Free Resources for Windows Perl Development 117
jamie pointed out an important announcement in the Perl community. Adam Kennedy, known as Alias, developed Strawberry Perl to "make Win32 a truly first class citizen of the Perl platform world." Over the last year, major CPAN modules have used Strawberry Perl to get to releases that work trouble-free on Windows. But the tens of thousands of smaller modules on CPAN are lagging, in many cases because of lack of access to a Windows environment for development and testing. Now Alias has worked with Microsoft's Open Source Software Lab to provide for every CPAN author free access to a centrally-hosted virtual machine environment containing every major version of Windows. "More information (and press releases) will follow, the entire program under which this partnership will be run is so new it's only just been given a name, so some of the organisational details will ironed out as we go. But for now, to all the CPAN authors, all I have to add is... Merry Christmas. P.S. Or your appropriate equivalent religious or non-religious event, if any, occurring during the month of December, etc., etc."
Re: (Score:3, Interesting)
If one is forced to use an Alias for online life, might as well make it obvious...
More like lack of interest. (Score:1, Redundant)
"But the tens of thousands of smaller modules on CPAN are lagging, in many cases because of lack of access to a Windows environment for development and testing."
I was born on a day, but not yesterday. I must admit to not using perl for anything serious in a very long time, but as I recall, many smaller modules in CPAN didn't even work trouble free on an up-to-date linux machine because they were either badly coded or simply didn't run with newer versions of other dependencies. Maybe things have changed, but
Re: (Score:2)
Actually, as far as linux, usually everything just works. I'm sure there's a lot of one off modules that never get anything but a first release and sit unloved but almost everything I've needed to install runs without issue and has done so for years. The only module I can think of on Linux that was a PITA was the SDL libraries, due to the different versions and many library dependencies.
As far as Windows, it depends, all the core stuff is fine, and if you can find it in a ppm repository, you're usually fin
Re: (Score:1)
Finally, I can make sure my module works on Windows 3.11!
Just last year, I tracked down a machine with Windows 95 and got it working there, but I really wanted to make sure it worked with ALL major Windows operating systems.
Re: (Score:3, Informative)
The CPAN testers [cpan.org] was conceived back in May 1998 by Graham Barr and Chris Nandor as a way to provide multi-platform testing for modules. Today there are 2,653,748 tester reports and more than 400 testers giving valuable feedback for users and authors alike.
CPAN modules are probably better tested cross-platform than libraries for any other language.
Re:More like lack of interest. (Score:5, Informative)
The problem with CPAN Testers is that while it can tell you IF your module is broken, it doesn't give you any way to actually get onto Windows to debug the problem. All you can does is guess the fix and upload a new release, and hope for the best.
CPAN Testers is the canary in the coal mine, which is handy, but doesn't actually help clear out the poison.
Re:More like lack of interest. (Score:4, Interesting)
Having a lot of testers doesn't seem to actually affect the quality of many of the smaller modules. Many of them have absolutely insane dependency chains, requiring both untested, unreliable modules, often from the same author, and completely deprecated modules for the same component, with massive duplication of modules to do the same small task in slightly different ways.
For the core modules, and those exciting modules likely to be included in the next release, I can see the results of testing work. But many of the smaller ones are one-off debris by sloppy programmers that unfortunately show up in the CPAN search engines. No one seems to test them, and they're apparently not tested again after their original publicaton for compatibility with new perl releases.
Re: (Score:2)
"...lack of access to a Windows environment for development and testing."
Reminds me of the Monty Python Cheese Shop sketch.
"Do you have any Windows machines?"
"We don't get much call for it around here."
"It's the single most common operating system on the planet!"
"Well not round here!"
I can accept many, many arguments as to why the Perl code hasn't been tested on Windows, but not that the team couldn't find a Windows machine. Surely at least one of them has at least one friend who at least has a dual boot
Re: (Score:2)
Why is ruby better? For one, it's object orientation isn't an after-thought, and it's actually useful. For another, it makes it easy to produce readable code (in perl, readable code requires hard work.) Jus
Re: (Score:1)
nice link there (Score:1)
http://tech.slashdot.org/Strawberry [slashdot.org]
Re: (Score:2)
This really works quite nice... (Score:1)
It includes a pre-configured CPAN-shell and compilation tools for c modules.
