I am currently working to develop a code which helps me to find a closest string match for a given string from a list of strings. I came across a Perl module Text::Fuzzy which helps me to do that. But i am working on cygwin and a having hard time installing it properly by using cpan commands. So how do i properly install it?
Can't exec "make": No such file or directory at /usr/lib/perl5/5.22/CPAN/Distribution.pm
This means that make is missing. You will have to install make to be able to install a Perl module.
Just do
apt-get install build-essential
Edit: Since you are using cygwin, please refer the link which xxfelixxx has suggested.
Related
While loading necessities into my crouton, apt-get recommended that I install libtemplate-perl. This seemed a jolly idea, and I obeyed.
Reading this answer, I see a fellow traveler install Plack and Starman via CPAN, but then use apt-get to install Dancer.
Minutes before reading said answer, I had installed Dancer via CPAN. And it had worked. It had worked real good!
What happens differently when I install a CPAN package via a non-CPAN package manager? Are there pitfalls I need to be wary of because my libtemplate-perl came from apt-get, or my Dancer came from CPAN?
On Debian or Debian based distros like Ubuntu, CPAN (/usr/bin/cpan utility) installs modules into /usr/local/lib/ by default. And Debian packages keep their files in /usr/share/perl5/ and /usr/lib/perl5/.
It's a better way to choose dh-make-perl tool to package any CPAN distribution not available in your apt repositories, to avoid making mess (serious risk of conflict between apt and CPAN):
dh-make-perl --build --cpan Some::Module
dpkg -i some-module*.deb
Also check out about local::lib and perlbrew.
A lot of tutorials, Stack Overflow answers, walk throughs, etc. will not want to assume that you know how to use /usr/bin/cpan and will instead suggest that you use your system's package manager, which you are more likely to be familiar with.
This is especially true if you are using a tool written in a particular language, but you don't know that language. For example many people wouldn't care whether crouton is written in Perl, Python or Lisp, they just want a tool they can use. Your average Debian or Ubuntu user is more likely to be familiar with apt-get than with cpan.
If you are looking to program in Perl I recommend installing things by cpan, using perlbrew and/or local::lib. If you are just looking to install a tool you can use, I would recommend using apt-get.
Additionally apt-get has the advantage that non-Perl dependencies will be installed automatically. For example the CPAN module XML::LibXML requires that the libxml2 headers are available on the system; the CPAN distribution has no way of stating that as a dependency, and will simply fail on install if it can find the headers to link against. The Debian package can actually specify that libxml2 is a dependency, and will install the dependency automatically for you.
Despite that if you are looking to use XML::LibXML as a Perl programmer, I would recommend installing it with cpan and installing the libxml2 package via apt-get. Having all your Perl modules installed in the same place - again, via perlbrew or local::lib - will help you keep your sanity in the future. CPAN is preferable to apt-get in this case because not all distributions have been packaged for Debian, and so you get a lot more options if you use CPAN directly.
In summary: TIMTOWTDI ;-)
Another difference to mention that I don't see in other answers, is that if you use CPAN to install a module then the module version you get will be set, until you decide to upgrade that module. Whereas, if you use apt and a later version appears in Debian at some point in the future, then apt-get upgrade will update it.
Maybe you want that, maybe you don't. It's neither necessarily an advantage, nor a disadvantage. Simply a difference to be noted.
The good things about apt-get are
that you can uninstall the packages afterwards should you need to
if you are maintaining any number of servers it is quicker to use apt-get as there is no building required
there are definite versions of packages which are tested for compatibility
The downsides are
not necessarily the latest version of packages
not all packages are available
apt-get installs are generally easier and better than cpan because of dependencies alone.
If you are so lucky as to need SOAP::Lite, for example, it is many dozens of dependencies, and a one-line "apt-get install libsoap-lite-perl".
It is sometimes not clear how Perl CPAN names map to their repository package names, but "apt-cache search " is your friend.
Another approach, that can be used for any distro in HOME install as simple user:
Pros
avoid conflicts
up to date version of perl and modules
when you reinstall, you keep your home/perl as-si
you can upgrade all installed modules with $ cpan -r
Cons
you need to upgrade Perl manually
How to
Edit ~/.bashrc or such, and add:
export PERL5LIB=~/localperl/lib/
export PATH=~/localperl/bin:$PATH
Then
source ~/.bashrc
Install latest Perl5 from sources: https://www.cpan.org/src/README.html
Now, you have an isolate Perl installation in your home.
