[us-commits] [ehb54/ultrascan3] 22f5fb: somo/grpy: build the module under C++11, and test ...

[emre brookes]

  Branch: refs/heads/ehb54-issue-1016-somo-merge-2
  Home:   https://github.com/ehb54/ultrascan3
  Commit: 22f5fbbfeccad48466fb20aa2b3af2b5b9cb9a1b
      https://github.com/ehb54/ultrascan3/commit/22f5fbbfeccad48466fb20aa2b3af2b5b9cb9a1b
  Author: ehb54 <brookes at uthscsa.edu>
  Date:   2026-08-18 (Tue, 18 Aug 2026)

  Changed paths:
    M us_somo/develop/grpy/grpy_shell.hpp
    M us_somo/develop/grpy/tests/run.sh

  Log Message:
  -----------
  somo/grpy: build the module under C++11, and test it at that standard

SOMO supports Qt5 as well as Qt6, so it has to build under the older
compilers Qt5 is used with, and nothing in its build raises the language
standard. The grpy module did not honour that:

  grpy/grpy_shell.hpp: error: no viable overloaded '='
  note: cannot convert initializer list argument to
        'grpy::ShellReport::Provenance'

ShellReport::Provenance carries default member initializers, which makes it a
non-aggregate before C++14, so the braced assignment

  rep.prov[ m ] = { ri.ok, ri.clamped_low, ... };

is not valid C++11. It is now assigned field by field. The initializers stay:
prov is filled by assign() and they are what make an untouched entry read
false/nullptr rather than indeterminate.

That was the only C++11 violation in the module -- grpy_shell, grpy_exposure,
grpy_types and grpy_process all compile clean at -std=gnu++11, as do both
test programs.

The reason it survived is the second half of this commit. grpy/tests/run.sh
compiled the module at -std=c++17, a HIGHER standard than the product it
ships in, so every test passed on code the Qt5 build would not accept.
Testing above the product's standard hides breakage instead of finding it.
run.sh now uses -std=gnu++11, to be raised when SOMO's own standard moves --
which is a question for the Qt6 migration, not for this module.

Compilers new enough to default to gnu++17 (the CI containers) built it
either way, which is why CI stayed green; the Qt5 macOS build did not.

Co-Authored-By: Claude Opus 5 <noreply at anthropic.com>


  Commit: d37eeb1678b8a4eba616a47c465ea4ce743ddc08
      https://github.com/ehb54/ultrascan3/commit/d37eeb1678b8a4eba616a47c465ea4ce743ddc08
  Author: ehb54 <brookes at uthscsa.edu>
  Date:   2026-08-18 (Tue, 18 Aug 2026)

  Changed paths:
    M us_somo/develop/src/us_hydrodyn_residue_builder.cpp

  Log Message:
  -----------
  somo/perceive: C++11 aggregate init in the residue builder

Same defect as the grpy one, in code that arrived with the perceiver:

  src/us_hydrodyn_residue_builder.cpp:47: error: no matching member
  function for call to 'push_back'
  note: cannot convert initializer list argument to
        'somo_volume::Sphere'

somo_volume::Sphere declares `double x = 0, y = 0, z = 0, r = 0;`, so it is
not an aggregate before C++14 and cannot be built from a braced list under
the C++11 that the Qt5 builds use. Filled field by field instead; the
initializers stay, since they are what make a default Sphere read zero.

This one is NOT new to this merge set -- it came in with the perceiver and is
already on main. It stayed hidden because the grpy error above aborted the
compile first, and because every CI container defaults to gnu++17.

With this, the whole library compiles at -std=gnu++11: 0 errors over the full
source tree, checked with make -k so nothing was masked by an early abort.

Co-Authored-By: Claude Opus 5 <noreply at anthropic.com>


Compare: https://github.com/ehb54/ultrascan3/compare/735d4f94f2cd...d37eeb1678b8

To unsubscribe from these emails, change your notification settings at https://github.com/ehb54/ultrascan3/settings/notifications