Installation is a bit screwy on Vista; it does require some manuevering there, but it works off to bat on xp.
I, for one, am quite happy that there are other options that function well besies ActivePerl; never liked it. No longer is one dependant on the proprietary ppm packaging so much! The value of CPAN is better put to use here.
Though, businesses that rely on perl might want to wait before abandoning activeperl; strawberry is rela
Re: (Score:2)
True, but after a morning of trying to work through ActivePerl PPMs, giving up, and having Strawberry Perl "just work", atleast my company is getting closer and closer to that tipping point ...
Netcraft confirms it (Score:1, Informative)
Re: (Score:3, Interesting)
That's not netcraft. Its the TIOBE index, which is notorious for being completely inaccurate. Google for "tiobe flawed" and you'll start to see just how worthless it is. That said, perl does have some serious challenges ahead if it wants to stay as popular or gain popularity.
Re: (Score:3, Informative)
it might help to look at the context of the ranking in order to see things in perspective. coming in 8th out of 50 [tiobe.com] is hardly "dying." and if Perl is dying then Python (7th) isn't far behind, and JavaScript (9th), Ruby (11th), Lisp/Scheme (19th), and Lua [wikipedia.org] (20th) are all definitely dead. not to mention the rest of the ranked languages:
21. ActionScript
22. MATLAB
23. FoxPro/xBase
24. Fortran
25. Ada
26. Transact-SQL
27. Prolog
28. RPG. (OS/400)
29. Erlang
30. NXT-G
31. Awk
32.
Re: (Score:2)
You left out my personal favourite: Logo is at rank 18, up from 21 last year, making it more popular than Lisp/Scheme, Lua, and Caml.
Re: (Score:1)
I'll give you one better. Apparently, Google Chrome is written in Delphi. TIOBE confirms it.
Re: (Score:2)
The current position is irrelevant if you want to determine whether something is "dying" or not; what matters is the long-term trend.
Re: (Score:2)
then i suppose the top 3 languages (2 of which have actually grown in the short-term) are dying since they've all decreased in popularity in the long-term?
Re: (Score:2)
You mean Java, C, and C++? I wouldn't call it dying (yet), but they're certainly on the decline overall - C/C++ is being pushed out of desktop app market, for example, replaced by front-ends in Python (on Linux) or .NET (on Windows). Of course both are going pretty strong in low-level, and will probably remain strong there for a long time to come
Java is just slowly declining overall. Sun's not feeling well, and IBM and Google are too conservative and not much interested in developing it further (look at the
Re: (Score:2)
You are correct. It is. [cpan.org]
Linky (Score:5, Informative)
Maybe this link will be a little more useful: http://strawberryperl.com/ [strawberryperl.com]
Perl::Windows (Score:1, Redundant)
Tag: "letsallkillthisbeast" (Score:2)
Both of them?
Re: (Score:2)
Microsoft and Windows? Good luck with that.
It seems I'm the first to say it but... (Score:2)
Adam, you're the freakin' MAN.
I actually didn't know the breadth of what you were arranging but this is just way beyond what I expected!
I will definitely be taking advantage of this... Next time you're in the Boston area, I want to personally buy you a case of your choice of beer.
(yes, I understand fully that this posting means it's "on the record" :D )
Judgement Day (Score:1)
WTF? (Score:1, Redundant)
Hello? Virtualization?
Re: (Score:1, Insightful)
"...lack of access to a Windows environment..."
Hello? Virtualization?
Hello, obtaining a legal copy of windows?
Re: (Score:1, Insightful)
"...lack of access to a Windows environment..."
Hello? Virtualization?
Hello? They would have to pay if the wanted to do it legally?
Re:WTF? (Score:4, Insightful)
On top of licensing issues, other accumulated comments included the fact many Unix greybeards have never used Windows before, so the accumulated time to find the right torrent, download it, work out how to install everything etc etc was something they greatly didn't look forward to doing.
It's not that they couldn't, it's just that they are busy people, like everyone else, and the time investment was too big for the relatively small win of closing one or two bugs on Windows.
Shortcutting that process by just letting them log directly into a running instance is considered a significant improvement for that group.
