use Perl

I've released, with the kind permission of Slando, Test::HTML::Form which makes testing HTML and HTML::Forms considerably easier.

The API is loosely based on Test::HTML::Content, but it use HTML::Treebuilder and re-uses parse trees and objects where possible, making it much quicker - it also uses filenames instead of passing strings around, and has additional methods for extracting strings and testing form elements easily

So far it's already made a lot of testing easier, but I already have a wishlist of extra features - like being able to extract HTML::Elements easily, provide urls instead of filenames, and allow extra tests on more form elements.

Anyway.. its in SVN at code.google.com and on CPAN

I tried posting another response to david at
http://journal.dedasys.com/2008/12/07/python-surpasses-perl but his blog doesn't cope with long replies. ..so I posted it here..

David,

I can't even post a response on your blog so I posted it here :

What those of us who actually have written and developed open source perl projects are saying is that sourceforge is a very poor indicator of anything but sourceforge usage in languages.

I think you're article would be much better if you turned it on it's head : Which languages use sourceforge, vs which languages have strong software repositories of their own.

Ruby has it's gems, php has pear, Python has the caverns of thingywotsit.

CPAN just happens to be far larger and better equipped than the alternatives.

As a free software author it's simply not worth the effort of using sourceforge for any of my projects - CPAN provides better tools and distribution, I have my own homepages for my projects.

I would probably even surmise that Sourceforge is actually the thing that is losing market share : you have better and more focussed alternatives like ohloh, github and google code, most significant projects will host their own version control, bug tracking and wikis.

"f CPAN shows some growth for Perl, that seems normal to me - if there's no growth there, the language really would be dead."

16% is more than just healthy growth - it indicates exactly the opposite of your conclusions!

"Freshmeat indicates that, relative to Python and other languages, Perl isn't seeing as much code released."

I'd turn that on it's head : the usage of freshmeat indicates the lack of better alternatives for other languages.

CPAN and PEAR have both moved forwards a great deal since sourceforge was created, yet sourceforge is pretty much identical to how it was in the 90s.

"The point that perhaps people are less interested in Freshmeat these days isn't a bad one (although "myself as a case study" isn't great statistics either), but why should that be different for Python, Perl or, say, Lua users? If it's going to drop off, it'll drop off for everyone."

Because other languages don't have a decent repository of their own and/or have different eco-systems.

Despite the existance of PEAR, PHP is still dominant on freshmeat, but PEAR and Freshmeat don't overlap much - one being for applications one being for libraries.

"Python, however, is growing, whereas Perl hasn't. I'm sorry if that upsets you, but those are the relevant facts."

No - all you've shown is that freshmeat is more popular with python coders than perl ones.

"However, it is not as popular as it once was, in terms of new code being written."

Again, you can't sensibly draw this conclusion when you know that CPAN growth is 16% year on year.

You probably don't even have figures to compare growth in the vaults..

A quick check of the web page shows "Parnasus Totals: 2025 items in 49 categories." CPAN yearly growth is larger than the entire Panasus repo!

Maybe Python has a new repo - Parnasus looks like it hasn't been updated or worked on since the 90s - that would certainly explain why pythonista's tend to rely on sourceforge.

Yup, there is a new repo : 5393 packages on cheeseshop.python.org

So python has 5393 packages on cheeseshop, 3200 on sourceforge, and 2000 on parnasus.

CPAN has 14793 packages on CPAN and 3800 on freshmeat.

CPAN growth is 16%, Perl growth on sourceforge is 1.1%, Python growth is 1.4%.

Looks to me like python is being held back by the lack of a CPAN.

I came accross two blog posts today on enjoying your open source / personal projects and it's a helpful reminder that I don't need to stress about the projects I have.

* http://nickelcode.com/2008/12/08/developing-for-fun-some-observations/
* http://weblog.masukomi.org/2008/12/14/on-coding-for-fun

both very good, with reminders of the fact that if you're not getting paid or some other reward and it's not fun, then it's probably not worth doing.

It's also worth remembering the flip side of the coin - nobody owes you support or bug fixes in a project they're doing for free on their own time.

And finally, I got what you always ask for as a writer or maintainer of open source (or heck even commerical) code :

an unsolicited patch with test and explaination.. somebody read the README, and did exactly as I wishfull thinkingly asked and then seemed surprised that I was surprised.

in the words of Patrick Nice[1].. "which was nice"

1] http://www.bbc.co.uk/comedy/fastshow/characters/patrick_nice.shtml

What's the betting all the people who turned up just to say that perl is dying won't be back to post anything at all tomorrow or the day after or ever again.

What's the odds that most of them never even bothered to join a perl monger group or attend a talk on perl "back when they used perl" before they discovered that python was all rainbows and unicorn. ?

Fairly low.

How insecure do you have to be to join use.perl.org just say how much you prefer python ?

lame, lame, lame.

I've been meaning to play with Padre (An Editor/IDE written in Perl) as it has perl extensions and therefore I can do with it, what lisp gurus can do with emacs, without the use of mind-altering drugs or lambda calculus.

This evening, in the space of 90 minutes I was able to add some very basic emacs keybindings, only a handful, and some of them work, while some still have their original bindings, despite using "$self->CmdKeyClear(ord('w'),Wx::wxMOD_ALT());" which should unbind them. Bah.

Anyway, that's enough for now, but it has encouraged me to hack on it some more when I get the chance and to start integrating autodia based tools for generating and creating documentation and code at the click of a mouse.

You can find the results of my efforts at : http://www.aarontrevena.co.uk/opensource/Emacs.pm

For the last few weeks, while my better half has been away on training for her new job, I've been doing a bit of housekeeping on Maypole.

This week I've been adding a DBIx::Class model, progress was pretty slow and painful initially, but now it's going much better and I end up "in the zone" only to realise that it's almost 1am and I will probably have to get up at 6am (again, done this 3 nights in a row, once reading a book, twice hacking code)

Headache, Gritty, burning eyes, but it's worth it - I just managed to get the list action working (apart from autocreating embedded forms and buttons) with the new DBIx::Class model.. soon I'll be able to use Maypole with DBIx::Class for both work and personal projects.

Test::FormValidator is wonderful, it works really rather nicely.

I needed to test some currency and price validation functions for a client's website, and it was both simple and painless.

With only a little effort, I think I can test all validation on the front end of the site now by checking each validation function, and then each profile with sets of known good and bad data.

This makes it really easy to add test cases for validation errors found by QA and User Acceptance Testing.

perlbuzz finally completes the transformation from an interesting blog about cool stuff in perl to andy lesters personal blog with plugs for his other blogs and random shit his friends IM'd him.

[edit]
Meh. Shit happens. Bloggers are human. Even heisenbugs can be fixed and overtired toddlers with colds can go to sleep.

The Offical Perl 5 Wiki at the perl foundation now has 899 pages, many of them are very good, and there is a steady if slow stream of improvements every week.

A while a go I cheerleaded it a bit and said we could reach 1000 pages - since then we've had about 115 pages added, and about 10 times as many updates - which is pretty good.

But there is still a lot of room for improvement - see the Most Wanted pages, or look at pages you're interested and see if there are missing links, content, mistakes or pages you could add

Just uploaded 2.08 to CPAN - now skip lists work correctly for most input handlers, Java has been entirely removed (it hasn't worked for most of this decade), and author tests are excluded - also graphviz options now include concentrate for tidier output.