Untouched open Tk bug tickets: Not a bug / RFE

Triage of 548 open Bug tickets: those with no comment or check-in from me on 2026-09-24, plus 38 added on 2026-10-04 (filed since then, Pending then, or reopened; 16 of them filed by me), plus 68 added on 2026-10-22 (all other open or Pending Bug tickets not filed by me: the 30 created before 8.5, 31 I had already worked on, and 7 filed or missed since 10-04). Reproduced on Linux/X11 under Xvfb (main 9.1 tip, 9.0, 8.7, 8.6 tip, 8.5, 8.4) and on a Windows 11 VM (9.1, 8.6, 8.5, 8.4). Rows marked patch link to a prototype patch (150 in all, in tk-triage/prototypes). Updated 2026-11-11 with what happened since: the emoji before each ticket id shows its progress. On 2026-10-24 every prototype for a still-open ticket was verified again on current main, on Linux and Windows, and given a test where practical; on 2026-10-25 seven more reproduced tickets got prototypes.

Not a bug / RFE 60

💬c0fb24b74e2026Windows binary build of wish 9.1 recognized as Win32/Wacatac.c!mlNOT-A-BUG (antivirus false positive in Microsoft Defender signatures. It hit only gcc-built static binaries, and obermeier reports that all VirusTo…

Open ticket c0fb24b74e · 85. Win Build · filed 2026-10-03

💬 Since triage (updated 2026-11-11)

Status now: Open

  • 💬 2026-10-03 anonymous: Can confirm with VirusTotal and my BAWT builds, that tcl91.dll and wish91.exe (static build) gives 1 error (from Microsoft) out of 70 scanners, when using a gcc 15.2 build. Using VS2022 builds, no errors are issued on VirusTotal. Paul
  • 💬 2026-10-17 obermeier: Gave it another try today on VirusTotal. Now all checks pass. Could you give your version another try using updated virus detection files?
Full verdict

NOT-A-BUG (antivirus false positive in Microsoft Defender signatures. It hit only gcc-built static binaries, and obermeier reports that all VirusTotal checks pass as of 2026-10-17.)

  • Platform Windows (binary distribution)
  • Tested nothing to test on Linux.
  • Cause heuristic signature in Defender (1.459.536.0), not anything in Tk.
  • Prototype none.
  • Suggested ticket comment This was a false positive in a Microsoft Defender signature. With current signatures all VirusTotal scanners pass (see 2026-10-17 comment). False positives can be reported to Microsoft at https://www.microsoft.com/wdsi/filesubmission. Closing; please reopen if it comes back with current signatures.
⬜37c6fcf92a2026print.tcl: _print_widget on Windows — isotropic mapping uses only one axis, distorts canvas aspect ratioNOT-A-BUG (NOT-REPRODUCED on 9.1 main with Microsoft Print to PDF)

Open ticket 37c6fcf92a · (unused) · filed 2026-06-24

Full verdict

NOT-A-BUG (NOT-REPRODUCED on 9.1 main with Microsoft Print to PDF)

  • Platform Windows
  • Tested code reading only
  • Cause library/print.tcl:_print_widget calls _gdi map -logical $lo -physical $ph with single values. GdiMap (win/tkWinGDI.c) then selects MM_ISOTROPIC with equal window and viewport extents on both axes. MM_ISOTROPIC keeps one logical unit the same physical size on both axes, so the aspect ratio should already be preserved. Using two values would switch to MM_ANISOTROPIC and could itself distort output on printers with resx ≠ resy.
  • Prototype none (the proposed change is not obviously right). Repro: win/37c6fcf92a.tcl (wraps ::tk::print::_gdi, prints a wide canvas with a circle and a square, computes the inch-per-pixel ratio on both axes from the mapping actually set, and prints a verdict)
  • Suggested ticket comment The single-value _gdi map selects MM_ISOTROPIC, which keeps the scale equal on both axes. Which printer or driver shows the stretching, and what are its x and y resolutions? A two-value mapping would switch to MM_ANISOTROPIC.
Windows VM, run 1
  • Platform Windows
  • Verdict NOT-A-BUG (NOT-REPRODUCED on 9.1 main with Microsoft Print to PDF)
  • Tested 9.1 main, win/37c6fcf92a.tcl, "Microsoft Print to PDF" (dialogs driven by work-win/printdlg.ps1). Mapping: (4155, 4155) -> (500, 500), MM_ISOTROPIC, 600x600 dpi; inches per canvas pixel x = y = 0.013850 (ratio 1.0000). In the rendered PDF the circle is round and the square is square. 8.6 has no tk print.
  • Cause none. MM_ISOTROPIC already keeps the axes equal.

