I'm writing a programmer's text editor (yes another one) in Perl called Kephra, which is also a CPAN module of course and bundled with Module::Install. Recently I saw that Module::Build has gone into core, so if I switch I could reduce dependencies. Is there any other reason to switch?
We use Module::Build in our group.
The main reason is Easy Extensibility.
Module::Build allows you to do more with your build process in pure Perl through subclassing. If you want to do more using Module::Install, you have to have knowledge of how Makefiles work, AFAIK. Since you presumably already know Perl, this can be an advantage.
As you said, using Module::Build removes the dependency on an external make program, which can be viewed as a good thing.
However, the main cons that I can think of are:
Although Module::Build has hit core, not everyone will be using an up-to-date version of Perl. For users with older versions of the core, you will be creating a new dependency.
Lots of veterans (not necessarily Perl people) are used to the perl Makemaker.PL; make; make install paradigm, and can be thrown off by having Build.PL instead. Hopefully this isn't a big deal.
Module::Build has occasionally broken our builds when its functionality has changed because the documentation didn't cover an edge case which we were using. The edge case was then changed and documented, but we had to re-code our subclass to get our build to work again (this happened for us at the recent upgrade from 0.2808 to 0.3).
All that said, though, I still recommend Module::Build simply for the extensibility. If that's not an issue for you, you may be better off sticking with Module::Install.
The cud as already been chewed a bit on this before in "Which framework should I use to write modules?"
After spitting out the cud I decided to go with Module::Build but clearly different answers are possible! (though I've been happy with M::B so far).
Well, Module::Build is a pretty good module, it's supposed to be a drop in replacement for ExtUtils::MakeMaker, that is, replace the Makefile.PL by a Build.PL, which generate a Build instead of a Makefile. It was also meant as "simple things should stay simple, hard things should be possible".
Module::Install takes a different approach and generates a Makefile.
Also, don't forget that not everyone runs the latest version of everything :-)
I don't remember any comparison of those modules, but I think you could find a few things from Module::Build and Module::Install respective cpanratings pages.
Related
I am writing a perl program that I want to share with others, eventually via cpan. it's getting to the point where I should start thinking about this on a bigger scale.
a decade ago, I used the h2xs package maker once. is this still the most recommended way to get started? there used to be a couple of alternatives. because I am starting from scratch with very little recollection, anything simple will do at this point.
I need to read a few long text files (not perl modules) for configuration. where do I put them and how do I access them, no matter where the module is installed? (FindBin?) _DATA_ is inconvenient.
I need to provide an executable (linux and osx). can putting an executable into the user's path be part of the module installation? (how?)
I would like to be able to continue developing it, run it for test purposes, have a new version, repack it, and reupload it easily.
before uploading to cpan, can I share a cpan bundle for easy local installation to downloaders and testers?
# cpan < mybundle.cpanbundle
advice appreciated.
regards,
/iaw
If anything I say conflicts with Andy Lester, listen to him instead. He knows more than I ever will.
Module::Starter is a good, simple way to generate module scaffolding. My take is it's been the default for this sort of thing for a few years now.
For configuration/support files, I think you probably want File::ShareDir. Might be worth considering Data::Section if it's just a matter of needing multiple __DATA__ sections though.
You can certainly put scripts in the bin subdirectory of your distribution, the build tool will put it in the right place at install time.
A build tool will take care of the work-flow you describe.
Bundles are something different. You make a distribution and share the tarball/archive.
If you set up PERL5LIB appropriately, then repeat make test, make install, make dist to your heart's content. For development/sharing purposes a lot of projects do their work on github or similar - makes it easy to share. They have private accounts for business purposes too. Very useful if you want to rewind and see where/when a problem was introduced.
If you get a copy of cpanm (simple to install, fairly lightweight) then it can install from a tar.gz file or even direct from a git repository. You can also tell it to install to a local dir (local::lib compatible - another utility that's very useful).
Hopefully that's reasonably up-to-date as of 2014. You may see Dist::Zilla mentioned for module development. My understanding is that it's most useful for those with a large family of CPAN distributions to manage. Oh - if you (or other readers) aren't aware of them, do check out autodie and Try::Tiny around errors and exceptions, Moose (for a full-featured object-oriented framework) and Moo (for a smaller lightweight version).
I think that advice is all reasonably non-controversial. I find cpanm to be much more pleasant than the "full" cpan client, and Moo seems pretty popular nowadays too.
Take a look at Module::Starter and its much more capable (and complex) successor Dist::Zilla.
Whatever you do, don't use h2xs. Module::Starter was created specifically because h2xs was such an inappropriate tool for creating distributions.
