Ubuntu :: Building Multiple Versions Of One Software?
May 26, 2010
I am not sure if this is the right section to ask this question (sorry if it isnt). Anyway, I want to build and install 3 different softwares:OpenCV, Player, Stage. These are required for my school projects and research etc. Now, I need to install them in this manner:
Opencv-2.0.0, player-2.1.3 and stage-2.1.1 and Opencv-2.1.0, player-3.0.1 and stage-3.2.2
The reason for this is, these are the sets of versions that talk to each other without problems. And I need the older versions for a project that I'm currently working on and its always nice to have the newest version installed on the system. If its only a single version install, its pretty easy and I can do it.Since I want to install multiple versions of the same software I could use some help. I am not sure whether I can install all of the them in the default directory (/usr/local/). So, I can set up a separate directory on my home folder. But thats as far as I've gotten to. I am particularly concerned whether one version would break another. Is there anything that I can do to avoid this?
I'm using latest Ubuntu version 10.10 and I want to try Firefox Beta, but don't want to replace the one I;m using now which is 3.6.13.
I have downloaded Firefox 4 Beta andd extacted file to a /Home Folder/(My username)/Firefox 4 (Beta)/
and I wen to that folder and clicked on the Firefox link bu version 3.6.13 came up. in another window. Do I have to close Firefox ad re open it or do I have to do something else?
I have installed GCC 4.5.1 using senaptic. But typing gcc in any terminal still calls the older version of gcc. I see that in /usr/bin gcc is a link to the older gcc. Would it be OK to replace that link with the new one poiting to the newer version of GCC?
I have a dual boot desktop with WinXP on C: and Ubuntu on D:. GRUB 2 handles the start-up. Having installed multiple updates to Ubuntu, my GRUB menu now lists:
and (recovery mode) for each of the above, along with the mem test and WinXP.Is there any reason I need to list all these variants of Ubuntu? If not, can I simply edit them from the GRUB config file or must I do something else to actually uninstall them?
On my lenny+backports+debianmultimedia desktop I see two annoyances. 1st: There are multiple versions of same packages. But while attempting to remove the older versions I see apt-get is removing some essential packages. Is there any way to kill this duplicacy of packages? Here is a list of packages whose multiple instances are installed:
For my work it is extremely useful to have multiple versions of gcc available. I've done this in the past under Ubuntu simply by enabling alternative or older repositories but in Fedora land I've not been able to find a standard automated way of doing this. Can anybody point me in the right direction?
The main problem is that the software was developed about 15 years ago and is incompatible with the most current version of Java. I would like to know how to install an older version of Java so that I can use it with just this one application, without changing the version of Java that I use for everything else.
When I turn on my computer, because of frequent updates it will display several versions of Ubuntu 10.10 that I can choose from. I wonder if it is possible to delete some of the older versions and how. I think having several versions of Ubuntu uses up a lot of space in the hard drive.
Some times i need the older version of some packages in my system...side by side with the new one.i.e mostly i found older version openssl package for supporting dependency of some programs Can I keep the following libraries at the same time in my system by installing multiple versions of openssl?
I do quite a bit of testing of a FOSS package, which I install from rpm. At any one time there may be 3 versions ( current release, new release candidate, and snapshot of trunk) that I wish to test. If I stick to normal naming conventions (i.e. current release fred-2010.0...rpm, release candidate fred-2010.2...rpm, and trunk fred-2010.3...rpm) it seems to me that I can only have one version installed at any one time. Is there a 'proper' way of managing this (using the package manager), so that I can have multiple versions installed at any one time?
I want to have 2 diffrent versions of a package installed at same time. This is the command i use "apt-get install myprogram=versionID" Problem is, that when i install the version i wanna add, it REPLACE it with the old I dont wanna replace, i wanna keep them both.
I have matlab installed on a network (I am not the administrator) and we usually start the program by typing "matlab", then choosing one of the version options from the menu i.e. typing "n". So because of this, trying to run matlab programs or commands directly like this... matlab -r ProgramName
...does not work. I just get the menu as usual and everything else is ignored. I assume the admin has overridden the matlab command with their own custom script. So my question is can I start a specific version of the program by specifying the folder that the version is in? I thought it might be something like /opt/matlab/version -r programname
this might look foolish, but I am a bit of a linux noob. Let me know if I should just ask my administrator but I thought there might be something easy I am missing.
I have a page which is going to be internationalized, and available in more languages. It contains PHP scripts to load, let's say, current user's data from database and the internationalized content itself, like "Welcome user" message. The problem for me is the fact, that internationalized content is not continual, and it's all over the page mixed with php scripts.
I don't want to use eval(). I've got 2 , they are, however, not good enough. 1. One file per language version, with scripts included - there will be many languages, so there would have to be many files with redundant data. Also if I wanted to change structure of script, I would have to change it in all pages. 2. Load international data from db, while scripts are on the page - not sure about good database structure I mean, how would I get the right content from database? (content would be split into rows, columns, or something?)
I'm missing a fundamental that I just cant seem to wrap my head around with setting up repositories with RepoSync. I set up a local repository when my system was running version 5.2. Everything ran like a top up until the release of 5.3 Upon release I ran the standard yum on my 5.2 server which upgraded it to 5.3 and thats where my mental block is catching me up.
