Programming :: Regarding Packaging A Software Release?
Jan 20, 2010
I have a C++ application consisting of several source/header files. As part of the release management procedure, I use a shell script to package all necessary sources/headers in a tar file for a successful compilation on end-user site.However, for example, each time a new directory is created within application's source tree, I need to edit the shell script to implement necessary changes, which is error-prone.What is the best practice to package the source/header files of an application into a tar file ? Is there any tool that I can use for packaging purposes rather than writing my own shell scripts ? For example, how is Linux kernel itself packaged into a tar file ?
The resulting debian package then has an empty "depends" field and no binary.
I used the Qt Hello World example application to start.
I changed debian/rules (generated by dh_make) so that it runs qmake as follows:
Code: Select all#!/usr/bin/make -f # -*- makefile -*- # Sample debian/rules that uses debhelper. # This file was originally written by Joey Hess and Craig Small. # As a special exception, when this file is copied by dh-make into a # dh-make output file, you may use that output file without restriction. # This special exception was added by Craig Small in version 0.37 of dh-make. # Uncomment this to turn on verbose mode. #export DH_VERBOSE=1
I have written a shared library and successfully used debhelper 9 to create a Debian package from source using a Makefile generated by cmake. I then went about writing a python wrapper to that library and wish to package that wrapper in with the library so I can have a single distributable rather than 2 separate ones.
All of my attempts so far have me placing my python source and a setup.py file in the same directory as the makefile at the time where I call debuild.
From here I have tried a couple different configurations to my debian/rules file as seen below:
This try ran make, but completely ignored the python stuff. From some research I have gathered that the --buildsystem flag tells debuild to ignore any makefiles in the directory, which obviously causes a problem in my case.
Another attempt was to modify the build dependency to first run make and then call the python build process that file looked like this
This appears to somewhat work as both processes do build, but a few of the python files are still not getting installed.
Is this the way I should be going about doing this? I've noticed that most python wrappers tend to package themselves individually and then make that package dependent on the library it is attempting to wrap.
Event 1............................................................ full_name: JENNY_JENNINGS genre: f age: 32
[code]....
But as you can see in the input, in the 2nd "Event", the line containing "age" is not present, but in the output my code is printing the 1rst age value twice. The correct output should be blank in the age field for 2nd line in the output like this:
I have faced a problem with my code (a small tcp server). After the thread returns, the memory not decreases, but when a new connection is made, memory not increases and the new connection initialize a new thread with a same thread id of the previous thread. When two connections are made at same time, two thread ids are created and when the respective threads returns the memory not decreases. The Valgrind indicates that not memory leak occurs, the pointers are released. Indeed, more memory are allocate when new thread id is created. i used gcc and debian.
After doing make I have the following questions? How can I see that the build is with debug information? What shall I do to build the release version? Can I have the release version in one directory and the debug version in another? I mean, when I change from debug to release, make will not overwrite the debug version.
I'm pretty new subversion, and have a basic question which I haven't found the answer to:Say that I on machine A modify the working copy of my code, create a tag "release-1.0", and commit it, how can I on machine B point the working copy to the tagged code? For example, say my "code" is actually config files for Apache, and I wan't apache running on machine B to update it's config files based on the files found in tag "release-1.0", how is this done?
I'm sure the solution is very easy, but I just can wrap my head around it.
I've been using Ubuntu so far. There is a system of launchpad repositories where there are lots of programs available that are not in official repo. I've already added rpm fusion but still I miss many programs that I had on Ubuntu. I found this page and I wonder where I can ask for making rpm's. I need several programs:
No package acidrip available. No package bombono-dvd available. No package cdcat available.
[code]....
I now that I can compile from source code but I'm new to linux and it's too difficult for me. I haven't even found program alien on Fedora that allows to install packages from other distributions.
I have an upstream source tarball which, by coincidence, already contains a directory named "debian", but which has nothing to do with the "debian" directory for packaging. Assume also that renaming this directory would be difficult, for example because many other parts of the code would need to be modified. On an rpm based system, this poses no problem to the packager, since the rpm build process is driven by an external spec file and never has to modify the original source. What would be the best way to package this on a debian system? Is it absolutely necessary to rename the original "debian" directory? Or can I make the dpkg build tools operate out of a different directory?
I have nearly completed a project.My team developed using C,shell scripts and qt.I would like to know how to package it as a .deb file??Also during installation,i want it to check for the existence of few packages in the system.If they are not there,i want it to download them using apt.
Code: Failed to initialize Failed to initialize packaging backend This may occur if other packaging tools are being used simultaneously. under more details i get:
Code: There are unfinished transactions remaining. Please run yum-complete-transaction as root. I get to terminal, and as a root I try yum-complete-transaction, but it just says there are no unfinished transactions left. I tried to update thru yum, but this is what I get:
I have a SUSE Linux traffic shaper that can manage bandwidth usage per user and etc.for use it, we must configure LDAP server &client ,edit sudoerr's file, create mysql database with some tables� . Its interface is base on cgi&perl language . How do I package this project? i don't know what part of proj is necessary for packaging (what is source code for creating a RPM package?)
