commit 64a49088eef58356f0e188d2093f5ee4ee630eb5

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 23 16:25:15 2026 -0400

Update version number to 7.44.0

commit e74b2fc1ddb0f51e7bd51ffa5c13502584a0fcb9

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 23 16:01:41 2026 -0400

TODO items addressed

commit 9463f18d6b28c0c86d1cfc31d1f6526dd8ece669

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 23 18:12:27 2026 +0000

A few fixes for more obscure build configs

commit 96ec562a9abf1139396e55cb83744e1ca740a535

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 23 12:32:41 2026 -0400

Make a stab at fixing MGED Windows I/O setup

Recent changes introduced a problem on Windows. Based on some work by Chris,

prior state was causing a double close that intermittently corrupted whatever

handle Windows has since recycled into that slot, trashing the stdout stream

when we quit. See if we can set things up to behave correctly.

commit 542aae6500113d3b2f7fcbd548bb3a17f30e4a38

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 23 12:04:26 2026 -0400

Update distclean logic for Windows build

We build with debug enabled, so we produce pdb/exp files in our build tree that

distclean needs to know about if it's actually going to scrub back to a clean

directory.

commit 544e3e2bf59edc9d8e5d0ca79c9112587ff80110

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 23 07:21:36 2026 -0400

Odd pathname hardening for Tcl tests

commit 8178117c349cf455ea55447e9d1da4d4c3e7032a

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Wed Jul 22 23:14:51 2026 -0400

Rename owner var to avoid X11 conflict when not including tk.h

commit 6a9608a95b1e393b422068bc6749da1e7b62ce06

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Wed Jul 22 21:59:22 2026 -0400

Use cmake --build rather than raw ninja

commit a79dcd6f8e6bea4a01bdaba11d0b0f1902b76c8a

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Wed Jul 22 21:47:43 2026 -0400

if we're going to run this on windows we need to match the parent build type

commit 65beec7321e64aa22a02e794c8c47b3a5c16e872

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 23 01:30:38 2026 +0000

list more files for disclean

commit bd3373bb706874fe392c0ce576329a9148e1ff90

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 23 01:08:14 2026 +0000

Make more use of prebuilt bext

commit df2aadf987346dafaecc44ac1ffeadce5ea91fbb

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 16:28:51 2026 -0400

a little more clarity on the rm change since there's now three paths

commit 59fd27654352093f828b936e5f4912282264f784

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 16:18:51 2026 -0400

replace remove with comb rm

commit def9f14b9a5804226754457142beeb005aed869c

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 16:18:16 2026 -0400

remove is no more, replaced with rm and comb rm

commit 57cdf57685d3ccfc2493f0157044e8cc7f7687d8

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 16:17:11 2026 -0400

update the man page with the new comb command usage. three forms that bring it more in alignment with the other object commands.

commit b12ab3d90bcb243f727f6e112e207a70da12a3f1

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 16:11:50 2026 -0400

update rm refs to new call pattern

commit 15e26dc1134ad3dac69771255510eefc5a3642bf

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 16:11:25 2026 -0400

redundant

commit cf4ee1011699a12f7bfc3b5e534f5f48cc0a31b8

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 16:05:44 2026 -0400

update the rm references. actually good examples of the rm obj/mem and comb obj rm mem1 mem2 forms

commit c79316f946e20e757fff6f533f9bfe82a9482b43

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 15:57:11 2026 -0400

fix mged kicking open a browser on the docs

commit 2a8532e078ff5ec295522d906b5a11954add18aa

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 15:55:58 2026 -0400

switch to the right var for rm (and leak oed preview)

commit 1e37d7469963ad7e2c4be83301c8be0c1ccd6ea7

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 15:54:38 2026 -0400

update archer's help also for rm

commit 0debfc5150d805143749db1f0bf221bd96f9217c

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 15:54:07 2026 -0400

update the old tcl help

commit 90015ace2418dbd3b8d93def36a11868e918bae3

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 15:51:37 2026 -0400

update pattern coverage on rm

commit 7f38eee8544b8195c9a72c8977b5b8f02c85f69f

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 15:50:43 2026 -0400

update to the latest calling pattern 'comb obj rm'

commit b45699446c237c218233ebd4568f19688acf5523

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 15:45:31 2026 -0400

update usage to simplify the three common patterns, and de-emphasize all the command switches.