ANSWERED: Basically, it can be done with no major side-effects if you compiled your own perl and you did it the same way your OS did. While it isn't a recommended practice, I've been able to run like this for more then a month. I would conclude it is relatively safe to do if you know what you are doing.
We came to the conclusion at work today that we needed to upgrade perl to 5.10.0
CentOS 5.x comes with perl 5.8.8.
We determined that the effort involved in maintaining scripts with #!/usr/bin/perl was futile.
According to some install stuff on CPAN and other places, it isn't a 'good' idea to replace the OS's version of perl. I already updated the link in /usr/bin/. So my question is, how bad is it really to replace /usr/bin/perl?
I've not noticed any adverse effects in our systems yet, but I'm prepared to correct the link (back to 5.8.8) as soon as there is a problem.
I'm worried that there may be some modules in the CentOS standard distro that aren't included in CPAN's source 5.10.0. I'm still trying to figure out what those modules might be.
Thanks in advance.
In my experience, the best practice is to compile your entire stack (Perl, Apache, ImageMagick, ...) from source yourself. That gives you complete control over which versions of everything are used and when everything gets upgraded.
Replacing /usr/bin/perl with one you compiled is a crap shoot. The OS might be using /usr/bin/perl as part of its maintenance or init scripts so changing it could brick your server or cause strange failures.
So ignore the system Perl, build your own, and fix your scripts to refer to your version of Perl.
Generally newer versions of Perl5 attempt to maintain backward compatibility with older versions. But that's not 100% assured. For example, a script that depends on an undefined behavior in Perl 5.8.8 (which shouldn't happen but sometimes does), that behavior may be different under 5.10.0. Nevertheless, it's usually fairly safe to assume that a script written for Perl 5.8.8 will run under 5.10.0 assuming there are no other factors involved.
But there usually are other factors (modules, byte compatibility for XS code, and so on). The list of possible gotchas is huge. That doesn't mean that any one of them will snag you on this go-around, but there is potential for problems.
If you've already got an upgraded Perl in /usr/local/bin, go ahead and use it. But don't dismantle or upgrade the old /usr/bin/ version. It's only a small chunk of hard drive (very small by today's standards).
By the way, a lot of people speak highly of perlbrew (App::Perlbrew on CPAN) as a tool to help maintain multiple versions of Perl.
Well, if you do decide to change the location of where Perl is installed, that is completely up to you and where you prefer it to be. But, keep in mind that any scripts that exist with a shebang line pointing to #!/usr/bin/perl will possibly break.
My recommendation would be that after you have installed it, create a soft link in /usr/bin/perl pointing to the executable for the new version of Perl that you installed. Just a thought. Its a work around to avoid breaking anything.
Creating the link above would certainly help to avoid the possibility of 'bricking your server' as #Mu pointed out.
Regards,
Jeff
I have read, at various times, both the documentation and a number of blog posts on Dist::Zilla. I have never felt confidence in my understanding of it.
In response to another question, #Ether raised the possibility of converting Crypt-SSLeay to use Dist::Zilla.
So, where can I find a concise guide showing me how to convert an existing CPAN module to use Dist::Zilla? Does the question even make sense?
Update:
The Makefile.PL for Crypt-SSLeay does a lot of work (a lot of it seems unnecessary and I am trying to prune it) to find platform specific include and lib directories, to deduce the version of OpenSSL on the machine where it is being installed. How can I include that functionality if I use Dist::Zilla?
The Dist::Zilla Choose Your Own Tutorial has a page on Converting a Dist to Dist::Zilla. One thing it doesn't mention there is my VersionFromModule plugin, which is useful if you want to replicate the way many people use MakeMaker, with the distribution taking its version number from the main module. (Many people use dzil the other way, with the version in dist.ini and a plugin to stick it into the module, but either way works.)
If I need a more complex Makefile.PL than the one dzil generates, I switch to Module::Build and use my ModuleBuild::Custom plugin, which lets me write my own Build.PL and have dzil drop in metadata like the prerequisites.
The MakeMaker::Awesome plugin lets you do something similar with Makefile.PL, but it wasn't quite what I wanted. Instead, I wrote a MakeMaker::Custom plugin that works much like my ModuleBuild::Custom plugin. The big advantage of
MakeMaker::Custom over MakeMaker::Awesome is that it makes it possible to build your dist for testing purposes without having to do dzil build. For an XS module that has to be rebuilt after every minor change, this is a big win.
These are the sites I have found the most helpful so far, as I'm in the middle of converting a CPAN distribution I comaintain to use it, as a learning exercise. I'm not there yet, but I haven't hit any super tricky bits so far!
The official Dist::Zilla site's Choose your own tutorial's guide to Converting an existing distribution
CPAN Dist::Zilla::Tutorial
Dave Golden's Why I'm using Dist::Zilla
Also, the #toolchain and #distzilla channels on irc.perl.org are full of helpful people, including the authors for Dist::Zilla and other related tools.
I started using Dist::Zilla several months ago. However, at YAPC::NA someone mentioned that they use ShipIt instead. Then today I noticed a .shipit file in miyagawa's cpanminus directory on github, so I decided to look into it some more...
My initial impression is that ShipIt has a subset of what is available with Dist::Zilla, but I don't want to jump to conclusions. So, for those who have had experience with both, what are the strengths/weaknesses of ShipIt vs Dist::Zilla?
crossposted at perlmonks
I'm the author of Dist::Zilla.
I evaluated ShipIt pretty extensively before choosing to go ahead and write Dist::Zilla, and initially they covered almost exactly the same problem space: doing all the boring grunt work of building and uploading a CPAN distribution. All of the features that Dist::Zilla now has beyond ShipIt are later additions, more or less.
If you only need the features of ShipIt, I still advise you to strongly consider Dist::Zilla, for one very simple reason: hackability. If I had been able to not write something new, I would've used ShipIt, but I found it to be underdocumented and difficult to extend. Its plugins were not generic enough and the core behavior made too many assumptions about how you'd like to work.
Dist::Zilla was inspired specifically by this problem: it turned everything into a plugin, and every plugin was given a very, very small interface so that its assumptions would be forcibly limited.
One benefit of ShipIt over Dist::Zilla is that ShipIt has (to the best of my knowledge) no plugins that will alter the way you actually write your code. This means your documentation will still look the same, you will still have a Makefile.PL, and so on. Some hackers don't like that so many DZ-based dists fundamentally change the assumptions of how to test and build CPAN code from its source repository. ShipIt will never change that.
It's possible to avoid using any such plugins with Dist::Zilla, but in general my experience is that people do use them, almost always, in one form or another.
As far as I can tell, my initial impressions were correct.
ShipIt provides functionality for releasing distributions:
keeping track of version numbers
integrating with version control
uploading to CPAN
displaying the changelog file in an editor so that you can edit it before release.
Dist::Zilla, by default, provides the ability to upload distributions to CPAN with a single command (i.e. dzil release). Dist::Zilla also has functionality for creating new distributions (i.e. dzil new My::New::Module). It also automatically generates so many of the files that I used to have to maintain by hand.
Using plugins, Dist::Zilla seems able to provide most, if not all, of the functionality available with ShipIt. It is also relatively easy to add brand new features using plugins.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 7 years ago.
Improve this question
What's the best framework for writing modules -- ExtUtils::MakeMaker (h2xs) or Module::Build?
NOTE This advice is out of date. Module::Build has been removed from the Perl core but lives on as a CPAN module. The pros and cons still stand, and my opinions about MakeMaker still stand.
As the former maintainer of ExtUtils::MakeMaker, I like to recommend Module::Build because MakeMaker is a horror show. Module::Build is so much better put together. But those aren't your concerns and I'll present my "least hassle for you" answer.
Executive Summary:
Because Module::Build support is not 100% in place through all of Perl, start with MakeMaker. If you want to do any customization at all, switch to Module::Build. Since their basic layout, options and interface are almost identical this will be painless. As seductive as it looks, avoid Module::Install.
Fortunately, Module::Build can emulate MakeMaker which helps some, but doesn't help if you're going to do any customization. See Module::Build::Compat.
For CPAN releases using Module::Build is fine. There's enough Module::Build stuff on CPAN now that everyone's dealt with getting it bootstrapped already.
Finally, the new configure_requires option lets CPAN shells know to install Module::Build before they can start building the module. Unfortunately only the latest CPAN shells know about configure_requires.
Oh, whatever you do don't use h2xs (unless you're writing XS code... and even then think about it).
MakeMaker Pros:
Comes with Perl and used by the Perl core (therefore it is actively
maintained and will remain so forever)
Everything knows what to do with a Makefile.PL.
Most module authoring documentation will cover MakeMaker.
Uses make (those who know make can debug and patch the build
process)
MakeMaker Cons:
Requires make (think Windows)
Difficult to customize
Even harder to customize and make cross platform
Difficult to debug when something goes wrong (unless you understand make)
Module::Build Pros:
Easier to customize/subclass
Pure Perl
Easier to debug (it's Perl)
Can emulate MakeMaker in several ways
The CPAN shell will install Module::Build for you
Module::Build Cons:
The Module::Build maintainers (and indeed all of the Perl Toolchain Gang) hate it
Older versions of CPAN clients (including CPANPLUS) don't know anything about Module::Build.
