: >How hard is it really to prg for the Amiga? (Just in case I get a new >computer someday!) I find it pleasantly easy and usually fun. The things I have trouble with I'd have trouble with on any machine. It's the mindset that most programmers are having trouble with. It appears to be a big jump for some to realise that the whole machine isn't there only for the one program they are working on. And then there are clutzes like me who can't keep track of structure pointers >4 levels deep :^( But neither do I call myself a programmer :^) ===================================================================== Kent Polk - Southwest Research Institute - kent@swrinde.nde.swri.edu "Would you rather die than think?" =====================================================================
chrise@hubcap.clemson.edu (Chris Everhart) (02/05/91)
In article <3680.27ad65e7@miavx1.acs.muohio.edu>, rlcollins@miavx1.acs.muohio.edu (Ryan 'Gozar' Collins) writes: > Are Programs so hard to write for the Amiga that every PD prg is bug > ridden? Or is a bad experience with a minute few. I've used many other > computer system, with tons of PD/Shareware sw, and I've only had problems > on very RARE occasions. Case in point, on my ST, I really only use one or > two pieces of commercial software, the rest is PD/Shareware, and have never > really had many problems. (I know the chance of crashing is greater in a > multi-tasking Environment, especially without hardware protection, but > wouldn't the programmers know how to deal with this?) Well, actually, simple programming of the amiga is quite easy with the appropriate compiler. However, when you mess with the operating system, as most of the really useful pd progs do, you can quite easily run into problems with multitasking. Other computers, like a VAX or some other such, are much easier to write programs for because if your program does something nasty, it aborts and tells you. On the Amiga, it will keep on running unless it really messes up bad. Unfortunately, it might run 99 out of 100 times in a seemingly perfect manner, and the 100th time it could lock up the system. If you've ever seen code for parallel processing that handles critical regions, you would be able to understand just how hard multiprogramming can get. And, actually, multiprogramming on other computers is just as hard. For instance, I'm working on a multitasking operating system for the IBM compatibles, and there are all sorts of problems with it. Unlike the amiga, it doesn't guru...it just locks up. I also can't use any DOS services unless I serialize them because they weren't written with multitasking in mind. Anyway, hope this all helps. Chris Everhart chrise@hubcap.clemson.edu
bairds@eecs.cs.pdx.edu (Shawn L. Baird) (02/07/91)
djohnson@beowulf.ucsd.edu (Darin Johnson) writes: >In article <t22918.665454268@ursa> t22918@ursa (Matt Ranney ) writes: >>Why can't Amiga's handle failed tasks like the Unix OS can? I'd be perfectly >>happy to have my computer tell me that a task failed, and all it's allocated >>resources had been returned. >Herein lies the problem. The OS does not keep track of allocated resources. >(sigh) However! the GOMF program (commercial, but old PD version on >a Fish disk) will attempt to recover resources from failed tasks, remove >the task, and let you continue on your way. It doesn't do a perfect >job, but works most of the time, with only a little loss of memory. The reason that Unix almost never crashes is usually because of the memory protection hardware which is required to protect processes from tampering with memory areas they don't own. In the Amiga a program can rampantly wade through memory trashing areas. In Unix you'll get a segmentation violation and thus avoid crashing any of the other processes and also make it easy to clean up the dead process. A stock Motorola 68000 is not capable of attaching an MMU (Memory Management Unit), although the 68020 and 68030 chips are. I have heard of a program called Enforcer which uses the MMU on a 68020 or 68030 to provide a more protected environment. It also attempts to provide cleanup. However, since the original OS wasn't designed to use it, it isn't quite as effective as the Unix method. I'm also not sure whether it runs under 1.3, 2.0 or both. | Shawn L. Baird | Or via US Snail: | | bairds@eecs.ee.pdx.edu | 17650 SE Cason Rd. | | ...uunet!tektronix!psueea!eecs!bairds | Gladstone, OR 97027 |
sbeagle@kennels.actrix.gen.nz (Sleeping Beagle) (02/07/91)
rjc@geech.ai.mit.edu (Ray Cromwell) writes: > I don't think the aMiga is hard to program, again, thats me. As for > bugs, I think bad-software gets publicized more than 'good' software, hence > it gives people the impression that a large portion of Amiga software is bad. > I've had IBMs, Macs, and C64's lock up on me plenty of times, the reason > the Amiga is infamous for crashing, is that the GURU stands out, its actually > kinda funny. The Amiga crashes more specular then the IBM(Lockup), Mac/ST > (bombs), so people take notice. I think there's a lot of truth in that. Multifinder on a Mac seems to lend itself to bugs and crashes too much. I just installed Windows 3 as well and I've been getting quite a few crashes on that. What worries me about Multifinder is that it will crash with software that is known to be good when not running multifinder. The Amiga tends to have bad pieces of software which will happily crash you any old time... -- ** Official Signature for Sleeping Beagle (aka Thomas Farmer)! ** sbeagle@kennels.actrix.gen.nz || Disclaimers are for sick societies ** Thomas.Farmer@bbs.actrix.gen.nz || with too many lawyers.
