Beefy Boxes and Bandwidth Generously Provided by pair Networks
No such thing as a small change
 
PerlMonks  

The number of ways I've independently discovered to make Perl segfault:

by Petruchio (Vicar)
on Sep 24, 2003 at 15:45 UTC ( #293886=poll: print w/ replies, xml ) Need Help??

vote on The number of ways I've independently discovered to make Perl segfault:

0
[bar] 279/49%
1
[bar] 95/17%
2
[bar] 51/9%
3
[bar] 33/6%
4
[bar] 10/2%
5-9
[bar] 16/3%
10 or more
[bar] 16/3%
Find them? Heck, I'm probably responsible for them.
[bar] 65/12%
565 total votes
Comment on The number of ways I've independently discovered to make Perl segfault:
Re: The number of ways I've independently discovered to make Perl segfault:
by fraktalisman (Hermit) on Sep 24, 2003 at 16:01 UTC
    Maybe I made Perl crash but didn't notice as I mainly use Perl scripts on my provider's webserver.

      Reason: BrowserUk Delete: I can't afford a loan.

      For more information on this node visit: this

Re: The number of ways I've independently discovered to make Perl segfault:
by hardburn (Abbot) on Sep 24, 2003 at 16:46 UTC

    Only one I could think of at the moment: trying to use PeekPoke on Cygwin.

    ----
    I wanted to explore how Perl's closures can be manipulated, and ended up creating an object system by accident.
    -- Schemer

    Note: All code is untested, unless otherwise stated


      Reason: BrowserUk Delete: I'm seeing doube trouble!

      For more information on this node visit: this

Re: The number of ways I've independently discovered to make Perl segfault:
by enoch (Chaplain) on Sep 24, 2003 at 17:07 UTC

       I can do it in no code!

       ActivePerl Build 635
       Built under MSWin32
       Compiled at Feb  4 2003 15:34:21
      

       With the following command:

      perl -P
      

       "Perl has generated errors ..." == a segfault in windows terms.

       Under Unix I've not seen Perl segfault unless I was playing with threads and fork()ing.

      Steve
      ---
      steve.org.uk
Re: The number of ways I've independently discovered to make Perl segfault:
by dws (Chancellor) on Sep 24, 2003 at 17:35 UTC
    I've discovered ways to make XS-based modules segfault, but not Perl proper. Since it'd be unfair to point the finger at Perl for some third-party's C problems, I have to vote zero.

      I've had particular success with segfaulting Perl 5.6.1. My latest was an overload which returned an undef as a string. Worked fine under 5.8.0...

      Another good place to look for segfaults is in the RE engine.

Re: The number of ways I've independently discovered to make Perl segfault:
by vili (Monk) on Sep 24, 2003 at 17:37 UTC
    Segfault, thats unPerl. unless, the user is being undude.
    ref. "The big Lebowski"

    ~vili
    "You are out of your element Donnie, the rug's not the point here Dude"
    -John Goodman
If I did...
by DentArthurDent (Monk) on Sep 24, 2003 at 19:16 UTC
    I don't recall ever doing it, but I can assure you that if I have, it was certainly my own fault for doing something stupid! ;)
Re: The number of ways I've independently discovered to make Perl segfault:
by chromatic (Archbishop) on Sep 24, 2003 at 19:32 UTC

    Okay, I'm not responsible for any segfaults in Perl proper, but I have really botched up the optree a couple of times with the proper modules. On the plus side, Artur told me that he has had the same experience with B::Generate — and he maintains it!

Had to vote for 3
by tbone1 (Monsignor) on Sep 24, 2003 at 20:53 UTC
    And Saint Merlyn raised the Holy Code on high, saying "Oh Lord Wall, please bless this thy Holy Hand Grenade of Perl code, so that thine followers' processes may be blown to bits, in thy mercy." And The Lord Wall did grin, and the people did feast on the lambs, and stoats, and orangutans, and fruit bats, and large chu-- ...

    (Skipping ahead a bit ...)

    And the Lord Wall spake "Three shall be the number of the segfaults, and the number of the segfaults shall be three. Four shall there not be, neither shall there be two, unless thou then proceedst to make segfault three. Five is right out!

    ...

    Oh, I've had enough of this. I always wanted to be ... a lumberjack!!!

    --
    tbone1
    Ain't enough 'O's in 'stoopid' to describe that guy.
    - Dave "the King" Wilson

Re: The number of ways I've independently discovered to make Perl segfault:
by Moriarty (Abbot) on Sep 24, 2003 at 23:48 UTC

    C I can get to seg-fault quite regularly, but I haven't done any serious programming in Perl ..... yet :-)

Re: The number of ways I've independently discovered to make Perl segfault:
by jonadab (Parson) on Sep 25, 2003 at 04:30 UTC

    I find it interesting that when I saw this poll, I was in the process of reviewing an old node of mine that mentioned one way to get 5.6 to dump core.

    I've also recently discovered an incredibly effective way to get Perl to dump core: run it on a filesystem that's corrupted in a way that causes stuff in general to dump core. (Perl dumps core... Emacs dumps core... ls dumps core sometimes...)


