Re: Bus error -- some education sought

[Puneet Kishor]

Edward Moy wrote:

> On Sunday, April 28, 2002, at 07:12 AM, Puneet Kishor wrote:
>
>> I am developing a web application, and have started encountering a 
>> new error (taken from the error_log)...
>>
>> [Sun Apr 28 08:50:40 2002] [error] (22)Invalid argument: getsockname
>> [Sun Apr 28 08:50:40 2002] [notice] child pid 493 exit signal Bus 
>> error (10)
>> [Sun Apr 28 08:50:40 2002] [notice] child pid 482 exit signal Bus 
>> error (10)
>>
>>
>> [..]
>>
>> In any case, here are my questions --
>>
>> 1. am I right about my assumption about failure to clear a memory 
>> pointer?
>> 2. that the above error is new to me means that this is a problem 
>> specific to the module I am using, and not some new gremlin I have 
>> exposed in my iBook or the "Perl on MacOS X" implementation?
>
>
> I've seen the bus error problem a lot myself, and after trying to 
> workaround the problem, occurring even in standard packages, I 
> concluded that it is a real bug in Perl 5.6.0, the default version 
> that comes with Mac OS X.  Upgrading to 5.6.1 solved the bus error 
> problems for me.

sorry, I am running Perl 5.6.1.


David Wheeler wrote:

>
>I think that it may be Mac OS X related, because I've run into it, too, but
>couldn't nail it down. What I did find was that if I ran the script that
>triggered the bus error in MallocDebug, the bus error never occurred!
>
>But if you can write a short test script that demonstrates the error
>happening every time, it'd be worth sending in via perlbug. If you post it
>here, we can also test it on our OS X and other systems, too.
>  
>
I'll see if I can do that, hoewver, this is happening in conjunction 
with a lot of related things, and there is a lot of data that gets 
munged by the scripts.

Anyway, it is good to know that this is not a one-off problem. I have 
also alerted the writers of the perl module I am using that is causing 
this error to occur.

Thanks everyone,

pk/

Re: CPAN shell error with DELETE key

[Craig S. Cottingham]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Friday, April 26, 2002, at 10:54 , William H. Magill wrote:

> I normally work with Tru64 Unix on Alpha and with Alpha Linux, 
> and I can
> say without qualms that the Open Source community simply does 
> not support
> anything but Sparc or Intel.

That's a little unfair. For example, if there was no open source 
support for Alpha, there'd be no Alpha Linux. I suspect it's far 
more correct to say that open source *developers* don't actively 
support any hardware that the don't personally have access to, 
and unfortunately Alpha boxes are far less common than Intel 
boxes.

> And it simply cannot deal with the idea of a
> 64bit clean environment.

Same argument applies.

> It takes exceptional work by a very small cadre of very 
> dedicated people to
> provide all of the ports to all of the other platforms that exist.
> That any Open Source software exists at all for non-Sparc and non-Intel
> platforms is a minor miracle.

What you've just described is the essence of open source. I 
write some code, think it might be useful to others, and throw 
it out there for the world to use. I'm pleased if it can be used 
on architectures other than what I use, and I'll accept patches 
for other architectures if they don't break the code for mine, 
but I don't consider it my responsibility to ensure that my code 
works on every platform known to man. That's why I open sourced 
it.

Want more open source software ported to Alpha? Ship me an Alpha 
box. You'll get it sooner if you can pay me a full-time salary 
as well. :-)

- --
Craig S. Cottingham
[EMAIL PROTECTED]
PGP key available from: 

ID=0xA2FFBE41, fingerprint=6AA8 2E28 2404 8A95 B8FC 7EFC 136F 
0CEF A2FF BE41
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.0.6 (Darwin)
Comment: For info see http://www.gnupg.org

iD8DBQE8zWT2E28M76L/vkERAorXAJ0TuquL4J6rK2qkj3LNjb3wn1ynBACfSKpj
WKavFlcbW7wa12GvDI40BeM=
=jkBp
-END PGP SIGNATURE-

Re: re-opening pipes

[Kee Hinckley]

At 1:33 PM + 4/29/02, Alan  Drew wrote:
>I am using perl to pipe commands to a compiled binary (long story - 
>35,000 lines of badly written fortran, with an equally badly written 
>user interface - and not written by me!). I am using Filehandle to 
>open the pipe in perl, but there are occasions when I want to 
>restart the binary (i.e close the pipe and open it again) without 
>restarting the perl script. I am a relative newbie to perl, so 
>please don't get too upset if the answer is an easy one!
>Al.

$fh->close();
and then do whatever you did to open it the first time again.
Is that what you mean?
-- 

Kee Hinckley - Somewhere.Com, LLC
http://consulting.somewhere.com/
[EMAIL PROTECTED]

I'm not sure which upsets me more: that people are so unwilling to accept
responsibility for their own actions, or that they are so eager to regulate
everyone else's.

Re: puzzling results from missing she-bang

[Michael Turner]

Peter,

Thanks again. Now, I have it clear.

/Michael

>
>> Thanks for your response. I remain puzzled, however. Your response 
>> seems to imply that if I invoke "print" from the command line, it will 
>> be interepreted. This is not the case, it errors.
>
> Sure it will - if you are using the sh shell.  Try this:
>
> zany:~/unix/avr/avr/include% sh
> zany% print "hello";
> hello
> zany% exit
> zany:~/unix/avr/avr/include%
>
> (ie, type: shprint "hello";)
>
> Enjoy,
>Peter.

re-opening pipes

[Alan Drew]

I am using perl to pipe commands to a compiled binary (long story - 
35,000 lines of badly written fortran, with an equally badly written 
user interface - and not written by me!). I am using Filehandle to open 
the pipe in perl, but there are occasions when I want to restart the 
binary (i.e close the pipe and open it again) without restarting the 
perl script. I am a relative newbie to perl, so please don't get too 
upset if the answer is an easy one!
Al.