commit 5c68be60db8bd35db3c864690d9125a865ae077c

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Wed Jul 22 12:58:08 2026 -0400

remove command is no more - yank from MGED's setup.c

commit 39becc26f3dac197bca6b275a2dde90dc0db6720

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 11:16:51 2026 -0400

few more rm tests, make sure it works in all three forms

commit 3ece9553eee58c24c0dea8dd25e7649bed5f9d97

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 11:15:59 2026 -0400

instead of hard-coded strings, lets switch to a proper command table since it looks like nearly all the options are actually commands

commit 050f49246ea6f4aca199a8acad329d94c5e94fe9

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Wed Jul 22 10:25:14 2026 -0400

Don't extern C something returning a C++ type

commit 83b3405b9f00d4b260a77c1c7ac71869540e6bbb

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 09:42:15 2026 -0400

reference comb rm instead of remove as alternative

commit d03868d1dacba0cef9f4636935aa22d9c822421d

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 09:41:31 2026 -0400

update tests to make sure comb rm works as intended, even if obj is named 'u'

commit a49321891b1cc68abc1ca5857fe3e035f02978e2

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 09:40:59 2026 -0400

add subcommands to comb including the backwards-compatible 'rm' for deleting member(s) but also wrap, flatten, and lift

commit e4c46de92b924676b2f6d7cc3476733055e32355

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 09:36:18 2026 -0400

remove the remove command, replacing with comb

commit fe88cc6fc7ff76643dfb0185f92b870bba6e2222

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 09:35:11 2026 -0400

taxonomy doc added

commit 1874b7ea6809acd35095799dbaf460f0320b56bd

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 22 01:17:16 2026 -0400

add a taxonomy doc to the repo that combines elements of the quick card ref, but enough to get to port regardless.

commit 0f7b9dc1e5a23716eca879b1b7898fbb3963f016

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Tue Jul 21 11:35:25 2026 -0400

mged manual is fixed (commit incoming)

commit aa0a9e89e7ab485351d3d4967db53c246701512c

Author: f4alt <christopherjmacgregor@gmail.com>

Date: Tue Jul 21 17:16:55 2026 -0400

document cf5c1237f5a39591be89eebcd62c0ea300b0eded

commit cf5c1237f5a39591be89eebcd62c0ea300b0eded

Author: f4alt <christopherjmacgregor@gmail.com>

Date: Tue Jul 21 16:44:46 2026 -0400

fix infinite looping in search

corner case where if we create a child of a matched filter, we'll

recursively create the child.

commit 735ea54b1f2402d1d8b54075d20c510956eab565

Author: f4alt <christopherjmacgregor@gmail.com>

Date: Tue Jul 21 15:32:27 2026 -0400

remove mged async commnads running via polling

Instead, use tcl events. This avoids the 10ms polling penalty

introduced.

NOTE: the SIGINT suppression is not strictly defensive. For instance,

the 'inside' (cmd_ged_inside) and '3ptarb' (cmd_ged_more_wrapper)

handlers arm SIGINT->sig3 via setjmp, then call cmd_draw -> edit_com ->

run_ged_async. A SIGINT during the pump would longjmp to the outer

frame (skipping thread.join() and the dtor - bad).

commit a2faeeb923f548444a873e1450177dffd5cf2256

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Tue Jul 21 11:36:34 2026 -0400

Call out RT_BOT_PLATE_NOCOS bot creation crash fix

commit 57bbca41a9ca349a993808e5236caedafd345f42

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Tue Jul 21 11:31:51 2026 -0400

fix search -exec multithreaded bug where rt/rtcheck/lint style execs would hang mged due to thread contention on Tcl printing.

commit ae941f0698ae7d3df02a9ae44945f55c7a8a19de

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Tue Jul 21 01:30:08 2026 -0400

consolidate all bext updates to the end since they're lessor-impact, and include from/to versioning.

