As I can see from https://github.com/canonical/ubuntu-maintainers-handbook/blob/main/PackageMerging.md#identify-logical-changes I have to go through *all* the git history from 2012 and split all the changes to a separate commits, then do a lot of other stuff. I can barely imagine doing all of that without making a single mistake in the middle which, effectively, will make all the rest of the a rebase job completely wrong (and I'll have to start everything from the beginning).
>In the debian/tests/control file you need to add a comma (',') between the two restriction names. A comment here is that if you merge the changes from Debian we will get 2 more DEP-8 tests (autopkgtest) in the package.
Thanks, will do.
>Nitpick: you committed the debian/files file which is not necessary.
Thanks, I'll remove it.
So, my question is, that if it's possible just to take a clean debian/sid branch, then merge *only* debian/changelog somehow and *manually* put all the necessary Ubuntu-specific patches on top without using all of this heavy machinery with rebasing and editing thousands of commits?
I understand that we have a very strict processes around importing packages, and I understand why. But it looks like this processes are suitable for regularly sync-able packages. Which is not the case for LXC, because most of LXC developers used to be a Canonical employees and Ubuntu LXC package was an independent from the Debian one (for 12 years!).
And another question is that if it would be possible to me to *only* fix the LXC_DEVEL issue without rebasing anything and updating anything to make things just work. Then one day, when we decide how to sort this rebase stuff out with such an deeply out-of-sync package I'll be happy to do a rebase on the debian/sid. I want to avoid Noble to be released with this LXC_DEVEL issue because it bothers developers/users of LXC/LXD.
Dear colleagues,
I have taken a look on this ubuntu- import/ rebase thing, thanks a lot for your suggestions and advice.
Unfortunately, I can't see any possibility to follow this Ubuntu-import way. Because the last Debian-based version of the LXC package was in 2012 (!) [ https:/ /git.launchpad. net/ubuntu/ +source/ lxc/commit/ ?id=b0c54a19dda 57a37294b244982 bcfd425db8dbd8 ].
As I can see from https:/ /github. com/canonical/ ubuntu- maintainers- handbook/ blob/main/ PackageMerging. md#identify- logical- changes I have to go through *all* the git history from 2012 and split all the changes to a separate commits, then do a lot of other stuff. I can barely imagine doing all of that without making a single mistake in the middle which, effectively, will make all the rest of the a rebase job completely wrong (and I'll have to start everything from the beginning).
>In the debian/ tests/control file you need to add a comma (',') between the two restriction names. A comment here is that if you merge the changes from Debian we will get 2 more DEP-8 tests (autopkgtest) in the package.
Thanks, will do.
>Nitpick: you committed the debian/files file which is not necessary.
Thanks, I'll remove it.
So, my question is, that if it's possible just to take a clean debian/sid branch, then merge *only* debian/changelog somehow and *manually* put all the necessary Ubuntu-specific patches on top without using all of this heavy machinery with rebasing and editing thousands of commits?
I understand that we have a very strict processes around importing packages, and I understand why. But it looks like this processes are suitable for regularly sync-able packages. Which is not the case for LXC, because most of LXC developers used to be a Canonical employees and Ubuntu LXC package was an independent from the Debian one (for 12 years!).
And another question is that if it would be possible to me to *only* fix the LXC_DEVEL issue without rebasing anything and updating anything to make things just work. Then one day, when we decide how to sort this rebase stuff out with such an deeply out-of-sync package I'll be happy to do a rebase on the debian/sid. I want to avoid Noble to be released with this LXC_DEVEL issue because it bothers developers/users of LXC/LXD.
Kind regards,
Alex