> come up with a minimal and self-contained reproducer, so that someone who does can try it and see what you see on your end. So far you have only provided some terminal invocations accompanied by error messages and vague descriptions of comparing the actual and expected output of pg_dump.
Did you miss my e-mail starting with "After taking a longer break from my problem, I have now made a fresh, clean, scientifically conducted experiment in order to truly get to the bottom of this annoying problem once and for all."? (I don't now how I'd link to it.) There I go through everything with full commands and a minimal test database (which I also describe how I created).
I don't see how I could possibly create a smaller or clearer example of the issue. Whatever "strange things" cmd.exe does clearly doesn't show up for my cross-platform PHP CLI script. (I guess it's theoretically possible that PHP for Windows does some sort of "processing" automatically which pg_dump lacks, but I doubt it.)
I haven't missed that. I'm trying to point out that your description of the problem is not self-contained: it lacks the schema definitions and test data.
Didn't you mention that once you could overcome the runtime errors you still have seen some unexpected output from pg_dump? A script that one can download and run on their (Windows) system would help to confirm the problem (if it does reproduce on others' systems) and speed up diagnosys and a fix, if required.
My expectation from such a script is, that it:
1. Connects to a postgres server on localhost:5432.
2. Creates a test database and re-connects to it.
3. Creates the test schema and table(s), and populatest the tables with the test data.
4. Issues the pg_dump command.
5. Compares the resulting output to the expected one (or runs checks against the output produced by pg_dump).
Kind regards,
--
Alex