NetBSD Developer Documentation:NetBSD pkgsrc Developer Information |
cvs add
" commands)
pkglint -v
" to test your package before
importing it. There must not be any fatal error and there should
not be any warning. If you disagree about any of these message
please contact "packages@NetBSD.org".
cvs import pkgsrc/category/pkg FREEBSD FreeBSD-current-1998-09-28
(where the date and pkg category and name reflect your package).
You should import the FreeBSD "port" unchanged, then commit
the modifications needed to get it working on NetBSD.
cvs import pkgsrc/category/pkg TNF pkgsrc-base
(where the category and pkg name again reflect your package).
NOTE: the pkgsrc/category/pkg path is a relative path for the cvsroot, NOT FOR THE FILES YOU ARE IMPORTING. YOU MUST BE IN THE TOP LEVEL DIRECTORY OF YOUR PACKAGE WHEN IMPORTING!
NOTE: "cvs import" is preferred to "cvs add" because the former gets everything with a single command, and provides a consistent tag.
Critical bug fixes, security updates, and build fixes can be pulled up to the pkgsrc 1.6 branch. Please forward the 'cvs commit' mail from the pkgsrc-changes mailing list to the 1.6 pullup address.
It is critical to ensure binary packages are built against the same set of depends to avoid install conflicts. The most practical way to arrange this is to use the bulk building system and upload the complete set of packages after every build.
Developers interested in assisting (all are invited), should subscribe to the bulk-packages@NetBSD.org list, and check localsrc/admin/bulk-packages for an available arch/osrev combination.
Note: This does not address the issue of new versions of binary packages not installing against a user's existing set of installed depends, or a updated depend potentially requiring an update of many other packages already installed in their machine.
Current options for that would be either freezing binary packages at tagged values, or having multiple binary package trees per arch/osrev. The former significantly reduces the utility of the binary packages, and the latter is not worth considering until we can get _one_ consistent tree per arch/osrev.
|
|