Module::Install Pros:
Slick interface
Bundles itself, you have a known version
Everything knows how to deal with a Makefile.PL
Module::Install Cons:
Requires make
Always uses bundled version, vulnerable to external breakage
Difficult to customize outside its interface
Mucks with the guts of MakeMaker so a new MakeMaker release will eventually break it.
Does not know how to generate a META file using the v2 meta-spec
(increasingly a problem with newer tools)
There are two questions here.
First, never use h2xs. It's old outdated nastiness, though I suppose if you're actually trying to turn a header file into XS code, it might be useful (never done that myself).
2011 update: I strongly recommend taking a look at Dist::Zilla, especially if you think you'll be maintaining more than one module.
For creating a new module, use Module::Starter. It works great, and has some nice plugins for customizing the output.
Second, you're asking what build system you should use. The three contenders are ExtUtils::MakeMaker (EUMM), Module::Build (MB), and Module::Install (MI).
EUMM is a horrid nasty piece of work, but it works, and if you're not customizing your build process at all, works just fine.
MB is the new kid, and it has its detractors. It's big plus is that if you want to heavily customize your install and build process, it's quite possible to do this sanely (and in a cross-platform manner) using MB. It's really not possible using EUMM.
Finally, MI is basically a declarative wrapper on top of EUMM. It also packages itself along with your distro, in an attempt to work around problems with users trying to install modules with old toolchain modules. The downside of the "package self" trick is that if there's a bug in MI itself, you have to re-release all your modules just to fix it.
As far as customization goes, there are some plugins for MI, but if you want to go beyond them you'll be back at the problem of dealing with Makefiles and build tools across a dozen+ platforms, so it really isn't going to help you too much in that realm.
I just uploaded Distribution::Cooker to CPAN. It's what I use to make new distributions. The nice thing about it is that your distributions can be whatever you like: you're just cooking some templates. I don't care if anyone uses it. For me it's simple, low tech, and doesn't cause extra problems.
You might start with something like Module::Starter to make your starter templates then add your own boilerplate and favorite way of doing things. You choose not only whatever you want in each file, but which files show up in the distro. As you figure out how you like to do things, you simply update your own templates.
As for Makemaker and Module::Build, the future is Module::Build. It's only us old guys using Makemaker anymore. :) There are ways to use both (or pretend to use both) at the same time. Look at the Module::Build, Module::Build::Compat, and Module::Install docs. Module::Build was kicked out of Perl's Standard Library and it's future is uncertain. It's back to Makemaker as a build system.
Although this is a bit of a cop-out answer, try using each just to get a little experience with each.
You also might want to look at Dist-Zilla which is a new author-only tool to create distributions. Because it just helps build the distribution, it doesn't ship with your code or do any installation, it can do a lot of powerful stuff.
The only trouble with compatibility regarding Module::Build is when a user tries to install modules without updating their CPAN client (CPAN.pm or CPANPLUS.pm) If they are installing your module from the CPAN, they can just as easily upgrade their client from the same mirror.
If you don't want to do anything complicated in your build process, sure: use EUMM. But if you have a build problem on a different target platform, you might end up in the Makefile, which is different on every variation of make.
Module::Build gives you lots of features (anything you can think of if you extend it) and is all perl so you never end up debugging a makefile. Module::Install gives you features, but you have to bundle it and everything ends up running through 'make' in the end.
I also recommend Module::Build and Module::Starter (with the TT2 plugin).
Module::Build is better by any means, but it is less widely supported than ExtUtils::MakeMaker (more specifically, older versions of Perl don't support it out of the box). It depends on your needs.
Personally, I recommend Module::Install, as do a lot of folks I know - the likes of the Catalyst and Moose folks also use it.
Here's a little clarification of the direction I hoped the responses would take:
pros/cons of various of frameworks
compatibility/install base of frameworks
suitability for internal (local) vs. external (CPAN) releases
not bare "use X" answers
Dave's answer has some good pro/con info. Leon's answer alludes to compatibility but isn't explicit. As brian d foy mentioned, only the old hats use EUMM, but I'm not convinced that MB is a good framework for things destined for CPAN due to it not being part of the core until 5.9.
There are pros and cons to both. These days I use and recommend Module::Build and Module::Starter.
EU::MM still seems to be the most widely supported and popular one, but Module::Build is catching up. Also, check out Module::Starter for a module that will help you get started.