I look at porting a large set of IBM Assembler Language programs, from vcs based on "file names, including version numbers," to modern vcs, which will give me, among other things, the possibility of branching and merging.
These files contain entries from 80 columns, the last 8 columns are almost meaningless serial number.
For a number of reasons why I really don’t want to waste space going inside, I need vcs to ignore (but reliably keep in a certain specific order) columns of the sequence number, as well as diff and patch based only on the contents of the first 72 columns.
Any ideas?
Just to make it clear “ignore but save”: I accept this a bit vaguely, as I have not yet fully compiled my ideas.
It would be something like this:
"When merging / patch, if one side has serial numbers, output them, if more than one side has serial numbers, use those that are in the file (1 | 2 | 3)"
Why do I want to keep serial numbers? Firstly, they really are ordinal numbers. Secondly, I want to return this material to the mainframe again, where serial numbers can be very significant. (Those of you who know what “SMP / E” means will understand. Those who don’t do this, be happy, but tremble ...)
I just realized that I did not accept the answer. Tough choices, but @Noldorin comes closest to where I need to go.