perl hunting

getting slash code

2012-01-30T14:17:00.000+02:00

this will work

git clone --branch live  git://slashcode.git.sourceforge.net/gitroot/slashcode/slashcode

diff between two mysql databases

2012-01-08T20:53:00.002+02:00

get differences between two iterations of the same database; requires the username and the password be set in ~/.my.cnf ; do not add the database name to ~/.my.cnf

... not perl, posting here so I'll have access to the script when not at home


#!/usr/bin/env bash

echo "" > differences.diff;

DB_ORIGINAL="db_old"
DB_MODIFIED="db_modified"

declare -A SKIP_TABLES
SKIP_TABLES[hits]=1
SKIP_TABLES[spam_reports]=1
SKIP_TABLES[searches]=1

for i in `mysql $DB_ORIGINAL --skip-column-names -e "show tables" | awk '{print $1}'`;
do 
    echo "looking at " $i;    
    echo "TABLE " $i >> differences.diff

    mysqldump --no-data $DB_MODIFIED $i > $i.definition.modified.sql; 
    mysqldump --no-data $DB_ORIGINAL $i > $i.definition.original.sql; 
    diff $i.definition.original.sql $i.definition.modified.sql >> differences.diff;
    
    #clean up
    rm $i.definition.original.sql 
    rm $i.definition.modified.sql
    # skip data
    if [[ ${SKIP_TABLES[$i]} = 1 ]]
    then
        echo "skipping data for " $i;
        continue;
    fi

    mysqldump --skip-extended-insert --no-create-info $DB_MODIFIED $i > $i.data.modified.sql; 
    mysqldump --skip-extended-insert --no-create-info $DB_ORIGINAL $i > $i.data.original.sql; 

    diff $i.data.original.sql $i.data.modified.sql | grep -v ' Host:'  >> differences.diff;
    
    # clean up
    rm $i.data.original.sql;
    rm $i.data.modified.sql;
done

testing Text::Clevery

2011-11-13T21:47:00.001+02:00

... so far, so good

wondering if there are other Smarty parser in the wild

some command line shortcuts

2011-11-03T22:55:00.011+02:00

dig for missed dependencies

egrep -r '^\s*(use|with|extends)\s+' * \
 | awk '{print $2}' | awk '{sub(/\;/, ""); print}' \
 | sort -u

jumpstart a "00-load.t"

egrep -r '^\s*package ' * \
 | awk '{ print $2}'  \
 | awk '{sub(/\;/, ""); print "use_ok(\""$1"\")"}' \
 | sort -u

quick syntax checks

for i in `find ./  | egrep '\.(pm|pl|cgi|t)$'`; do perl -wc $i; done

bad tests

2011-10-24T14:29:00.004+03:00

bad tests are worse than no tests, and it's better to have no tests than to have tests designed to pass

there it is, I've said it publicly

sorting utf8 with Perl

2011-09-14T13:59:00.000+03:00

What's Wrong with sort and How to Fix It

informative perl utf8 article

2011-09-14T12:42:00.001+03:00

use unicode by default in Perl

how to do utf-8 with HTML::Mason

2011-08-06T00:22:00.002+03:00

how to do utf-8 with HTML::Mason

4 easy steps to make Mason utf-8 Unicode clean with Apache, mod_perl, and DBI

PerlSetVar MasonPreamble "use utf8;" fixed it for me

yes, I know Mason 2 is live :)

fuck iso-8859-16

2011-07-15T02:45:00.002+03:00

... and the ignorant bastards that cannot distinguish between an aesthetic choice when designing font faces and character encoding.

iso-8859-16 is not for Romanian, and the traditional and correct way to write in Romanian is with s+cedilla, t+cedilla ... which means iso-8859-2. Want ş with a tiny comma below the "s" ? Fucking design a new font face and give the rest of us a break. Rumanians wrote with s+cedilla for 150 years and did not lose sleep over it, but now we have to deal with two fucking character sets that don't translate and which look so much alike on screen that you need fucking glasses to distinguish between them.

Yes, you know who you are.

Net::SSH::Perl install

2011-07-12T17:15:00.002+03:00

Very nice module but a bit of a pain to install: the pain will be caused by Math::Pari wich fails automatic installation with cpan(m).

To fix it get the Math::Pari archive from CPAN, then ftp://megrez.math.u-bordeaux.fr/pub/pari/unix/OLD/pari-2.1.7.tgz and extract pari-2.1.7.tgz in the same folder where the Math-Pari-2.* folder is, then perl Makefile.PL Configure, make, make install. pari-2.5.0 won't work.

perl code deployment on shared hosting

2011-03-24T08:46:00.000+02:00

on perl code deployment on shared hosting

getting started with Perl and BASH

2011-03-15T11:23:00.002+02:00

... was asked for pointers in a private mail; here is my quick and dirty answer:

I am no shell expert (have read and modified lots of scripts, but have not written any advanced stuff from scratch); for BASH I recommend http://tldp.org/HOWTO/Bash-Prog-Intro-HOWTO.html;

for Perl: the documentation packaged as man pages is very good (do a "man perl" to get the index), and in book form almost anything from O'Reilly is good (they are the unofficial sponsors of Perl ;) ) , this one is a must http://oreilly.com/catalog/9780596000271 , this one is a good place to start http://oreilly.com/catalog/9780596520113 , and http://www.onyxneon.com/books/modern_perl/index.html as soon as you're done with "Learning Perl" .