Re: (Score:1)
I'm one of that group. And yes, my beard *is* going grey. I've not used Windows for over a decade. I have no idea how to set up and configure a current Windows to be secure and to have a reasonable development environment. Nor do I care to learn, as I have better things to do with my time, like making tasty booze and grumbling about The Youth Of Today with their ghetto blasters and hard core pornography.
But if this lets me test my code on Windows before releasing it, and spot and fix stupid errors, then
Free Resources for Windows Perl Development (Score:1, Flamebait)
>> Free Resources for Windows Perl Development
The definitive one: http://python.org/ [python.org]
Re: (Score:2)
Hey! Python can run Perl? That's awesome. Do you have a link for that? I'd love to be able to use my favourite language on what is by all accounts a great interpreter.
(If you want to run Python on Perl instead, you can [cpan.org]. Perl: the choice is yours.)
Re: (Score:2)
>> If you want to run Python on Perl instead, you can
That you can run Python from other languages (not only Perl) talks very good about the Python's modularity and adaptability. BTW, I understand that Perl programmers logically would want to call and use Python, so this module has a lot of sense.
>> Python can run Perl?
From your interesting link:
"perl is a Python module bundled with Inline::Python that gives you access to Perl from inside your Python code."
BTW, I don't see the need to call Perl f
Every major version of Windows? (Score:2)
I see in the list Windows XP and Server 2003 (they're almost the same...), Vista and Server 2008 (they are exactly the same!). Where's Windows 2000? there's quite a large chunk of users left on that platform. Hell, even a few users on Win98 or NT4 still.
Re:Every major version of Windows? (Score:5, Informative)
If you'd Read My Fucking Article you'd see that the current set is the MINIMUM initial launch set.
The plan once we are running is to start adding more variations as needed. I'd certainly like to have a 2000 instance.
As for the Windows 95 family, as I understand it support was dropped from the current Perl core for anything older than 2000.
Re: (Score:2)
OK, maybe that was overly cruel...
But yes, I'm aware the launch set only covers well-supported versions, and yes, there's more variations coming later.
Damn Dude. (Score:2)
Do they make you hang out with Ballmer for a week just to get access?
What about ActivePerl? (Score:5, Interesting)
I've used ActiveState's ActivePerl on several windows boxes over the years and have had "trouble-free" experiences with it. Granted, some of the more bleeding-edge modules weren't at the latest revs but the mainstream software I used didn't strictly require those either.
Re:What about ActivePerl? (Score:5, Interesting)
My issues with ActivePerl have been that it is fundamentally different to all the other Perl platforms (you don't get the full CPAN, just binary packages) and that because one company is the central gatekeeper of all the binary packages, there was never a reasonable way for CPAN authors to debug their modules.
I for one wrote 150+ modules, of which a grand total of 7 were available on ActivePerl, due to various bugs in the ActivePerl build farm that went unfixed for years.
To be truly first-class, you should be the same as the other platforms, not similar-but-different.
Re: (Score:1)
I for one wrote 150+ modules, of which a grand total of 7 were available on ActivePerl
Perhaps the problem is with the way you code your modules and not with the ActivePerl build farm? ActivePerl has like a brazillion modules available for it.
Re:What about ActivePerl? (Score:5, Informative)
Not really. ActiveState made bizarre PPM binary compatibility decisions, which meant that their version of Scalar::Util [cpan.org] didn't include the XS components. As the 5.8.x series continued, more and more modules relied on that XS component, which meant that increasing amounts of the CPAN weren't available as PPMs. I don't know the exact figures, but it wouldn't surprise me if one-third to one-half of the CPAN were unavailable from AS's repositories.
(Did you know Alias is one of the CPAN administrators?)
Re: (Score:2)
What makes ActiveState Perl different from Perl as seen on Linux et al? Why cant you just run the same Perl code from Linux on ActiveState? And if it sucks so much, why has no-one done a more straight port from the Linux code (or is that what Cygwin Perl is for?)
Re: (Score:1)
In core Perl itself, nothing. ActiveState uses the same source code. The problem is ActiveState's PPM repository.
It's the same source code.
ActivePerl and Strawberry Perl are distributions of Perl. They don't change the source code. The diffe
Re: (Score:2)
Perl modules may include C code and the module installers assume the availability of a compiler. Since that is not a given under Windows, ActiveState makes binary versions available in their PPM repository, but they support only a subset of modules for a number of reasons and are often behind.
