Fix typo in PDFSet.cc use of default uncertainty CL in percent. Thanks to Christoph Schwan for rapid detection and debug.

Add parser the extended error-type string format

See merge request !32

Reduce distutils dependency

See merge request !31

Move distutils to being first-preference Python installer again, and fix a couple of small C++ issues flagged by clang

Remove distutils dependency from configure, fall back to using setuptools if possible in Python extension build, and update Python version detection macros.

Disable FileIO thread-locality in MPI mode. Add thread locality feature record and attribution to ChangeLog

Fix data races when using multi threading

See merge request !27

Assemble the PDF-set index from _all_ pdfsets.index files found in the search

See merge request !28

Lazily cache the PDFSet error-info struct. Chain the in-place and return by value PDFSet::uncertainty() methods in the more efficient order. Add PDFSet::uncertainties() methods to compute multiple observable uncertainties at once.

Add Python support for the extended error scheme, and extended/new PDFUncertainty and PDFErrInfo structs

Resolve "Locale mechanism for numeric parsing not supported on GCC11"

Closes #6

See merge request !22

Use the PDFSet::lhapdfID() function by preference to return a PDF::lhapdfID(), thereby avoiding the index-file lookup if possible.

Add pdfsets.index and pdfsets.dox for PDF4LHC21

See merge request !30

Update pdfsets.index and doc/pdfsets.dox for MSHT20qed and nCTEQ15

See merge request !29

Convert PDF-set check, correlation, and sampling functions to extended error scheme, and add PDFErrInfo::coreType() to clean up code

Rework the error-info struct to explicitly store the envelope components and use their single/symm markers. Error computation now looks ~correct, and user code is a bit more elegant.