Hi,
I'm wondering why you guys are using quilt for managing patches to the nx-libs repository.
I see that the general purpose is to have the original nomachine source base intact/unchanged to track changes easily.
However, quilt is a pain to use (i.e., omitting a patch in the series or changing it), especially taking into consideration that git _already is_ a SCM.
Wouldn't it be easier to actually use multiple branches for that purpose?
[nomachine]--{3.5.0}------------{3.6.0 (future)}----... \-->[x2go]-------\-->------------...
where [name] is a branch named "name", {name} is a tag named "name" and the arrows denote merges.
Note that we don't want to merge x2go project changes back into the original nomachine source, thus not merging the x2go branch into the nomachine branch. However, we _want_ to merge new versions of nomachine's source into the x2go branch at some point.
Also, this setup is very flexible. If, for instance, you decide to still support the x2go nx 3.5.0 branch, even though nomachine nx 3.6.0 has been released, simply branch off the x2go branch, like this:
[nomachine]--{3.5.0}------------{3.6.0 (future)}----... \-->[x2go]-------\-->------------... \-->[x2go 3.5.0]--------...
Afterwards, the nomachine branch (tag 3.6.0) can then be merged into the x2go branch and the x2go 3.5.0 life on happily ever after.
It is way easier to rebase a branch such as to not include a specific commit (cherry-pick), than omit a patch in the quilt patch series.
Not sure if it's only me, but this quilt stuff is really messing with my work flow.
Please drop your thoughts in here.
Best regards,
Mihai