Bundling the C compiler (mingw gcc) with Perl is what Strawberry Perl is all about so you can use the standard build process for CPAN modules.
Re: (Score:2)
> Perhaps the problem is with the way you code your modules
At the lowest point, just after the release of ActivePerl 5.10.0 the were ZERO modules available starting with the letter S...
To their credit, ActiveState did finally recognise the old PPM build farm was a piece of shit, and they've now rewritten it from scratch and junked the old version entirely. The situation is nowhere near as bad as it used to be.
Re: (Score:2)
That is all nice, but wait until you get a module that *isn't* in PPM. I can name several, especially the little guys written by the parent (Image::Math::Constrain or Image::Delivery for example). I use FreeBSD and it is easy to get such CPAN modules into the ports tree (and actually, most of the ones I've added are for Adam's stuff). Lord knows how you'd get these things added to Active Perl.
Of course, the real "exciting" part of Active Perl is getting anything related to mod_perl* working.
PS: The use
Re: (Score:2)
Hey Adam how mature is strawberry perl compared to activePerl? Can I run it yet on my vista laptop or is it still beta?
Re: (Score:2)
Vista should work fine now.
Only caveat is that there is currently no 64-bit port.
At long last (Score:1, Troll)
People that hate perl and people that hate windows can *finally* find some common ground!
Honestly, even a shop teacher can count the number of users that care about this on one hand.
Re: (Score:1, Funny)
Honestly, even a shop teacher can count the number of users that care about this on their one remaining hand.
Re: (Score:1)
Sign me up! (Score:4, Interesting)
I maintain a few modules on CPAN. Nothing big, I'm the sole author.
In August, I got email from someone complaining that one of these modules doesn't pass its self-tests. After some back and forth, it turns out that it passes on Linux and fails on Windows. They even submitted a patch, but I don't want to integrate it unless I can test it on Windows.
I've got some Windows machines in my house, but I'd have to put together a usable development environment, and it's a hassle, and I've got a day job, and it just hasn't happened in 4 months.
If Alias et. al. can get me access to a Windows environment, this module could get cleaned up a lot sooner.
what could possibly go wrong (Score:1)
Re: (Score:2)
A decade ago Perl was a 1st class citizen on Win32 (Score:3, Informative)
Perl has been a first class citizen on Win32 starting with the GSAR port back in late 90s, then Perl for Win32 and now ActivePerl.
In fact ActivePerl was more up to date than unix Perl during the late 5.005 and 5.6 because the pumpkin was primarily a Win32 developer.
If you want to find the second class citizens in the Perl world look at OS2, Aix, Hpux, and other strange unixes. I know you want to make Perl better and are working hard on it, but insulting the people who put together the foundation you're now working on is misguided. They did a damn fine job.
Re: (Score:2)
Perl could be a first class citizen on Windows if MS bundled a C compiler and library like any decent OS does.
Re: (Score:2)
Or, you could go and download [microsoft.com] them, geeze! Why bloat installations with bundled stuff that the vast majority of users won't ever need?
Re: (Score:1)
Re: (Score:2)
Your post makes no sense. If Microsoft bundles other things they should also bundle gigabytes of C++? This doesn't follow in any logical way, would make no business sense and would inconvenience most users. Can you even name something comparable that Microsoft bundles with Windows? Most stuff they ship with Windows is very minimal, just enough to get you started; I don't believe any of the componen
Re: (Score:1)
No C compiler requires gigabytes of disk space. C++ isn't used to build Perl.
Explain how this would inconvenience anyone.
The .NET Runtime. It comes with compiler executables for C#, VB.NET and JScript.NET, and it is installed by default since Windows XP S
Re: (Score:1)
Re: (Score:2)
Spoken like a developer
Re: (Score:2)
And how is the build process for a CPAN module going to automate this if VC isn't installed? ..."
"Dear user, please go here:http:..., then click this and that, wait for the download, then click setup.exe, wait for the install to finish, then come back here and retry. Oh, and mare sure you use only browser X or Y because Microsoft won't support anything else, and make sure you turn on Javascript, and cookies,
Re: (Score:2)
Not a very good question: first, on Windows, most stuff is distributed as binaries. There are lots of tools and facilities for creating installers; compiling source is not a requirement. Anybody who plans to distribute his product in source format to normal Windows users, expecting them to build and install it, has already failed.