commit f45bda7e9a7e56a8dc1baa7f1b4c254647161fec

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Tue Jul 21 11:27:24 2026 -0400

search -exec behaviors fixed

commit 2f635f88027bf34ada7000783ceba20f201e74bb

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Tue Jul 21 01:20:44 2026 -0400

nix the older bugs section

commit 30b2870d44cda8d1730e84c4ac177220378122cc

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Tue Jul 21 01:20:10 2026 -0400

fixed locally, close enough

commit fab5edd61955a1b834ee3c400166ce8fae224e73

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Tue Jul 21 00:12:24 2026 -0400

bump the deferred

commit ecf04b795999f4d391a5b2df50b1fd83a21ca99c

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Tue Jul 21 00:11:49 2026 -0400

wrote notes, reviewed bullets

commit ff2d2990f976845e6f8c788594fcb34325bc93c4

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 20 23:54:50 2026 -0400

not technically bugs and todo already has an entry to improve oed

commit 4f016ea121f79d8f12c75d3fea3a8630f3b708bb

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 20 23:48:07 2026 -0400

after commit review, found ~15 user-visible changes that were missed including submodel, jt, gcv appends, rt auto-obj, metadata, rt pnts, g2asc of more obj, all the proc-db additions+changes, tire tread, fbserv/rtsrv IPC/TLS, bot exterior, and better area/vol/lint validation

commit 633398b0405c4fe47e89589c29ffe8baa049b55e

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 20 22:52:11 2026 -0400

over 220 new man

commit 410bb7e8ff6e22a2b9cd900b8b973d4149c5660a

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 20 22:31:01 2026 -0400

add some release prose summary highlights for 7.44

commit 2394081300fb0fd9f2b3d5493d6e39a6f2905762

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Mon Jul 20 20:14:02 2026 -0400

Update bext SHA1 to latest

Whoops, should have put that in a branch. bext state should match

the update to latest (working) patch version

commit 00a01dce26c7b9233035906d036cbfda453935d3

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Mon Jul 20 21:07:34 2026 +0000

Stand up a libged "db" command

Well, that was more complicated than expected. The Tcl

db "object" was actually filling a lot of the role of

a db command, so a lot of its logic shifted to the libged

source as pass-throughs to other libged commands.

commit 9f72c86436d8761ac081dc44c8fff68c44ad93b9

Merge: 2de691836a 12edeac6e2

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Mon Jul 20 12:37:48 2026 -0400

Merge pull request #247 from Meghna45-code/fix-bot-nocos-only

libwdb, libged: fix mk_bot crash in RT_BOT_PLATE_NOCOS mode

commit 2de691836a71e0fd404f6fd5a60405931563729e

Author: joshbaker22 <141415098+joshbaker22@users.noreply.github.com>

Date: Mon Jul 20 11:07:16 2026 -0400

Update search.adoc fix table

add some escape characters to table (pipes were screwing things up)

commit 1e0b74a63036a7924eea66ac92574d6a9ff317a5

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Mon Jul 20 09:45:40 2026 -0400

Back out the bext ap214 stepcode changes for now.

commit 12edeac6e20ff725e6aab9fe6a2de1f7c5e5b482

Author: Meghna45-code <vidhims111@gmail.com>

Date: Mon Jul 20 17:32:15 2026 +0530

libwdb, libged: fix mk_bot crash in RT_BOT_PLATE_NOCOS mode

commit 34e2546b77901d209e5f98bdfcb7613c518d0bae

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Mon Jul 20 07:56:22 2026 -0400

Need to change how we're validating a bext checkout

commit ccda8b67ade93dae5407ea608cb4f7e8eeaa7d99

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Mon Jul 20 00:33:31 2026 -0400

Unexpectedly, updating shannonbooth patch (apparently) caused breakage running on the CI - back it up for now.

commit e366225907d542917e4b6da3616a3094090584d2

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Mon Jul 20 00:20:57 2026 -0400

fix for stepcode patches

commit 0e76d9b4d9359fc9248b831f31bffbdb8f46a5da

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Sun Jul 19 21:07:10 2026 -0400

