I have a script that builds our redistributable packages. Currently, the script reads a text file containing filespecs for the package, builds a File Set containing files that are newer than the package date and, if any are found, rebuilds the package.
The problem is that files are not always promoted into a release shortly have they are built. So a file may have a date/time that is before the package date, but the version in the package is still an earlier version.
So I want start writing the package build history, including package contents, to a database. When the script runs to rebuild the packages, it'll compare the current file set date/times against the date/times stored in the database for the last package.
Has anyone stored File Sets & compared them in a similar way? I've thought of two approaches:
- Build File Set & ADO Dataset Iterator, walk the ADO Dataset and compare it against the File Set. I don't like this idea very much, which is why I spent time thinking of another (and writing this post).
- Build File Set and immediately populate it into a new "working" table. Then write SQL to compare the two tables and locate any mismatches. If any mismatches are found, the "working" table will be promoted to the "current package" table and the package will be rebuilt.
Any thoughts on these idea or alternative suggestions?
Thanks!