Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/lelutin/pkg-scout

Debian/Ubuntu packaging system and files for Scout project
https://github.com/lelutin/pkg-scout

Last synced: 6 days ago
JSON representation

Debian/Ubuntu packaging system and files for Scout project

Awesome Lists containing this project

README 

        
How to use

==========



First of all, make sure the values for the variables are right (See

configuration section below). Then, update any file in subdirectories named

after the distribution versions that you need to modify. Make sure to create

all the pbuilder environments that are necessary. If they already exist on your

computer, update them with either "prep" for all distribution versions or

"prep-*name*" for a specific distribution version name:



    $ make prep

    $ make prep-jaunty



To build all packages, simply call:



    $ make



To build only a subset of the distribution version, specify them as make

targets:



    $ make maverick lucid



To upload packages to Launchpad, either use the "upload" target for all

packages, or "upload-*name*" for a specific distribution version name:



    $ make upload

    $ make upload-karmic upload-hardy



To remove all built files and residues, use the following in the top directory:



    $ make clean



When developing

---------------



When working on modifications for the packaging, it is useful to have a work

copy set up. To assemble the *source* + *debian* directories under the

build/*distro_version* directory, Use the dev-*name* target (where *name* is

the name of the distribution version you'll be working on):



    $ make dev-testing



You can then 'cd' to the build/*distro_version* directory (in this example,

build/testing) and hack away your modifications.



While working in this directory, you can use the Makefile that is automatically

copied by the dev-*distro_version* target. To build the .dsc + .changes files,

use the 'dsc' target and to build the .deb package, use the 'deb' target. To

build them all, use the default target ('all'). To upload the currently built

package, use the 'upload' target. Finally, to remove files that were created by

the build process, use the 'clean' target:



    $ make

    $ make dsc

    $ make deb

    $ make upload

    $ make clean



One last target that you can use when working in the build/*distro_version*

directory, is the 'apply' target. This target copies all changes made to files

inside the *debian*/ directory in the dev environment to the topmost

*distro_version*/ directory:



    $ make dev-testing

    $ cd build/testing/scout-0.3.2/debian

    $   # and test out your modifications ;)

    $ cd .. && dch -i

    $ cd .. && make apply



Using this last feature is dangerous for your files -- it overwrites your files

inside *distro_version*/ -- so if you want to use this, make sure to have

everything under a versioning system.



Configuration

=============



If any detail changes for the project, or if you'd like to use this makefile

system for another project, here are the values you can adjust.



Makefile

--------



The names of the distribution versions that the "all" target builds are set in

the DISTROS variable:



    DISTROS := unstable lucid



The project name is set in the PROJECT variable:



    PROJECT := scout



Makefile.stage2

---------------



The repositories where packages are pushed are set in UB_REPOS_URL and

DEB_REPOS_URL for Ubuntu and Debian, respectively:



    UB_REPOS_URL := ppa:your-name/ppa-name

    DEB_REPOS_URL := some_url



Organizing pakage contents

==========================



The directories at the first level that are either named 'common' or have the

name of a distribution version represent the contents of the debian directory

for each package.



While putting up together the build directory for a distibution version, the

makefile first copies all of the files under the *common* directory. It then

copies all the content of the version specific directory over what is already

there. So you can think of directories named after a distribution version like

the overrides on the common files.



For example, say you have the following structure:



    common

     | compat

     | control

     | copyright

     | pycompat

     | rules

     | watch



    dapper

     | changelog

     | control



    hardy

     | changelog



This means that in the end result, the *debian* directory for Ubuntu Dapper

will have 5 files from common (compat, copyright, pycompat, rules and watch)

and two files from the *dapper* directory (changelog and control). Also, the

*debian* directory for Ubuntu Hardy will have all files from common plus the

changelog file from the *hardy* directory.



This mechanism lets you centralize files that shouldn't differ from one package

to the other, like the *copyright* file while having the flexibility to

override files that can differ like the *control* file.