Bump to latest bext sha1

commit 7ef25d649210eda702ece4c1c3695e82e73ebb5b

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Sun Jul 19 20:02:24 2026 -0400

Bump to latest bext SHA1

commit f1a465c45d92b16128f8cdd9fc96cf7ee4d614f1

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Sun Jul 19 16:56:17 2026 -0400

Make sure we're ready for group process termination

Timeout termination raced with child process-group creation and

returned before reaping the killed child. We now establish

groups from both parent and child.

commit 1634bc6b801e2db82a4144c7357ec980c360bc0d

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Sat Jul 18 21:23:04 2026 -0400

More Windows robustness work

commit 42b0cc824545b08fc953ba8b8d0488e1948f261f

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Sat Jul 18 15:38:00 2026 -0400

regress-remrt was still hanging sometimes on Windows

commit 29c14c3006407c2d84fe75328ec6fb9c55d823ae

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Sat Jul 18 13:50:21 2026 -0400

Need to stop subprocesses before resources get taken away.

commit d53f005a897e355d86e3609b110f9c5faae380a0

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Fri Jul 17 17:02:14 2026 -0400

Add a Windows backend for bu_backtrace

commit fee80b492261d7c8a1b4588cb3152885f90b319b

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Fri Jul 17 16:18:27 2026 -0400

Don't bu_bomb when log fails

Logging failure is a function of enviornment (observed with qged during

shutdown - hooks are unregistered and no stdout/stderr available.)

commit 58365e34539f74edd978816fd73b34e36ba56487

Author: f4alt <christopherjmacgregor@gmail.com>

Date: Fri Jul 17 14:00:39 2026 -0400

fix helpcomm path sourcing

need to guard the source with an existance check first. Depending on

ordering during the build, [bu_dir data] can/does return empty "" - which

collapses the path to a relative tclscripts/helpcomm.tcl. Instead,

verify the file was found (bu_dir succeeded) before sourcing; if it's

not, this keeps the file sourceable at index time where we're guaranteed

to resolve [bu_dir data]

commit 42322a692a3e4fa8c4b530c34b88ba9312dc991f

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Fri Jul 17 10:29:00 2026 -0400

Make asc2g faster for large input reads

commit 795fafb9bb42349cb7d94aab47fbe3a7a40ef7a8

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Fri Jul 17 08:39:03 2026 -0400

Our teardown process for MGED had some issues (a few of which were libbu/libged level.) Make a stab at cleanup.

commit 8fe203adb08ecdde023503c68dfc17c0447e9ca0

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Fri Jul 17 12:12:14 2026 +0000

Put back the other view subcommands

commit 5951e3363693c1a13bf556c9d7c4d2c438220849

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Fri Jul 17 07:04:53 2026 -0400

List file for CMake

commit a48e2027eb1753da743fbee35ac17fe3586a101c

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Fri Jul 17 01:25:24 2026 -0400

Fix regression tests

Looks like list is a no-go for MGED, and concat and glob have similar

theoretical problems.

commit 0bebd04d7e24f60ce8e8daf143f52d29f4f6d7c9

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 16 23:10:14 2026 -0400

Try some MGED memory cleanups

MGED valgrind exits are usually noisy - try a few cleanups.

Part of the noise is due to how Tcl is normally built. Finalization

removes allocator roots without returning the arenas, causing Valgrind

to reclassify memory from “reachable” to “possibly lost.” Tcl documents

that freeing all memory for leak-checking requires TCL_FINALIZE_ON_EXIT;

it is not the normal allocator behavior. Tcl finalization documentation

(https://www.tcl-lang.org/man/tcl9.0/TclLib/Exit.html), Tcl source

discussion (https://core.tcl-lang.org/tcl/wiki?name=%28Initi%2FFin%29alize).

commit e90ef9cd9622cebf349d36a3eb556ec5569a46c2

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 16 23:08:40 2026 -0400

Make a note that the mapping[MAX_PSW*MAX_PSW] array is what's triggering valgrind large range warnings when it loads

commit a71ddcd0a8106d52887b4cf8311b5d587e3ba67a

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 16 22:34:59 2026 -0400

Correct and optimize search hierarchy evaluation

Tighten up when we can use the optimized -above/-below paths - some

of them were problematic. Any -exec specifier currently disables

the fast paths because the database could be altered and invalidate

the caches.