jdickson@jato.jpl.nasa.gov (Jeff Dickson) (02/09/91)
In article <t22918.665454268@ursa> t22918@ursa (Matt Ranney ) writes: >david@dogmelb.dog.oz.au (David Le Blanc) writes: > >>Unix boxes 'segmentation fault' or 'core dump' or sometimes have >>to reboot!, Macinsloshes give 'System Error'. IBM pc's simply lock, > >Why can't Amiga's handle failed tasks like the Unix OS can? What happened to IBM PCs and Macintoshes? >I'd be perfectly >happy to have my computer tell me that a task failed, and all it's allocated >resources had been returned. I'm not an OS programmer, so I don't know why >this would cause a problem, or how hard it would be, so could someone answer >this? > >I had my Amiga for 45 minutes, and had GURU'ed it just by opening messing >with things on the Workbench 1.3 and Extras disk. These are commercial >programs, and I wasn't really doing anything out of the ordinary. > >-- >Matt Ranney >t22918@ursa.calvin.edu Software development on the Amiga can be a long drawn out process depending on how often the Amiga Guru's in the interim. But the Guru number is informative and often helps you steer in the right direction to fix the problem. Better than the competition - they just lock up snd don't tell you s___! What really bugs me, is that some of the Amiga utilities can cause the Amiga to guru. Take the note pad utility. If you select the menu item "save as" then "save", the machine will guru. Sure it doesn't make a whole lot of sense, but because one is above the other sometimes you can goof. Seems like some kind of resource tracking could be implemented since permission to use many resources is through a particular interface (i.e. OpenDevice(), OpenLibrary()). But this has flaws, because sometimes you just clone objects that correspond to devices (i.e. Serial Device). To do memory tracking seems expensive. I still slave unto my 7.1 MHZ A2000. I don't want the overhead of the OS to be increased much more than it is already in V2.0. Jeff a good deal of overhead (CPU and memory).
rshaw@theborg.mlb.fl.us (Ron) (02/09/91)
What makes it harder for new Amiga would be programmers to learn programming
on the amiga is the lack (especially in the past) of adequate information on
how to get started. I have found that most new C programmers in the amiga
environment (without any prior knowledge of C) actually though that they could
pick up an amiga specific C programming book or purchase SASC or Manx & learn
C language. They had a real shock when they realized that was not the case.
In this area we have quite a few IBM C programmers and students with amiga's
that took a C programming Couse & still were unable to grasp Amiga C.
IBM C programmers in the area, just gave up.
Hopefully with the new 2 tech disk/mag journals coming out, more of the
problems will be allieviated
Ron Shaw..... The only good 8 bit computer is a
Dead 8 bit compter....
-----------------------------Mathematics is a state of mind,
Electronics is a state of being.darrell@comspec.uucp (Darrell Grainger) (02/13/91)
In article <rshaw.0365@theborg.mlb.fl.us>, rshaw@theborg.mlb.fl.us (Ron) writes: > What makes it harder for new Amiga would be programmers to learn programming > on the amiga is the lack (especially in the past) of adequate information on > how to get started. I have found that most new C programmers in the amiga > environment (without any prior knowledge of C) actually though that they could > pick up an amiga specific C programming book or purchase SASC or Manx & learn > C language. They had a real shock when they realized that was not the case. > In this area we have quite a few IBM C programmers and students with amiga's > that took a C programming Couse & still were unable to grasp Amiga C. > IBM C programmers in the area, just gave up. > Hopefully with the new 2 tech disk/mag journals coming out, more of the > problems will be allieviated > > Ron Shaw..... The only good 8 bit computer is a > Dead 8 bit compter.... I have learned what I know about C language from reading the manual which come with a compiler and scanning over other people's source. I actually found the Amiga C easier to grasp then Turbo C on an MS-DOS machine. The hardest part with the MS-DOS machine was trying to get good examples of how to address the hardware without breaking portability rules (i.e. no using X = 0x278; Y = 10; *X = Y;). I guess I'm just different. Darrell Grainger
cmcmanis@stpeter.Eng.Sun.COM (Chuck McManis) (02/19/91)
(Sleeping Beagle) writes: >I think there's a lot of truth in that. Multifinder on a Mac seems >to lend itself to bugs and crashes too much. I just installed Windows >3 as well and I've been getting quite a few crashes on that. As most people find out, there are just as many if not more programs that crash on the Mac and PC when "multitasking" is present. More so because the original programmers of those applications didn't have to worry about the possibility that some other task might be running. Also the most common time for a PC program to crash is on exit because it has screwed something up. With Windows or MultiFinder or Desqview etc, these crashes become problems because it takes out the whole machine. Prior to this behaviour, most PC users hit ctrl-alt-del so often that some think it is the standard way to exit programs! -- --Chuck McManis Sun Microsystems uucp: {anywhere}!sun!cmcmanis BIX: <none> Internet: cmcmanis@Eng.Sun.COM These opinions are my own and no one elses, but you knew that didn't you. "I tell you this parrot is bleeding deceased!"