v4.0.1

• Retain miniSEED v2 sequence numbers when re-packing records during pruning

v4.0.0

A new major version that supports both miniSEED v2 and v3.

This release also removes many of the more esoteric features of the program in order make it more maintainable.

• Port to libmseed 3.1.2 with support for both miniSEED v2 and v3.
• The modification log now includes SourceIDs instead of Net,Sta,Loc,Chan.
• Byte range annotation on input file names now uses a - between start and end offsets to match libmseed. Legacy delimiter : also supported.
• Add -m option to match source ID against globbing pattern.
• Add -snd to skip non-miniSEED data, otherwise quit on unrecognized input.
• Add -VCHAN archive layout option that uses publication version.
• Allow -Q to accept publication versions in addition to RDQM codes.
• Remove skipping of channels on decoding errors, output untrimmed record.
• Remove the -szs (skip-zero-sample records) option.
• Remove the -lso (longest segment only) option.
• Remove the -msl (minimum segment length) option.
• Remove the -sb (staging buffer) option.
• Remove capability to replace input files and the -rep and -nb options.
• Remove -R (source ID rejection) option.
• Remove match/reject list file capability.
• Remove -sum (basic summary) option.
• Remove -S[dhm] (split boundary) option.
• Remove -rls (split on record length change) option.
• Remove -mod (modification summary) option.
• Remove warning about LIBMSEED_LEAPSECOND_FILE environment variable, now included in libmseed.

v3.24

• Fix heap corruption fault due to free'ing just-allocated archive file entry.
• Avoid needlessly closing open files when seeking to remain under open file limit.

v3.23

• Properly handle larger file sizes.
• Update libmseed to 2.19.8.

v3.22

• Update libmseed to 2.19.6
• Treat archiving writing failures as fatal errors.

v3.21

• Update libmseed to 2.19.5, with fix for leap second calculation

v3.20

• Update to libmseed 2.19.4
• Reorganize miniSEED output writer such that re-packed records that result in two or more records are properly included in the output. This handles the odd cases where removing data from a record results in more Steim compressed data than before.
• Add -SDS layout option with PR from @bonaime
• Add warning for impossible combinations of selections that specify multiple, distinct time ranges from a single record.

v3.19

• Update libmseed to 2.18.
• Read any leap second list file indicated by the LIBMSEED_LEAPSECOND_FILE environment variable and use it when determining data coverage.
• Fix edge case of pruning data and splitting on time boundaries when resulting record only contains a single sample.
• Reformat source according to included clang-format profile.
• Remove inappropriate Makefile.wat, src/Makefile.wat and src/Makefile.win. This will not build in Windows without a replacement for system regex.

v2.18

• Update libmseed to 2.17.
• Add error check in trimrecord() for errors returned by msr_pack().
• Add Warning in trimrecord() when data are repacked into more than one record, data loss is certain. There is no easy work around in dataselect to output more than one record at this point, hopefully this is rare.

v3.17

• Each input file may be specified with a byte range to read. Useful to limit dataselect's operation to specific parts of files.