Exomerge field sorting fixes

Quick release to make the exomerge field sorting fix available.

Main change in this release is to fix the sorting of higher-order (tensor, vector, quaternion, ...) field names in exomerge.
Few minor changes related to Trilinos build/configuration changes.

What's Changed

• New implementation of _sort_field_names based on Python list sorting by @Damian-AI in #682
• Modifications to exomerge field name sorting by @gsjaardema in #683

New Contributors

• @Damian-AI made their first contribution in #682

Full Changelog: v2025-06-27...v2025-07-07

Conjoin refactor, IOSS parallel mapping error improvments, Catalyst2

Conjoin

• Each part mesh can have different variable counts and names
• Add use_all_times option

Catalyst2 update

What's Changed

• Build system can handle paths that contain regex characters (e.g. c++)

Full Changelog: v2025-06-07...v2025-06-27

Ejoin improvements, Exodus.py, io_shell

EJoin

• Much closer functionality to gjoin
• Can merge blocks/sets with same name
• Can merge specified blocks/sets into a specified output block/set
• Nodeset merge -- limit nodes considered during merge operation
• Offset -- can specify per-part
• Scale -- can specify per-part

Exodus.py

• Fix num_blob() function

io_shell

• Option to sort output times to workaround bug in some downstream apps not handling non-sorted times
• Option to randomize output times. For testing apps to make sure they can handle non-sorted times

What's Changed

• EXODUS.PY: Fix num_blob() function by @gsjaardema in #669
• Ejoin combine sets blocks by @gsjaardema in #662
• Ejoin combine specified names by @gsjaardema in #670

Full Changelog: v2025-05-22...v2025-06-07

Aprepro improvements, EJoin nodeset merge, decomp

Improvements to aprepro handling of if/include/loop blocks that have one or more #, $, or // at the beginning of the line followed by whitespace. Previously this text would be output following the block with no newline and then the following text appended to that line which could cause issues due to the unintended appearance of a comment character at the begnining of the line.

Aprepro now strips those leading characters and they are not output.

• Also added {#} as an "aprepro comment" -- ignore everything following on the line. Output '#' and then the remainder of the line with no aprepro processing.
• Added APREPRO(OFF|ON) which ignores all lines in the block.
• EJoin has nodeset merge option
• EJoin has beginnings of combining of sets -- currently if requested, nodesets with same name will be merged into single output nodeset
• Exodiff has option that permuted element connectivity is not seen as a diff.

What's Changed

• Ioss decomp omitted blocks by @gsjaardema in #652
• switches from wget to curl by @andrewcorrigan in #655
• EJOIN: Add nodeset node match option by @gsjaardema in #658

Full Changelog: v2025-05-05...v2025-05-22

Periodic Snapshot -- Change in name handling defaults and others

What's Changed

• ALGEBRA: fix build on newer compilers
• Change defaults for internal variable name case changes.
• better step selection and printing in io_shell and io_info

Full Changelog: v2025-04-29...v2025-05-05

intermediate release for variable case default change

The default for lowercasing input variable names has been changed to leaving the case as it is on input mesh.

What's Changed

• ALGEBRA: See if this fixes type error by @gsjaardema in #648
• Change default setting for lowerCaseVariableNames to false

Full Changelog: v2025-04-29...i2025-05-05

Periodic Snapshot - cpup fixes, numbers, compare, db names

Bug fixes on some frequently used codes.

cpup improvements
io_shell --compare improvements
database name case handling changes.

What's Changed

• Do not lower case database names by @gsjaardema in #642
• Numbers support tet meshes by @gsjaardema in #644

Full Changelog: v2025-04-14...v2025-04-29

[NOTE: Code is the same as the i2025-04-29 release]

i2025-04-29: Intermediate Release

Intermediate release to get changes into sierra

• cpup improvements
• io_shell --compare improvements
• database name case handling changes.

What's Changed

• Do not lower case database names by @gsjaardema in #642
• Numbers support tet meshes by @gsjaardema in #644

Full Changelog: v2025-04-14...i2025-04-29

v2025-04-14

Periodic Snapshot...

Major changes:

• Kluge to support file-per-rank and multiple-rank-single-file read/write in CGNS, other CGNS-related changes.

• Catalyst changes

• Update to latest TriBITs

• Improved static library build

• EPU: Handle case where no elements, but add_processor_id specified

• CPUP: Handle decomp-created zgc between zones better

• IOSS: Add filessystem type function (lustre, gpfs, nfs, ...)

• CPUP: Fix handling of assemblies

• IOSS: fix io_shell compare of db with no changesets

• IOSS: Cgns - handle missing assemblies correctly

• IOSS: Clean up owning_processor calculation

• EXO2MAT: Add -i option to not transfer info records to mat file

• Full Changelog: v2025-03-13...v2025-04-14

Periodic Snapshot - bug fix, new functionality, portability

Some bug fixes and new functionality:

• Fix problem with file-per-rank input assigning incorrect node ownership
• Change aprepro default output format to full precision
  • Added legacy_output_format() function and --legacy_output_format command line format to use legacy format instead.
  • Cleaned up aprepro help/usage output
• Various analyzer recommended changes
• Windows portability fixes.

What's Changed

Full Changelog: v2025-02-27...v2025-03-13