Derive -maxdepth pruning only where the boolean plan guarantees a

traversal bound, and fall back when depth constraints or -exec prevent

safe reuse.

Other changes:

Preserve DFS callback ordering and cyclic path results, fix relative

-below depths, nested hierarchy parsing, regex compilation, unique

result collection, -exec cleanup, and several predicate error paths.

Batch libged local roots, preserve plan arguments exactly, propagate

search errors correctly, and expand librt, libged, and MGED regression

coverage.

Document hierarchy depth semantics and plan forms that may require

quadratic evaluation.

commit bc97bf3e211d85dc2f6941273ba4685a74545e0a

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 16 21:03:44 2026 -0400

Use the sub-interp only for -exec searches

Since there is a fixed cost to search Tcl interp setup, see if we can

avoid paying it if we don't really need it.

In principle, for non -exec searches, the async thread running should

suffice for what we need.

commit c4ac4d37d5c941ecdf27e91bd38d05500bf28d24

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Thu Jul 16 20:36:34 2026 -0400

commit some temp working notes to pick up where left off

commit fe21ee4e2ba441efa9d708b1b313c4a3dfefde54

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Thu Jul 16 20:34:52 2026 -0400

add the 53 missing ged commands that 'should' just work through the plain wrapper. also added 21 of the approx 30-40 other commands missing that are essentially aliases (copy -> cp). left the rest that affected the display (zap, clear, etc).

commit fa36187f1c28119eec9bff55c7c217da15e33018

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 16 20:12:47 2026 -0400

search performance optimizations, sorting optimizations

Looks like 53563dcc39 was incomplete (either not done or I missed saving

some changes at some point, not sure which - either way needs fixing.)

- src/librt/search.cpp:506 now carries match state parent-to-child,

reducing the common case from O(N·depth) to O(N).

- Mixed -above/-below searches receive the same optimization.

- Depth-constrained and side-effecting -exec/-print expressions retain

the correct ancestor-walk fallback.

While we're in here, fix a second hot path - we now render every path

once into pre-sized shared storage for final sorting. Comparisons no

longer reconstruct paths or allocate strings, which showed up in

profiling for larger outputs.

Compared to the previous 7.42.2 release, in MGED we are setting up,

syncing and tearing down a separate Tcl interp per command run, as well

as launching search in a separate thread and managing I/O between the

threads. The overhead is real, but seems to be in the

less-than-a-second range (depending on platform.) For very large

numbers of search calls in a script that may add up to an observable

runtime slowdown.

commit 08e39fbee001b255b1d0c1382860faaa75679499

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 16 16:30:25 2026 -0400

Make sure search's Tcl is shut down win or lose

The Tcl Inteprp used by search is created by the thread that runs it async, so

it must be shut down by that thread. Leaving it up and a pointer to it in

search_interp resulted in a crash on Windows when a failed command didn't

clean it up and a subsequent initialization tried to clear it out. Tcl prints

a message about an async delete from the wrong thread and wipes out.

commit bd53790f691cfd9bf3bac07e4eb7324c8535dd63

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Wed Jul 15 17:29:07 2026 -0400

final release actions

commit 26777a18325e8995919d9a84989b6037578c9957

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 16 00:37:16 2026 -0400

More qged work

commit 67d7fa2d392f25b09a539c18e494b318c33b8bef

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Thu Jul 16 00:20:18 2026 -0400

Make a stab at fixing/improving some QGED behaviors

commit 091ed7f9c07ccc9310d5acf8ad10ebd6a32c65e5

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Wed Jul 15 13:06:12 2026 -0400

New bext SHA1 - get latest FreeBSD updates and fixes

commit 07b9fe1573f5a0168ef7afc9a1fb845eae5172c5

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Wed Jul 15 12:26:02 2026 -0400

Add the find_package_zlib for the gcv jt plugin

commit e51b4817a95e6823a656314615d73dbc8290ea3d

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Wed Jul 15 14:33:03 2026 +0000

Make MGED aliases available without Tk

Provide the alias and unalias Tcl helpers independently of Tkcon and load

them during MGED's base interpreter setup. Aliases defined in .mgedrc now

work in classic and headless MGED sessions without attempting to initialize

Tk or requiring an available display.