    $;=sub{$/};@;=map{my($a,$b)=($_,$;);$;=sub{$a.$b->()}} split//,".rekcah lreP rehtona tsuJ";$\=$ ;->();print$/
Re: The number of ways I've independently discovered to make Perl segfault:
by castaway (Parson) on Sep 25, 2003 at 05:15 UTC
    Currently by using perl 5.8.1-RC5 and ithreads.. still not sure quite why (but still, it only segfaults on exit, 5.8.0 was doing it on share() .. )

    And another before that, but I forget what that was..

    C.

Re: The number of ways I've independently discovered to make Perl segfault:
by Phaysis (Pilgrim) on Sep 26, 2003 at 00:35 UTC
    Although it's not particularly all-Perl, I have made it segfault/crash about once or twice in a coding session about a year ago. See, there's something that just doesn't work right when you mix AS 5.6 and ImageMagick on Win98se. Was it I::M? Or was it Perl? My mind boggles.

    One of these days, I'm gonna use a proper platform for that. :grins:

    (Ph) Phaysis (Shawn)
    If idle hands are the tools of the devil, are idol tools the hands of god?

Re: The number of ways I've independently discovered to make Perl segfault:
by scrottie (Scribe) on Sep 28, 2003 at 12:16 UTC
    sub f { my $x = shift; *z = $x; } f({}); f();
    Dave Mitchell solved this within a week and gave this description: "the second call to f() has $x as type RV (due to the previous call), but !ROK - due to the undef assignment. sv_setsv_flags() didn't check for ROK in this context." This was discovered while writing Object::Lexical. It stuffs closures into a stash created to represent an object - each object is its own blessed stash. I assumed (incorrectly) that the problem had something to do with blessing stashes. The fix is in 5.8.1. Hats off to the Perl 5 Porters! You guys rock! -scott
Re: The number of ways I've independently discovered to make Perl segfault:
by Aristotle (Chancellor) on Sep 30, 2003 at 20:36 UTC
    I've made vanilla Perl (ie no non-Perl extensions loaded) segv a number of times. I voted three only because I can't remember precisely; in reality it's probably more like 5 or 6. Almost all of them involved slushing around in the symbol tables one way or another (local trickery ranks high on the list); one of them was DESTROY related. I never took notes though. (In retrospect - maybe I should have perlbugged them?)

    Makeshifts last the longest.

Re: The number of ways I've independently discovered to make Perl segfault:
by shenme (Priest) on Oct 01, 2003 at 11:30 UTC
    I need to change my answer, to one.   Perl v5.6.1 for i386-linux using modules CGI, CGI::Application, CGI::FormBuilder, CGI::Session, and (depressing the dromedary's dome) Storable, as in
    $session = CGI::Session->new( 'driver:File;serializer:Storable', $q, {Directory=>'/tmp'} );
    But throwing in a gratuitous $session->close() cured all that.   Sigh....
Re: The number of ways I've independently discovered to make Perl segfault:
by Anonymous Monk on Oct 01, 2003 at 11:36 UTC
    I've found two, but they are my fault. One way was to call an ioctl function with too short input buffer, but then perl printed a nice warning about this and dumped core only after that. (This one was because I wrote 8092 instead 256*32, I must admit that perl can multiply better than me). The other was by calling the regexp engine inside a /(?{CODE})/ construct.
Re: The number of ways I've independently discovered to make Perl segfault:
by photon (Novice) on Oct 02, 2003 at 14:13 UTC
    i've crashed perl a lot of times. but fortunately i can not remember how ;) what i can say is that it was almost everytime my own fault. now i seem to be able to write less dirty code because i even can't remember when the last core dump was ;) most times it had to do with bloat caused by inefficient recursions or eating a lot of memory etc. that really isn't perl's fault. i like perl because it's fast and stable. reliable on many platforms. but of course only if your code is clean ;)
Re: The number of ways I've independently discovered to make Perl segfault:
by etcshadow (Priest) on Oct 04, 2003 at 20:34 UTC
    I understand that some of this may have been fixed in more recent versions of perl (this is in 5.005), but forking and signal handling seems to do the trick, fairly often. Maybe somewhere around one in twenty executions I can get a segfault.
Re: The number of ways I've independently discovered to make Perl segfault:
by gwhite (Friar) on Oct 06, 2003 at 13:22 UTC
    Those are caused by _my_ code... Huh, never made the connection...
    g_White
Here's my bit
by aquitaine (Novice) on Oct 08, 2003 at 01:01 UTC
    perl -wle ' print unpack ("p*", "202.54.9.1");' Null terminated string pointer , got it by accident.
      I've never crashed it, I mainly use Perl in *nix :-P, even using sockets and TCP/IP stuff, maybe Im just too conservative while coding

        Me thinks I am also to conservative while coding to create such errors. They are definitely cool to check out and figure out why they do what they do.

        Personally for me, it gives me a greater understanding of what is actually happening behind the scenes. Sorry no new code!

View List Of Past Polls


Log In?
Username:
Password:

What's my password?
Create A New User
Chatterbox?
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others avoiding work at the Monastery: (4)
As of 2014-09-21 11:46 GMT
Sections?
Information?
Find Nodes?
Leftovers?
    Voting Booth?

    How do you remember the number of days in each month?











    Results (168 votes), past polls