maintaining UNIX system source code

[Kevin C. Smallwood]

In article <874 at hsi.UUCP> stevens at hsi.UUCP (Richard Stevens) writes:
>How do you maintain the source code files for your UNIX ??

At the Purdue University Computing Center, we maintain the source code by
checking all of the distributed source code in when we get it.  For 4.3BSD
UNIX, we checked everything in as revision "4.3" setting the "state" to
"dist" (for distribution).  At the same time, we also forced a check-in of
the same file as revision 4.3.1.1 (on a side-branch).  We might also set a
symbolic name of something  like "4_3BSD" as the 4.3 revision.  Now all local
modifications and fixes occur on the side-branch of the distributed version
of the source file.  Typically, we use the "state" (-s) to either "Local"
or "Fix" or something else more meaningful than "exp" :-).  When 4.4BSD
UNIX (or whatever it will be called) comes out, we will probably check
everything in as revision 4.4 and use the side-branch 4.4.1 for local
modifications.  BTW, we got this technique from Purdue's Engineering
Computer Network (ECN); a technique they used for the kernel code.

Another technique we are using is to actually track the revision that come
from the source distributor.  That is, if the source has an RCS or SCCS
revision of "7.1", for example, we check-in the source file with this
revision number.  We still use side-branches to do the local work.

Kevin C. Smallwood			Internet: kcs at j.cc.purdue.edu
Manager of UNIX Systems			UUCP: ...!{pur-ee,purdue}!kcs
Purdue University Computing Center