Installing from repositories
 

I want to install HDF5 on my linux computer.

Eh, can I do this from repositories- if so how, and why?

I downloaded the source package from github but i'm not sure how to build it...

there's a setup.py file. What do I do with this?

Thanks...

Hi,

There are hdf5-tools and hdf5-helpers packages in universe.

To build from source run
Hope this helps

Thanks, what's this error mean?

Installed /home/chris/anaconda2/lib/python2.7/site-packages/Pydap-3.1.1-py2.7.egg
Searching for hydroffice.bag
Reading https://pypi.python.org/simple/hydroffice.bag/
No local packages or download links found for hydroffice.bag
error: Could not find suitable distribution for Requirement.parse('hydroffice.bag')

What does it mean? That you are missing hydroffice.bag.

Usually for python stuff. I use pip to handle python application installations.

http://docs.h5py.org/en/latest/build.html

okay cool.

conda install h5py seems to be working.

What's going on there- is it pulling it from the repositories?

Also this probably seems stupid but after it's installed... how do I open it?

wait, what happened here?

Please use code tags to post code.

Sorry, have edited. Will do in future.

Chris, have you tried installing from repositories?
Does it work at the end of the day?
Here are lists of files for hdf5-tools and hdf5-helpers if you are lost.

with this user, you really have to look at their posting history.
started about a dozen threads about installing python stuff, ignoring the obvious advice of "use the repos" and always starts a new thread instead of trying to solve the problem first.
all in the last 2 weeks.

I'm sorry, I'm trying my best to learn. This is a forum for helping new users, not for criticizing them surely?

But I still can't get my head around this repository stuff guys. No one has took the time to give me a proper explanation, instead just swatting me with "use the repos".

If I'm correct, can you just use the command

will this install stuff straight from these repository things? I have 'googled' this, I know they're archives of online software programs, I know there's 4 of them... but I'm not exactly sure how to use them.

Looks like Chris is not yet familiar with what repositories are, hence attempts of building. For newcomers it's okay.

Each distro is tied to a pool of software called a repository. The software in the repository is verified to work properly with the distro. Installing software from outside the repository can and does cause problems. While Ubuntu and Debian both come from the same base (Debian), they are not the same. Installing a Ubuntu based program in Debian may work or may not and vice-versa.

Pick a distro and only install from the repository until you understand what you're doing. As Ubuntu (and all of its derivatives) has a huge repository, it's likely that all you need is there. It can be installed through Software Center, Synaptic (which I use) or apt-get (which I also use).

You really need to take the time to do some reading so you actually understand how the whole package management philosophy works. I've seen all your posts and you have yet to show you have paid any attention to the advice you've been given and soon folks are going to just give up helping you...

Hi NGIB,

Thanks for the advice. So if I wanted to install a program from the repository on my system using apt-get, I would say something like

and it would install it?

Yes, but you need to know the exact package name.

Also, you need to stop posting multiple threads about the same issue. As I said in your other thread, install Synaptic and you can use it to figure out what you're looking for. Rather that continue to ask the same thing fifty different ways, perhaps you need to follow the advice you've been given and actually put some effort into learning what you're trying to do.

If you continue the path you've been on, I suspect folks will add you to their ignore list - I know I will...

A short guide (I prefer using plain apt instead of apt-get, apt-cache etc.):

Prior to all you have to update information about packages in repositories:
First you want to search for packages; their names can differ from the names of source packages. Repositories contain precompiled binaries, which means that you don't have to search the world wide web for sources. So, to search:
An example output:
Second, to install:
Consult with apt search to type a proper name of a package.

Third, to remove a package:
or to remove completely, including configuration files:
When removing something, consult with dependencies - packages that will be removed as well because this particular package is needed for them to run. An example:

Finally, someone who has taken the time to explain things clearly.

Thank you Mitt! This makes a lot more sense than what I've previously heard. I'm assuming that you could download other package managers like Synaptic to do a similar job, but using apt, which is the standard package manager for Debian, is also basically the same thing, it's just a matter of user preference?

And I just have one more question:

I guess that trusty-updates is a particular type of mirror repository? Why is my system using this instead of either Universe etc.?

Ok, I'm done with you. You do realize that we all expect you to actually put in some effort to read and understand on your own? I spoon feed my one year old grandson and that's my linit...

dpkg is actually default in both Ubuntu and Debian but it doesn't deal with repositories, and is only used to deal with local packages (for all features type man dpkg). APT is kind of a front-end for dpkg with some additional functions, like online repositories.

The file /etc/apt/sources.list contains information on repositories that APT uses. Please, post output using Ubuntu divides software into four main sub-repositories:

Main - contains only supported by Ubuntu free software;
Universe - contains unsupported but free software;
Restricted - contains supported non-free software;
Multiverse - contains unsupported non-free software.

Updates contain newer versions of packages, provided by these four.


So, print the output, we'll see what you have.


By the way, man pages are a great source of information about a programme. The descriptions there are highly technical and informative. It's a good habit to read them fully, though it surely requires some time.

Just my two cents.

Using command-line tools is by far the best way to get used to a command-line interface. They have advantages:

• they are simple
• they are fast
• they will help you if your graphical user interface is broken
• it's a traditional Unix way
• they will help you understand your computer better

Thanks Mitt. This stuff is pretty overwhelming to get your head round at first, your help is definitely speeding up my learning curve! Okay,

Put where lacking universe, multiverse, etc., so it will look like this:

Ummm ... instead of being rude and stating that you're done with him, why don't you save your breath
and my time and simply remain quiet?


Thanks,
Tink