gph@hpcc01.HP.COM (Paul Houtz) (03/28/90)
Does anyone know what types of problems are associated with using diff -e and then storing the result and the original and using it as backup file, using ed to restore the original. I know that a file with a line with a single . in it will confuse ed, and I have taken care of that. Is there anything else? Thanks! gph@hpsdesis.HP.COM
gph@hpcc01.HP.COM (Paul Houtz) (03/29/90)
I wrote: >Does anyone know what types of problems are associated with using >diff -e and then storing the result and the original and using it as >backup file, using ed to restore the original. > >I know that a file with a line with a single . in it will confuse ed, and >I have taken care of that. Is there anything else? > >Thanks! > >gph@hpsdesis.HP.COM ---------- Ooops, I guess I wasn't clear in my original question. I know that if the file changes underneath the delta, one will have problems. Let me clarify my question: I would like to use this as a method of transmitting only changes to a file over a network, rather than the entire file. I have other methods (checksum, datecodes, etc) that will make absolutely certain that the changes generated by diff -e will be applied to EXACTLY the same file. What I want to know is, has anyone heard of anything that could cause diff -e to produce output that is not correctly applied by ed? Are there any other characters or sequences of characters like the single dot mentioned above that can blow up diff -e or ed? Are these programs limited by record size or file size? For example, I heard that someone had problems with a linelength of 124 with a "\" character followed by a number of tabs. I have not been able to reproduce this problem. Thanks again! gph@hpsdesis.HP.COM