Enjoy the power to install any libs via:
$ cpan -i Whatever::Module
I have perl 5.8.8 installed in /usr/bin/perl
I need to use a later version, so am trying to install another version in a different place.
(nb. I started out trying to install perlbrew but on the linux server I'm using but that's not working - am getting all sorts of certificate problems).
I logged in as root and followed the instructions here, to install perl from source:
http://www.cpan.org/src/
This gave me an install of perl in
/root/localperl/bin/perl
I thought that didn't look right so I copied that the localperl directory to /usr
cp -r localperl/ /usr/
Now I can run a script in my /home/myusername/ directory by using
/usr/localperl/bin/perl
So I guess that looks more normal for an alternate install of perl, though:
a) Am not sure this is correct. So the question is if I stick #!/usr/localperl/bin/perl as the first line in every script, will all be fine?
b) Have no idea what to do to install modules for this new version. So:
i) What to do to build latest versions of modules for this version?
ii) Can I copy across all my existing modules that work with 5.8.8?
(Yes, I did attempt to read the doc and saw there were lots of options for configuring the install, but having tried one or two found this was even more confusing). Any specific help on the above appreciated.
I was using ubuntu 12.04 until 12.10 was released. I used ubuntu for software development and after installing 12.10, i noticed that the perl version (5.14) shipped with 12.10 does not include the Switch.pm module needed while building WebKiT-GTK.
Looking around on the internet i found few suggestions indicating that i should install something call p5-switch from something called ports. I have looked around and was not able to get this done. I am not a perl guy and have no idea where i can get this package.
Can someone please help me as to
1. Where to download the package for ubuntu 12.10
2. In case it is not a .deb, How do i install it.
OR
1. At least be able to downgrade the perl installation to something lower than 5.14
Thanks and Regards
~Sameer
sudo apt-get install libswitch-perl
will install it for you.
"ports" is a *BSD packaging system of sorts, not what you should be looking for.
You can find what package has a particular perl module by going to packages.ubuntu.com, entering Module/Name.pm (in this case, Switch.pm) in the "Search the contents of packages" form and checking "packages that contain files whose names end with the keyword" and selecting the desired distribution, then making sure you ignore false hits like CGI/Switch.pm in the results. Debian has the identical search for its packages at packages.debian.org.
(Note that Switch.pm has serious limitations, was never really intended to be used in production, and should certainly not be used in new code.)
Do make sure you've checked properly that there isn't the Switch module available via apt. If it is available, that's the one you want.
No, then you've two options the longer, correct way and a shorter way that's not quite as clean.
1. Longer
Install cpanm and perlbrew with apt. The perlbrew tool lets you install a complete version of Perl from scratch in a separate directory. Set up a user for your webkit building, run perlbrew as that user, install your perl. Then, use cpanm to install required modules and you are done. A bit of googling will get you step-by-step examples of how to use these tools. If anything goes badly wrong, you can just delete all the files in that user's home directory and start again - all you waste is a little time.
The reason experienced Perl people prefer this is that it keeps the perl you want for webkit-gtk separate from your system perl that ubuntu's packages will expect to be unchanged from the one they ship.
2. Shorter
Install cpanm with apt. As root, run "cpanm Switch" and it will install the Switch.pm package and any dependencies. It will also upgrade any already installed packages it thinks it needs to. This last step is why this option isn't ideal. In the (rare) case when the update isn't compatible with something else on your system uninstalling is fiddly.
I tried:
cpan> i /xls2csv/
No objects found of any type for argument /xls2csv/
cpan> install xls2csv
Warning: Cannot install xls2csv, don't know what it is.
Try the command
i /xls2csv/
to find objects with matching identifiers.
I haven't had any problems installing registered modules, e.g.
cpan> install DateTime
It is just the xls2csv that is providing problems.
Thank you for any insights provided.
P.S. Constrained To Windows with ActivePerl 5.10 (I think) Also I have the xls2csv-1.06.tar.gz file downloaded and tried
cpan> install C:/...path to file locally/xls2csv-1.06.tar.gz
That did not seem to work, and I must install locally.
The indexes used by CPAN.pm are all module-oriented, so they don't see distributions that contain no modules. (xls2csv has only a script.) You can still install them by giving the full path (under authors/id) of the distribution, just as you would to install older or development versions of a distribution:
cpan> install K/KE/KEN/xls2csv-1.06.tar.gz
If you are using my cpan(1) command, you can also install whatever is in the current working directory by telling it to install ".".
% cpan .
That should trigger the normal CPAN.pm process for handling dependencies and so on. You should be able to do that in the CPAN.pm shell too.
If you want to fetch the file from a CPAN mirror, you can do it as ysth suggested. You give it the path under authors/id in the repository and CPAN.pm will find it for you. You're skipping the step where it looks in the PAUSE index because you've already given it the path to download.
It's a 5 year old script with few tests (7) none on a Mac, and one bug report (a year old). It appears to not have been updated since it was written so use this script with realistic expectations. In other words, don't be surprised if you can't get it to work.
Is it possible?
If you download the source code, and read the README file. This will probably tell you you should do
perl Makefile.PL
make
make test
make install
or
perl Build.PL
./Build
./Build test
./Build install
If you download the source code, it will generally have a Makefile.PL. You run "perl Makefile.PL; make; make test; make install" and it will build and install for you.
Obviously if you're not using CPAN.pm, you're going to have to deal with dependencies yourself.
Also, if the reason you can't use CPAN.pm is that you don't have permission to install into /usr/lib/perl, you can force CPAN.pm to install locally, but I forget how.
If you are on a Linux box, a very large portion of the packages can usually be obtained using the built in package manager. For instance, on an Ubuntu system, if you want to install the PostgreSQL Perl module you'd simple do:
sudo apt-get install libpg-perl
You can see a list of the modules for Ubuntu here: http://packages.ubuntu.com/hardy/perl/
I find I can often guess at the names myself. Not sure if this helps at all, but for myself I often find this easier to use than CPAN as it does a lot better at resolving dependencies.
See here: How to install perl modules using CPAN without root
I have just set this up on a server without root access and CPAN does everything automatically.
But if you really wanna install a module without CPAN and you don't have root (assuming this since you don't wanna use CPAN), you can do it as follows
perl Makefile.PL PREFIX=$HOME
make
make install
You're gonna have to hunt down dependencies yourself so it's better to use CPAN.
If the problem is no root access, I would recommend looking at local::lib and also this webpage for CPAN.pm and non-root installation.
But to answer the question as asked, CPAN or CPANPLUS are helpful, but they aren't required. You can always do it the old-fashioned way as Leon says - though usually, it's easier not to.
If you are using Red Hat (Fedora, CentOS), you should use RPM for Perl dependencies wherever possible. Perl packages are almost always named perl-Module-Name, e.g. perl-DBI, perl-Spreadsheet-WriteExcel, etc.
On Ubuntu the naming scheme is libmodule-name-perl.
If the .pm file is pure Perl and doesn't need to be compiled you can just put it in your application's lib folder and use it as normal.
We can install all perl modules both from and even with your terminal in ubuntu. If you are using a ubuntu server then execute the following command ,
'sudo apt-get install "perl_module"'
The modules which you want just give the name in "perl_module" means If you want to install Apache2::Cookie it will be in "libapreq2" so you have to give like,
"sudo apt-get install libapreq2"
I, as others have would highly suggest using CPAN.pm. It is a breeze to use and can resolve any dependencies associated with the module you need automatically.
On the other hand, I would suggest that you read the perlmodinstall document over at perldoc as it gives details on other os' as well.
Regards,
Jeff
If you're asking this because you're having problems with CPAN... you're probably running out of RAM that's why you can't use CPAN.
Maybe you don't have a swap file. Try this:
$ sudo su
# dd if=/dev/zero of=/swap bs=1M count=1k # create a 1GB file
# mkswap /swap
# swapon /swap
Otherwise... stop some services.
$ sudo service mysql stop
$ sudo service nginx stop
...And try again
$ cpan install CPAN
$ cpan install MIME::Lite::TT::HTML