Have Tkcon use the shared implementation so its existing alias behavior is

preserved without duplicating the helper procedures. This also allows user

aliases to be captured and replayed cleanly in the Tcl interpreter used by

search -exec.

Expand the search -exec regression coverage to exercise the public alias

command, global and namespaced aliases, aliases targeting Tcl procs and GED

commands, fixed and dynamic arguments, and worker-side alias queries.

commit 937775803223c15904a97590f105a935e7e7989c

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Wed Jul 15 10:04:45 2026 -0400

Pull in the latest FindZLIB.cmake from bext

Will need to see if BRL-CAD's usage needs updating, but as a rule we need to

keep our Find modules in sync between bext and brlcad so find_package calls in

each do the same thing. bext has the more modern one, so we should be using

it.

commit b4d072dd511cb163f119a304799c7d2aca69fb4b

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Wed Jul 15 07:01:07 2026 -0400

For brledit print usage if -h or -? is supplied.

commit f81a81e72c88a75b39aacb5b4451943f757a34d0

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Tue Jul 14 22:52:05 2026 -0400

Fix header

commit 1bf5eb9d6b6eae12dd1414d4cd499ce3de6e84e9

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Tue Jul 14 18:36:33 2026 -0400

Add/restore view dir subcommand

commit d25579a6ba2f2a0b768cb3ba608dec46e07031eb

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Tue Jul 14 17:19:46 2026 -0400

Robustness fixes for non-GUI MGED I/O

commit 24a4603b39 set classic MGED stdin nonblocking. With a PTY,

stdin/stdout/stderr can share status flags, so large ordinary search

output hit EAGAIN; bu_log treated it as fatal and bombed. This is

unrelated to search -exec or the secondary Tcl interpreter.

Changes:

- Classic cbreak input now stays blocking, reads one guaranteed-ready

byte, then drains only Tcl-buffered bytes— avoiding both event-loop

stalls and paste inefficiency. Non-cbreak pipe input remains

nonblocking. src/mged/mged.c:1763

- bu_log/bu_putchar now use raw standard-stream writes, retrying EINTR,

EAGAIN, and EWOULDBLOCK under the existing serialization lock. It uses

write on POSIX and _write on Windows. GUI hooks and threaded

search -exec handling are unchanged. src/libbu/log.c:65

- Added a nonblocking-output regression test that verifies a 128 KiB log

message is complete and uncorrupted. src/libbu/tests/test_log.cpp:142

commit 515819a7f5ef2049721693762ad541ca1c1bd9dd

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Tue Jul 14 18:44:22 2026 +0000

Try to make MGED bu_log behaviors more robust

- src/libbu/hook.c:71 moved one hook too many, causing

out-of-bounds memory access. A delayed, platform-dependent

failure in a later bu_log is consistent with this.

- The global log hook list was mutated without synchronization.

It now uses a dedicated semaphore, with a thread-local

recursion guard in src/libbu/log.c:62.

- An invalid output_hook replacement removed the functioning

MGED hook before validation. src/mged/cmd.cpp:1427 now preserves it.

- Worker output is drained after join(), ensuring no output can

arrive after the final drain.

- GUI stream redirection now uses dup2() rather than assuming

close() followed by dup() reuses descriptors in src/mged/mged.c:3219.

- Pipe data is passed to Tcl using Tcl_DStringAppendElement, handling

braces and other Tcl-sensitive output safely.

