irssi-cvs moving to irssi-svn

Today I’d like to bring you some information on the pending irssi-cvs to irssi-svn move. Upstream Irssi developers changed from CVS to Subversion as their version control system already a couple of months ago and we’re now following this change with our live ebuilds.

Now to the point that makes this a little bit tricky. irssi-cvs is currently marked stable on hppa, ppc, sparc and x86. This is just inappropriate for a live ebuild, because we (as in Gentoo developers) have no way to verify the state of the source code. Moreover we can’t verify the integrity of downloaded files, that means we expose our users to some security risk.

To cut a long story short, we’re going to keyword irssi-svn ~arch only and I’m going to push this move out as an automatic move via our regular updates files. The new irssi-svn ebuild has a version number higher than the old irssi-cvs ebuild. For our ~arch users this means that they will see this as an update to irssi-svn after the move is done on their local system. For arch users with current portage versions the move will be done and they will see no update and no error message, if no package depends on irssi-svn. They will stay at the same version of irssi they had before. If they want to emerge irssi-svn, they will get the familar ‘All ebuilds that could satisfy “irssi-svn” have been masked.’ message. Stable users are free to take a look at “Mixing Software Branches” from our handbook, on how to use irssi-svn on their stable system.

This entry was posted in net-irc. Bookmark the permalink.

6 Responses to irssi-cvs moving to irssi-svn

  1. Would it be more useful to you for ~arch users to stay or use the -svn version?

  2. swegener says:

    The only difference is that we now use Subversion to get the source, but the CVS code has not been updated since they moved to Subversion. That means by using the -svn version you will get access to more recent source.

    The -svn version has only been tested by developers and only for a limited time. It’s still package.mask’ed and will be up to the point when I push that move out. All irssi ebuilds have blockers against the other irssi packages, so no user should have several irssi packages installed at the same time and hence I should already be able to remove the package.mask entry.

    I don’t expect it to break, but having ~arch users test it would be good. That’s why the version is higher than -cvs versions, so that it gets shown to ~arch users as an update.

  3. Lisa Seelye says:

    I’ve been using irssi-cvs for some time now. I’ll be happy to give the -svn version a go.

  4. hm says:

    Shouldn’t you use -arch for live ebuilds ?

  5. swegener says:

    A keyword signifies that the ebuild actually works on a specific arch. -arch means that it doesn’t work, which isn’t true, even for live ebuilds. package.mask would be the right way to go to “prevent” users from installing irssi-svn. And yes, I thought about package.mask’ing it, but that’ll make the transition even more complex.

  6. Annan says:

    A keyword signifies that the ebuild actually wodks on a specific arch. -arch means that it doesn't work, which isn't true, even for live ebuilds. package.mask would be the rkght way to go to "prevent" users from installing irssi-svn. And y3s, I thought about package.mask'ing it, but that'll make the transition even more complex.;

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>