When I run a reposync It seems that all I'm able to download is the 5.2 packages. I'm trying to figure out how I can manage both a 5.2 repo along side a 5.3 repo with reposync and figure out what needs to be done to properly make reposync pull down the latest updates for the newer version. My original thinking was that once my repo server was upgraded to 5.3 it would start pulling the 5.3 updates but obviously I'm way off base since that is not happening.
I'm having a problem building some applications (e.g. Qsynth and FreeCAD) with sbopkg relating to an unfound library named "jscore". I've been keeping my system up to date with slackware64-current.Google searches have turned up very little about this library. There is a sourceforge page that has no files and seems to be related to a google hosted script jsCore.js... but I don't think that's what's missing here. I've come across some vague references to this jscore library being part of webkit. Can anyone verify this? I updated my webkit library itself and found no apparent results. Perhaps jscore was removed in an update to Qt (both of the example applications mentioned above link to Qt) as it seems to have it's own webkit library, libQtWebKit.
I am building a "Windows-less" HD (no versions of MS Windows). I have acquired the first of two 500GB Seagate Momentus drives for this purpose. The Seagate was selected because the drive being replaced is the exact same geometry, however the new drive is 7200rpm vs 5400rpm. If this process is successful, I may repeat with an SSD. The current drive is partitioned as follows:
I'm not to clear on the difference between LTS versions and other versions, but think I may want to go with LTS. Can someone tell me if my thinking is correct given the following situation: I have some very cool, but very expensive software installed with a group license from my school, a school which I am not going to be attending for too much longer. So I want to go as long as possible without reinstalling Ubuntu, because once the product is licensed it will be licensed until I reinstall Ubuntu (or I uninstall the program). So I think this is going to require me keep the Ubuntu version I install as long as possible.
So in this case, should I go with 10.04LTS or should I just install Natty Narwhal and keep that as long as possible? It looks like 10.04LTS will be "supported" longer, but I'm not exactly clear on all that "supported" entails. Presumable it means security and software updates will be available for 10.04LTS for much longer than 11.x versions? So I'm thinking I should go with 10.04LTS
Is my thinking correct in going with 10.04LTS? Edit: It was pointed out that this would be against my contractual agreements. Which I suppose is probably true.
We are working on a project to create a display wall of 8 monitors arranged as 2 high by 4 wide. Each monitor is connected to a single machine and all machines are networked with a master machine with its own, seperate monitor.
Our goal is to get the 8 machines to share a single desktop, with the master machine acting as the server. We have looked at using Xinerama or NMM, but we are unsure about how to get started configuring the multi-machine, multi-head display.
My ubuntu 10.04 has two monitors connected both being treated as individual desktops.But only one desktop is showing the wifi network icon in the notification area.Is it possible to have more than one wifi icon going at one time?
I often use the rpl command to make changes to multiple html files at once. For example:
rpl -R '<br />' '<br /><br />' mydirectory However, I haven't been able to figure out how to change multiple lines. For example, let's say I want to change all occurrences of :
I'm trying to connect one computer to two others in an ad-hoc infrastructure.
[computer 1] ---- [computer 2] ---- [computer 3]
computer 2 is running Linux and has a single NIC wlan0. I want to it to connect to both computer 1 and computer 3 so each computer can talk to the other. No switch is available so it needs to be an ad-hoc setup.
So essentially, I have 3 monitors and i'm attempting to run 2 different window managers/sessions at once split up on the monitors. It is intended to be like this:
Code: [GNOME/Compiz - Screen0] - Left Monitor[code]......
I'm starting to have A LOT of opened windows in my machine. Sometimes within a project, I have e-mail/task management/personal e-mail/twitter, and a lot of different opened applications/terminals in my Linux workstation.Sometimes it would be interesting to have different workspaces to projects instead of this configuration I have nowadays that are classes of work (bad name, I know, but I think you got the idea).I'm starting to think about using two monitors: one with Corporate Management, Work and Personal. The second monitor is only the development state: each workspace here is about a project being worked on instead of groups of works like before. A workspace may be implementing different classes for example.
My question is: I just want to change to a second monitor using the mouse. I want to still be able to change workspaces in the same monitor using keyboard shortcuts. The keyboard shortcuts wouldn't change monitors, just worskpaces on the same monitor. All the tutorials I read (like this one) only tells how to use multiple monitors but doesn't answer my question about keyboard shortcuts.Does Linux (Ubuntu 10.04 Lucid Lynx or Debian 5.0.5 Lenny) support this envisioned setup (Different workspaces in a way that keyboard workspace switching only works in the current monitor) ? If so, how?I haven't tested this setup, that's why I'm asking. In this question the user says it works exactly how I want it to behave, can someone else confirm it?
installed the latest version of R ( R.2.11.1) in my machine and I can run that by simply typing "R" in terminal but I need to install the old version of R ( R.2.9.2) since some packages don't support the new version. I have the .tar.gz file of old version. How can I install that? How do I run in terminal ?
Just tried to run an apt-get update on my 8.10 Ibex server and got lots of errors about missing Packages.gz files. Checking the archive site [URL] shows that all Intrepid directories are missing!