Added concurrent hook mutation, parallel logging, recursion, hook deletion,

and invalid MGED output-hook tests. The sourced-proc search regression now

exercises the preserved-hook path.

commit f110978fe8d75a5f2e57e7c507f8196d637dcf61

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Tue Jul 14 13:56:48 2026 -0400

Update to Qt 6.11.1

commit 5d596a2ddfc422d6495ef4b580d6930f2c2e5a64

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Tue Jul 14 15:47:07 2026 +0000

Remove completed TODO item

commit 0308de3f25e27cadbe5afcbab63f146f02cdb63d

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Tue Jul 14 14:48:44 2026 +0000

Trying to make the search -exec Tcl sync more robust

commit 1e851ac6ccdb5872b46938e1b0864c25472b0c69

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Tue Jul 14 10:43:52 2026 -0400

Use older strclear/plief until the main CMake updates drop

commit 488afd00920006923a337daeb97bee94e6e7ae0c

Author: joshbaker22 <141415098+joshbaker22@users.noreply.github.com>

Date: Tue Jul 14 10:19:23 2026 -0400

Update search.adoc Add examples to mimic dbfind and dbfindtree

Add 2 examples that specifically mimic behavior or soon to be removed commands dbfind and dbfindtree.

commit 3d6c1635a62c71f63d3f1226760982c0d62fa62a

Author: Clifford Yapp <238416+starseeker@users.noreply.github.com>

Date: Mon Jul 13 10:40:48 2026 -0400

Various cleanup details related to recent logic changes

Haven't tried distcheck itself yet, but distclean should do better now.

commit 8d9f87202ee92e00b99537974b9a0e268f55d8fc

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:45:51 2026 -0400

one char nudge for symmetry and so we're not under the logo

commit 47d6480afa18869bf161f869468d6d324973ff80

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:38:57 2026 -0400

enable existing boolweave test even if it's inadequate

commit e315010cce85f120c4139a86ef79377dbbadd677

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:37:58 2026 -0400

oops, add main change providing mirror support to brep objects

commit 5f3b8397fc68c894b71e5e5f6fc7e27e26229d63

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:36:19 2026 -0400

improve rendering performance with lights, minimizing malloc/free calls a whole lot more (128 -> 8192). timing issue exposed in new rt heat map outputs.

commit 1d306bde6b7786cfcf3e98970bec52433995b13a

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:32:50 2026 -0400

gist pages now include the file size

commit b09e44a84710c2ebdcfbc822f55feae25800c560

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:28:53 2026 -0400

add a test for icv's file handling

commit 22cfd4833becdfd342b61d386a88e0a934d1ba3a

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:22:50 2026 -0400

added ability to mirror brep, minimally tested but looking good.

commit e442c6be476d5666c8372c60f405227f83edbe71

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:20:09 2026 -0400

related to the mirror bug, need to flip some transformed comb leaves

commit ab1e29ec806d1f72dbc6a11fcfefd60daf80e17f

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:18:22 2026 -0400

fixed corruption of rhc entity upgrading. was a 1-char typo causing it to read the wrong field. technically was likely a v4 import failure on any RHC.

commit 118e45cdda01fab46bac9ef447ae5c08cc699748

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:14:24 2026 -0400

fixed a really old bug in archer if you tried to open a database named the same as a command. issue was an ' ls' call.

commit 329d436946c9efebfe243a7d6c3908820c223687

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:10:48 2026 -0400

more keyword search guts

commit 4c5e4f8eb60b7aa92c6294d0ff357f1c541156f2

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:10:17 2026 -0400

fix a bug in file tracking (typo)

commit 70c3824d85342f64451f6e9b09cccf08a30eb623

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:07:36 2026 -0400

add search to the man browser gui too.

commit 469337f06df1230d772c726a1d8dd290360065bf

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:04:57 2026 -0400

if they want keyword searching, route it a little differently

commit b4b907949cb7e9502dc2611ec1a3c684bfba5230

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:03:55 2026 -0400

more keyword searching guts and passthrough

commit ebf81d3744447484fc5413b7b89f23da56484c9f

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:03:11 2026 -0400

oops, add the new helper proc for searching man pages

commit f15efb022828a922664dc4a7e5bfb36c878508ae

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:02:08 2026 -0400

add searching of the manual pages to the mged GUI

commit 3d07ce6e970b44836ff549a1ad86614185ad7cd7

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 04:01:06 2026 -0400

fixed an mged ancient bug where pasted text would sometime get captured underneath the yellow cursor box

commit 29f862276c13985d9b7fee0a8805ddd59274d519

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:50:17 2026 -0400

baseline affirmation, jt mesh successfully imported

commit f739f74d487b41e921a359e531b8cabd3e34a1a5

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:47:03 2026 -0400

add full text keyword searching of manual pages.

