Resurrect pdf_builtin

[Juergen Reuter]

It's not working. Seems to be the setup of the paths, which traces down through integrations to processes.

[Juergen Reuter]

With r4551, the path is correctly set again for the installed version. To prevent the dispatch unit tests to depend on external path variables, there a dummy path "." is used. Still, the PDF set cannot be correctly set yet.

[Juergen Reuter]

One more thing connected to comment1: At the moment, the unit test is copying the PDF table over from its place in share/pdf_builtin or so. But here also the proper test path should be used! This is connected to the qcd_parameters ticket (#513).

[Juergen Reuter]

The previous (2.1) pdf_builtin SINDARIN file is still not running through as still it is not possible to set other PDFs. But this might be because they are set as optional arguments? (cf. #516). Ranking it up anyhow....

[kilian]

Done in r4651, including alpha_s access. The PDF set can be defined globally. Local options are indeed not implemented yet, but this is a different ticket (see above). Closing this one.

[Juergen Reuter]

This seems not to be correct in tests/Makefile.am, WHIZARD should know the correct paths (now):

PDF_FILE = cteq6l.tbl
$(PDF_FILE): $(top_builddir)/share/pdf_builtin/$(PDF_FILE)
        cp $< $@

sf_pdf_builtin.run: $(PDF_FILE)

So I guess this has to be removed. Right?

[kilian]

Yes, this was obsolete. Removed in r4668.