Anything signed by Randal Schwartz, Damian Conway, chromatic, Larry Wall, Tom Christiansen (and others, but those are the names I remember now) is worth reading, whether in print or on the web.

you can find books online here http://www.perl.org/learn.html , and some other links here http://oreilly.com/perl/

If you're on windows get Active Perl http://www.activestate.com/activeperl or Strawberry Perl http://strawberryperl.com/ , both are very good. Indigo Perl used to be in vogue years ago, might not be a bad choice http://www.indigostar.com/indigoperl.php ... a long time ago I used to use Indigo Perl when exiled in M.S. land, so I am a bit partial to them :), but Strawberry Perl appears to be the "official" Windows port of Perl.

If you know Unix/Linux reasonably well, you could try http://www.cygwin.com/ : you get a lot of unix tools running on windows including perl, or install Linux in a virtual machine (I recommend VirtualBox, but Parallels and VMWare make good tools, too).

interview with Damian Conway

2011-02-25T12:49:00.001+02:00

might be old, I just found it

interview with Damian Conway

Mojolicious 1.0

2010-12-29T12:38:00.001+02:00

finally with documentation :)

http://mojolicio.us/

memory leaks

2010-12-26T19:58:00.001+02:00

been hunting Catalyst/Template Toolkit memory leaks for a few days ...

unicode again

2010-10-07T20:52:00.001+03:00

nice explanation

http://www.ahinea.com/en/tech/perl-unicode-struggle.html

Padre again

2010-08-24T23:05:00.005+03:00

Been using it for a week every day, and it feels great.

There are a few issues, but it has crashed on me only once, syntax highlighting in css is great, knows when it's editing html without filename hints and does not balk at Mason ... besides having perltidy and perlcritic only one keyboard shortcut away.

Catalyst unicode-related error

2010-08-16T21:52:00.003+03:00

Caught exception in engine "Wide character in syswrite at
 /usr/local/share/perl/5.8.8/Catalyst/Engine.pm line 675."

the solution is to add the Unicode plugin to the application:


use Catalyst qw/
# ... some other plugins
  Unicode
  /;

"Desperate Perl and Large Scale Perl"

2010-07-26T22:16:00.002+03:00

a tale of two languages: desperate Perl and large scale Perl

not enough links to this screencast :-P

2010-07-26T21:48:00.000+03:00

Perl Myths 2010-03 Ann Arbor Perl Mongers

"amusing" @INC trouble

2010-07-01T12:22:00.002+03:00

got into a bit of panic today:


$ perl -e 'use LWP'
Global symbol "%Config" requires explicit package name at ...

This happened when running anything that used perl, and this error I have not seen before.

Asking around and googling resulted in one conclusion: core Config.pm is busted, Perl is damaged, must reinstall. Lucky for me I realized there was a Config.pm in the folder I was in at the time, and './' was in pole position in my @INC.

tried Ubuntu 10.04

2010-06-11T10:04:00.002+03:00

... not bad, though I won't switch: shivering at the thought of moving all the configs, instant messenger logs, mail accounts and setting and other useful junk accumulated during 5 years of updating a Debian in-place.

perl on wikibooks

2010-06-02T11:41:00.000+03:00

Perl book on wikibooks

nice one, did not know about it

strange apache error

2010-05-27T15:46:00.002+03:00

apache failed to start with an error:


[Thu May 27 13:41:29 2010] [crit] (28)No space left on device: mod_rewrite: could not create rewrite_log_lock
Configuration Failed

Carlos Rivero has a solution using perl.

Here is what I used, as somebody suggested in the comments to the post linked above:


ipcs -s | grep emilper | awk ' { print $2 } ' | xargs ipcrm sem

then restarted the server

installing padre 0.62 on debian sid

2010-05-24T14:13:00.005+03:00

Got an unexpected "bank holiday", so I am playing.

Padre installs nicely on Debian Sid when using CPAN on 2010-05-24 ... I wonder if I am doing something wrong :P ... or maybe Padre is getting closer to being ready for prime-time.

Plugins that installed cleanly:

Padre::Plugin::PAR

Padre::Plugin::XS

Padre::Plugin::Encrypt

Acme::Padre::PlayCode

Padre::Plugin::Nopaste

Padre::Plugin::DataWalker

Padre::Plugin::HTML

Padre::Plugin::HTMLExport

Padre::Plugin::JavaScript

Padre::Plugin::Kate

Padre::Plugin::CSS

Padre::Plugin::Ecliptic

Padre::Plugin::SpellCheck

Padre::Plugin::PHP

Padre::Plugin::ViewInBrowser

Padre::Plugin::PerlCritic

Padre::Plugin::Debugger

Padre::Plugin::Encode

Padre::Plugin::Alarm

Padre::Plugin::Swarm

Padre::Plugin::ClassSniff

Plugins that fail tests:

Padre::Plugin::Shell::Base

Padre::Plugin::CommandLine

Padre::Plugin::InstallPARDist

Padre::Plugin::Vi

Padre::Plugin::Template

Padre::Plugin::Filter

perl editors

2010-05-24T10:44:00.002+03:00

Eclipse

When using subclipse with Aptana Studio or Eclipse, it will complain about missing "libsvnjavahl-1". On Debian, the package libsvn-java needs to be installed. After that /usr/share/java/svn-javahl.jar needs to be added as "external jar" to the JRE in use.