Second, if somebody needs to install perl modules, they belong to the "knowledgeab
Next... (Score:1)
Promote Microsoft and undermine Linux (Score:1, Insightful)
Whatever other benefits there are, the main beneficiary of porting open source software projects to the Windows platform is Microsoft.
Bear in mind that if there were no Linux, many of the other open source software projects would not have existed.
Microsoft don't even have to embrace and extinguish when the open source projects are doing it for them.
Programming on Linux (Score:1)
Re: (Score:2)
Re: (Score:1, Redundant)
I know what Win32 is. I also know, having written my own abstraction layer for it, that it's a real POS. Thankfully, I haven't had to use it in over a decade, and have no plans to in the future, just like I don't see myself switching back to Windows either at work or at home.
KDE 4.1 makes Vista look like an old Buick.
Re: (Score:2)
Re: (Score:2)
Good point - the old "Windows is like a car with the hood welded shut" strikes again! Darned :-)
The frisbee cost is another good point. Open up the wrong email, or surf the wrong site, and you'll spend more to fix the damage than the OS cost in the first place - and that's even if you had
Re: (Score:1)
I recently updated to Ubuntu 8.10, and discovered that the sleek, fuel-efficient KDE programs I love were ripped out and replaced with hummers. Honestly, Kolourpaint takes up 50% of my cpu drawing a line.
And I do use 32 bit, because my single 2.8 Ghz processor has served me well for three and a half years, and I don't see any reason to upgrade it.
And just to be clear, that 2.8Ghz is more than most cores nowadays, so though I may be stuck in the past, I'd still say that for the most basic of drawing applicat
Re: (Score:2)
KDE 4.1 makes Vista look like an old Buick
In your eyes. In mine, they both look like the rear end of a dump truck. I choose my UI for how easy it is to work with, not how easy it is to look at.
Re:win32 a first-class citizen? (Score:4, Insightful)
Re: (Score:2)
I sincerely hope so. But until it kicks the bucket, there's software on it which needs to be supported, and improved Windows testing and debugging of current CPAN modules will make the lives of those of us who use those modules on Windows a whole lot easier.
Re: (Score:2)
Yes, it's impossible to find a 64-bit only CPU for use with Windows.
Fucking impossible.
Re:win32 a first-class citizen? (Score:4, Insightful)
YOU may consider Windows a second-class citizen, but the market sure doesn't. In an above post, you compared Microsoft to GM, and declared that they were dying. Yeah, sure, they're dying exactly like GM, except for the fact that they're making money hand over fist, have over $70 billion in assets, and haven't required a government bailout.
Re: (Score:2)
Windows Mobile is everywhere.
No
So is Windows Embdedded
No
Most (embedded / server) places that have linux are not advertised as such.
Re: (Score:1, Interesting)
I think someone needs to get their head out of their ass and face reality. Outside the desktop, Windows isn't visible? Really? Windows Mobile is everywhere. ...
Yes, and that someone is ... you! I'm guessing you didn't see the news Friday that Windows Mobile was just passed in market share by ... the iPhone!
http://letmegooglethatforyou.com/?q=iphone+q3+2008+market+share
The problem for fanbois who point to Microsoft's size as a sign of its invincibility is that ... size is not a sign of invincibility.
Re: (Score:2)
You're joking, right? Are we talking about the same Windows Mobile that got its doors blown off in the mobile market by OS X on the iPhone in a mere two quarters?
http://blogs.zdnet.com/microsoft/?p=1163
* * * * *
"Time flies like an arrow. Fruit flies like a banana." —Groucho Marx
Re: (Score:1)
Windows Mobile is not everywhere, Windows Mobile is on a bridge to nowhere!
The iPhone has more worldwide market after a mere 2 years, WinMo had over a decade headstart, since it was called WinCE or whatever, and is still irrelevant.
Blackberry dominates in NA, Symbian dominates globally, and Android is going to bite into WinMo's market share before anything else!
Re: (Score:1)
How much of the server growth is due to capacity, though? I'd like to see some numbers comparing how many users a Windows Server host for a web application can support versus, say, BSD, for the same application. I looked around a bit but couldn't find any.