| dkr | This is perl, v5.8.8 |
| twb | Aw, cperl-perldoc doesn't have completion like bash does. |
| dkr | Caelum: weird, it's not deterministic though, ran that exact command above multiple times, and sometimes it doesn't sefault |
| QtPlatypus | dkr: Is it possable that the regular expression is degenerate and sometimes causes perl to allocate more memory then is advalable? |
| roadfish | The game frozen-bubble is broken. I cut down the perl program to: use SDL::App; print SDL::App->new(); any tips on how to debug this SDL? |
| Caelum | dkr: doesn't segfault here... dkr: maybe you have bad ram in the server |
| dkr | Caelum: that commandline test was on my workstation, not the server, :) |
| Caelum | oh... |
| dkr | QtPlatypus: any idea how to test if that is the case? |
| Caelum | dkr: ahh, it doesn't segfault on amd64, just tried it on i386 and it does segfault dkr: err, actually what I meant to say was, it doesn't segfault on bleadperl, but does on 5.8.8 too bad 5.10 isn't out |
| Woosta | if I open a file and then while( defined( my $line = <fh> )) over it, I'll never end up with the whole file in memory right? just each $line at a time? |
| dkr | hrm, I guess I could test in chunks, but with multibyte chars chopping it up in arbitrary palces might be bad |
| Limbic_Region | Woosta - that depends |
| dondelelcaro | Woosta: depends on what $/ is |
| Caelum | Woosta: not exactly one line at a time because of buffering and such, but yeah |
| Limbic_Region | if there is only one line, then you will end up with the whole file in memory |
| Woosta | Caelum: ta |
| Limbic_Region | if $/ is set to something not in the file, you will end up with the whole file in memory |
| Woosta | It's a maillog so many many 'lines' |
| Limbic_Region | if $/ is undef, you will end up with the whole file in memory and if the file size is less than perl' er if the file is smaller than perl's buffer, you will end up with the whole file in memory regardless of the number of lines |
| dkr | I wish DBD::mysql worked that way, :) also having trouble with usng too much ram getting results back from db queries, even when fetched a row at a time, :) |
| Khisanth | damn pedants ;) |
| Woosta | Khisanth: yeah, I stopped listening an hour ago :-P |
| Limbic_Region | hrm, I can't seem to think of any more ways the whole file would end up in memory |
| Khisanth | dkr: there is an option to change that |
| Limbic_Region | I must be slipping |
| dondelelcaro | Limbic_Region: gnomes? |
| dkr | Khisanth: DBI or mysql specific? |
| Caelum | dkr: there must be a module or something to test for unicode |
| Limbic_Region | dondelelcaro - sure but I would want it to be repeatable |
| Khisanth | dkr: the option is "mysql_use_result" :) |
| dkr | Caelum: yeah, was looking at http://search.cpan.org/~markf/Test-utf8-0.02/lib/Test/utf8.pm btu not exactly what I need. I should look at the source anyway though :) |