EPIC should install with no trouble.

In Eclipse I am giving a try to QuantumDB Eclipse Plugin for database browsing: it can be installed easily from the Aptana plugin manager.

Eclipse is a pain when working on remote files (even with ssfs) if the link is not lightning fast, so other editors I use are Komodo Edit (kind of an unofficial standard at work), Emacs (at home, always come back to it after exploring alternatives), or Kate.

daemontools install on centos 5.4

2010-05-17T10:48:00.002+03:00

after getting this error:

/usr/bin/ld: errno: TLS definition in /lib/libc.so.6 section .tbss mismatches non-TLS reference in envdir.o

need to edit

src/conf-cc

add this:


-include /usr/include/errno.h

... via Installing Bernstein’s daemontools on CentOS 5

tested CPAN last night

2010-02-25T17:37:00.002+02:00

If you want to test CPAN works as advertised, install MojoMojo. I did it last night, and it worked, even though there are hundreds (or so it seems :) ) dependencies. I installed it in my home, so this evening I'll delete the ~/perl5 directory and start again, this time taking notes: once a few of those dependencies, some from CPAN and some *-dev packages, are installed, the whole "cpan MojoMojo" should work nicely, and you'll have only to watch from time to time for prompts from tests.

Gtk2-perl docs

2010-02-09T07:10:00.004+02:00

main site

gtk2-perl demo, recommend

presentation, pdf file, go here for source code

muppet's site, with a bunch of useful stuff

Gtk2-perl recipes

use Devel::NYTProf with CGI and apache

2010-01-28T15:50:00.002+02:00

put these lines in the apache virtualhost config file:


SetEnv NYTPROF "file=/tmp/nytprof.out"
SetEnv PERL5OPT "-d:NYTProf"

the output from Devel::NYTProf will go to /tmp/nytprof.out

undefined symbol: apreq_handle_apache2

2010-01-23T16:49:00.005+02:00

solution from here

symbol lookup error: /usr/lib/perl5/auto/APR/Request/Apache2/Apache2.so: undefined symbol: apreq_handle_apache2

Make sure you load the apreq module in your apache2 config file:

LoadModule apreq_module /usr/lib/apache2/modules/mod_apreq2.so

In debian there is a file in /etc/apache2/mods-available called
apreq.load. Just link that to your
/etc/apache2/mods-enabled directory and restart apache2

HTML::Mason and apache issues

2010-01-23T15:18:00.002+02:00

this was copied from somebody else, but I did not keep the reference ...

===
while atempting to upload files in html::mason the response is only "End of file found":

solution: add this to the vhost configuration file

PerlSetVar MasonArgsMethod mod_perl

explanation: it seems HTML::Mason is using CGI to deal with post/get by default, and when CGI is done with post/get, there is nothing left for apreq to work with; the line tells HTML::Mason to use mod_perl to deal with the args

I18N for HTML::Mason

2010-01-23T13:48:00.004+02:00

in the folder where I keep the virtual host data:

locale/ro/LC_MESSAGES/ is where the messages.mo file goes


drwxrwxr-x 1 emilper www-data 4096 2010-01-22 04:53 apache_conf
drwxrwxr-x 1 emilper www-data 4096 2010-01-22 05:48 htdocs
drwxrwxr-x 1 emilper www-data 4096 2008-07-16 18:37 lib
drwxr-xr-x 1 emilper emilper  4096 2010-01-22 02:33 locale
drwxrwxr-x 1 emilper www-data 4096 2008-07-19 02:13 log
drwxrwxr-x 1 emilper www-data 4096 2010-01-22 04:54 mason

in my Apache VirtualHost config:


PerlSetVar  GettextLocalesDir /home/www/mysite/locale/
PerlSetVar  GettextDomain messages

in my Mason request class (MasonRequestClass set in apache conf) :


$self->{'i18n'} = Locale::gettext->domain($self->apache_req->dir_config('GettextDomain'));
$self->{'i18n'}->dir($self->apache_req->dir_config('GettextLocalesDir'));

and


sub _t {
  my ($self, $string) =  @_;
  return $self->{'i18n'}->get($string);
}

in my code:


$page_title = $m->_t("String to translate");

As of now, I have difficulties persuading xgettext to extract the strings from autohandlers, .mas files, and .html files. Until I write my own extractor, I add the strings by hand to a "messages.po" file, which I compile to a "messages.mo" file that can be used by gettext with:


msgfmt messages.po