Skip to content

API: Update code to match SIMPLNX MessageHandler API changes - #24

Merged
imikejackson merged 2 commits into
developfrom
topic/message_handler_rewrite_2
Sep 25, 2026
Merged

imikejackson merged 2 commits into
developfrom
topic/message_handler_rewrite_2

Conversation

@imikejackson

Copy link
Copy Markdown
Contributor

No description provided.

imikejackson and others added 2 commits September 24, 2026 09:51
…thms

Brings these filters in line with the simplnx progress-messaging rules.

InterpolateValuesToUnstructuredGrid had three defects. Its thread-safe seam read
the clock itself and the comparison was inverted, returning when a message was
due, so it flooded for one second and then went silent for the rest of the run
while the filter kept working. It reported 100 percent even when the workers had
returned early because the user cancelled. It reported an array as complete
before copying it. The seam now uses the shared throttle, completion is reported
only when the run was not cancelled, and the array count follows the copy. The
worker keeps its count stride but no longer reads the clock, and it flushes its
pending count on the cancellation path instead of discarding it.

MergeColonies, ComputeMicroTextureRegions, ComputeSaltykovSizes and
ComputeLocalAverageCAxisMisalignments ran whole-volume and per-feature passes
with no messages, and in several cases no cancellation check either. Each now
announces its phases and reports its outer work unit. The Saltykov fitting loop
has no honest denominator, so it sends throttled free-form status, while its
quadratic matching phase reports a real percentage.

Two corrections from review of that work. The per-cell loops in MergeColonies
reported and polled cancellation on every cell, which on a large volume is a
billion added reporting paths per pass; they now use a count stride. The final
remap writes a cell parent id and its feature-level parent id together, so
returning part way leaves the two inconsistent with nothing to restore either;
that pass now honours cancellation before it begins and then runs to completion.

Several phases declared N units while iterating a range starting at one, so the
bar could never pass (N-1)/N, and reports were placed before the work they
described rather than after.

Signed-off-by: Joey Kleingers <joey.kleingers@bluequartz.net>
@imikejackson
imikejackson merged commit 0029577 into develop Sep 25, 2026
3 of 6 checks passed
@imikejackson
imikejackson deleted the topic/message_handler_rewrite_2 branch September 25, 2026 11:37
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants