Whenever work on a new task starts, the developer updates his
illumos-clone to current source code, and creates a clone of it (several methods are available and discussed below) named like
For people who understand pictures better, here is an illustration of proposed repository cloning and tiering architectures:
After some development and testing, the developer has committed his changes to the
ws1 repository (with the comment like "
1234 Bug description one-liner"). He should then pull and update the
illumos-clone repository from official sources and update the
ws1 repository from
illumos-clone, and merge any changes. This way the
ws1 repository now contains his work as if it was done against the current revision of the "official" public source codes. Then the developer prepares a webrev and provides it for public review.