added to brlman, man, and apropos, with -k and -K options that saerch synopsis+options+description and entire manpage respectively.

commit 93d3ed01e2c295a118ee75b517b93c2c5e8d69ef

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:41:36 2026 -0400

add full text keyword searching of manual pages

commit 2f454bf4c7ae6694740c054e9d4ef8ab1ac565de

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:39:26 2026 -0400

make sure we can v4 export rhc

commit 982e9abe5d3e9480aeb95551f808ba32b7951963

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:38:20 2026 -0400

make sure sphere case works

commit 13925659548295716c714a6be0048c4398688098

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:36:53 2026 -0400

add jt file format

commit 6a1f0ab73d8c1ab266026f5f0783144a98b8aca5

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:36:23 2026 -0400

KISS. lets not pollute API w/ inconsistently long names just because of some minor detail on the clean one that got their first. just icv_create().

others introduced at some point have the same problem, but they can get cleaned later. immediate issue is not halting our own builds because of deprecation labels. removed deprecation labels as unnecessary (MI, icv isn't published/documented) and actively messy.

commit c63ef3688513dd3fe02340391e0656c3a79123e8

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:26:50 2026 -0400

fix indent

commit aedf574e5558fac6a67b61102d2ae470da1903a6

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:07:50 2026 -0400

todo fulfilled, but migrate the common cases to the manpage

commit 71c6e324e005b5e2c6bf1ce0349257d5b97ac60c

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:06:02 2026 -0400

add a deeper unit test that captures much of the new behavior including the safety exclusions

commit e1749fa9e755e833d038d2fbb957029e7be2cbd9

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:04:40 2026 -0400

rm has a safeguard

commit bd84801d5e949172f135db98d83e37cefb150ddc

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 03:00:30 2026 -0400

ged rm is a big deviation from the old behavior to the point of needing multiple protections, but ultimately landing more secure and expected for those familiar with linux/posix/unix/bsd/etc. it now kills but only when not externally referenced (unless forcing). see manpage for details.

commit 3cd7cf7c7d871cd8d0c61debb1c41b6b871e9e9b

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 02:52:24 2026 -0400

update rm command with the documented changes.

now rm is effectively kill or killtree but also still sometimes just removing a recipe instance. to get the prior 'rm comb foo' one must insert the slash: 'rm comb/foo'. if someone new really does want to kill comb and foo, they'll have to force it rm -f. if a comb or member are referenced elsewhere in the db, rm aborts (unless it's forced). only controversial dec was to not allow rm -rf to kill more than the comb specified when referenced externally, but not the underlying children also -- justification is they weren't named, so we're doing the safe action especially considering they could literally be any assembly easily (intentinoally or not).

commit 9ce51a97f96e4ea05da8dd28ffee4892836d76ff

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 02:40:31 2026 -0400

no longer a bug, but still needs more work to go faster.

commit ab6523e871123b25fc354187bcd372f421f9f1ee

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 02:37:02 2026 -0400

fixed a bug in gqa where perfectly symmetric objects (like a simple sphere) reported wrong volume.

gqa was terminating iterative volume estimation early because all three ortho views are in perfect agreement no matter the resolution. Now it checks against the previous volume to determine whether we were also in agreement with the previous iteration.

commit 20c86177d1a3ab2e094fed9b04ddd080a19838c5

Author: Christopher Sean Morrison <brlcad@gmail.com>

Date: Mon Jul 13 02:29:27 2026 -0400

verified an old infinite loop bug was not actually inf, just wildly quadratic time.

this partially improves the situation though still not performant beyond about 200k cloned entities. the test case originally documented is 1M and still takes several minutes:

make sph sph

clone -i 1 -t 1000 0 0 -n 100 sph

kill sph

g line sph*

clone -i 1 -t 0 1000 0 -n 100 line

kill line

g square line*