0.4.8

• Merge in pointer changes from #66 (thanks to @jayhesselberth !)
• Merge in changes mentioned in deeptools/pyBigWig#152 , wherein a mismatch in the order between the chromosomes in the header and in the data section caused wrong results. This mismatch should now be supported, though the library will (correct, in my opinion) refuse to create such files.

0.4.7

Various small consistency changes and bug fixes.

0.4.6

Handle non-empty files that have all 0 values in their summaries. Previously a summary indicating 0 bases covered was used to indicate an empty bigWig file. This is apparently not the case for very old files.

0.4.4

Fix the uninitialized value from deeptools/pyBigWig#85

0.4.3

Fixed issue #35

0.4.2

• bwStats() and bwStatsFromFull() can accept the sum type, which is the per-base sum.
• Regenerated the documentation, since I don't think I've done that recently.

0.4.1

• For remote files accessed with https, most error checking of the certificate is now ignored by default.
• You can now create and access bigWig files without entries. It's unlikely that other tools/libraries will support this.
• Support for >65535 contigs has been fixed. Files with up to ~1 billion contigs in the header should be able to be produced and still parsed in java programs (e.g., IGV).

0.4.0

• Renamed io.h to bigWigIO.h, fixing #30

0.3.4

• Fixed issue #26, where an inappropriate bin size was being used for sparse data.

0.3.3

The method used for HTTPAUTH is now negotiated.