How can I use the preinst script to check to see if the version that is installed older than the version trying to be installed and if so, take an action?
For example: if the software installed is at a version less than 6.14 then take an action, else return 0?
preinst: Code: Select allset -e case "$1" in install|upgrade) # if version installed <6.14 then do something else return 0 ;;
I have error in KPackageKit when i try to install anything or check for updates:
Failed to initialize packaging backend. This may occur if other packaging tools are being used simultaneously.
Details: There are unfinished transactions remaining. Please run yum-complete-transaction as root.
I don't know anything about other packaging tools - I think I don't using any. Yum works fine in terminal. Reinstall PackageKit and KPackageKit files doesn't helped me. I also try to rebuilt rpm db:
rm -f /var/lib/rpm/__db* rpm -vv --rebuilddb yum clean all
" p, li { white-space: pre-wrap; }Cannot get the exclusive lock on the packaging backend. Please close any other legacy packaging tools that may be open." I cannot update anything, and I haven't opened another package downloader. The only one open is KpackageKit. Please help, is this a big or have a I got a virus or something? I don't know what information is needed so if I need to expand tell me what I need to retrieve and ill post it.I'm not good with Kubuntu problems.
I've got a CentOS 5.2 installation with broken package system. When I try to do "yum update" I get lots of "Missing Dependency" errors. How could I restore it to a usable state?
From the GUI, software update gives the error "Failed to Initialize - Failed to initialize the packaging backend. This may occur if other packaging tools are being used simultaneously." In the more details, "There are unfinished transactions remaining.
I run a terminal, switch to root and do as it tells me. It returns saying that there are no unfinished transactions. I've run it several times with the same result. Then I run "yum clean all" and "yum upgrade." The upgrade processes runs and reports "there are unfinished transactions..." However, the upgrade process still completes. Basically it is now a severe annoyance, but the system does seem to upgrade. I just cannot figure out why there are somehow unfinished transactions that I cannot get rid of. I tried reinitializing the rpm database - no luck. How can I somehow reset yum?
I recently installed Fedora 13 on my system, and I'm still in the process of setting it up as I like. As the FCEU and snes9x versions on rpmfusion are old (fceultra 0.98, snes without GUI), I wanted to compile my own RPMs Everything went OK, but I found myself having to build the packages several times because of missing LDFLAGS. I got several errors like "symbol X missing, it was found on Y so you may want to add Y to the linker flags", which I fixed by modifying & exporting LDFLAGS.
So, 2 questions:
1.- why are the LDFLAGS missing? How can I avoid having to set them up manually? 2.- There should be a way to avoid re-compiling the source and only run the rpm packaging stuff with the already built things (some option like 'don't build, only package'). how can I do that?
BTW, I can share my (pretty much hacked) .spec files if you are interested
I was told I need to build / configure a RHEL 5.1 server. I did so using the old ISO I downloaded from Red Hat's site. My question is if I do a 'yum upgrade' command, will that then change my release date from 5.1 to 5.5?
I Have installed oracle 11g release 2 in fedora 12 but after completion of the installation there is no launch shortcut has generated nor any where any icon for launching oracle is given so now how can i run oracle.
Since I use this for work, I don't want the DE to get unstable with updates, but want to do other updates like security updates. Should I disable the KDE Factory Desktop repos? How can I work this so everything updates without taking my KDE into 4.4.1 beta? I ask because I'm not sure yet what comes down through OSS, NON OSS, and Updates repos....or, if I can only do those. I'm just trying to make it dummy proof..I don't tend to break things in linux
I have heard its going to be July. One concern is KDE4.4 and whether this is going to be the KDE default environment for 11.2. Over at KDE, 4.4 is out with a 1 click install for 11.2. The main problem I have discovered from reading the KDE forum is poor multi monitor support with separate X screens. From what I read this is extremely problematic. I would like to update to 4.4 to get the equalizer back in Amorak but attempted once, the day before they officially released 4.4 with no success. I wonder if anyone here is using 4.4 in a separate X screen setup?? Additional thought-- when 11.3 is released will 4.4 be the only option or will we be able to elect to use 3.xx??
I just updates to 4.5.4 from 4.5.3 from the openSUSE 4.5 stable repo and when I open aboutKDE (from most KDE apps; Help->About KDE), it says I have KDE 4.5.3 Release 1!
I was wondering, is there ever more than one build of a given Ubuntu release (but with the same major/minor version)? That is, if I download X.10 a month after it comes out, and then I download it a few days before the next version is released, say, will it be bit-for-bit the same as the one I originally downloaded? Or does the Ubuntu team do new builds over time, fixing things, etc.?
What is the reason to drop Xdialog from the new release; is any replacement on the horizon? (beside zenity that is really crippled version of Xdialog).