Side finding: _print_canvas.rectangle and .oval replace an empty -fill with the printer background, so unfilled items print filled white and hide the items under them (the oval in the repro is covered by the square). Also, the switch in _print_canvas.TransColor uses a braced pattern, so $vtgPrint(canvas.bg) is never substituted.

  • Prototype none
  • Suggested ticket comment With "Microsoft Print to PDF" (600x600 dpi), the single-value _gdi map gives MM_ISOTROPIC with equal scale on both axes, and circles print round. Which printer shows the distortion, and what are its x and y resolutions? Note that dpi_x is taken from dmPrintQuality, which may not be the real X resolution on some drivers.
⬜682ff8f6372026print.tcl: _print_canvas — unqualified info commands may skip canvas item handlersNOT-A-BUG (the handlers are resolved correctly)

Open ticket 682ff8f637 · (unused) · filed 2026-06-24

Full verdict

NOT-A-BUG (the handlers are resolved correctly)

  • Platform Windows (library/print.tcl, Windows-only code path) — logic testable anywhere
  • Tested info commands _pc.$type called from a proc in a child namespace, with and without a same-named global command, on tclsh 9.1 (main) and 8.6: it always returns the unqualified name, so the eq test is true and the handler runs.
  • Cause none. info commands with an unqualified pattern searches the current namespace (::tk::print), which is where _print_canvas.<type> lives. The only theoretical problem is an item type name containing glob characters. Minor cleanup if wanted: if {[namespace which _print_canvas.$type] ne ""}. The call also passes $printargs(hDC) where the $hdc argument is available; this is harmless.
  • Prototype none (no bug)
  • Suggested ticket comment info commands with an unqualified pattern resolves in the current namespace (::tk::print), so _print_canvas.$type is found and the check works; checked on 9.1 and 8.6. Closing as not a bug unless someone has a canvas item that is actually skipped.
⬜c2f501a5a02025[SOLVED] Tk apps get progressively slower on Debian 13 running X11NOT-A-BUG (in Tk). Per the reporter, the _server%d_%d/_client%d atoms and the slowdown come from the ibus XIM server (IMdkit, ibus issue 2924).…

Open ticket c2f501a5a0 · 67. Unix Window Operations · filed 2025-11-01

Full verdict

NOT-A-BUG (in Tk). Per the reporter, the _server%d_%d/_client%d atoms and the slowdown come from the ibus XIM server (IMdkit, ibus issue 2924). Workaround: XMODIFIERS= wish.

  • Platform X11
  • Tested Xvfb :124 without an IM server, 5 runs each of main and 8.6 with echo exit | wish: the atom count stays at 232, so Tk itself does not leak atoms.
  • Cause ibus XIM IMdkit (external)
  • Prototype none (not a Tk bug)
  • Suggested ticket comment Tk does not create these atoms itself. They come from the ibus XIM server, see https://github.com/ibus/ibus/issues/2924. The workaround is to start wish with an empty XMODIFIERS. Closing as third-party issue.
⬜97d9ac3eb12025Adopt Tcl 9 interfacesNOT-A-BUG (refactoring tracker/RFE). Branch dgp-adopt-tcl9 has 2 check-ins (980977e8c6 Tcl_FreeInternalRep, 4cad747bae Tcl_GetNumberFromObj; 2025-0…

Open ticket 97d9ac3eb1 · None · filed 2025-09-03

Full verdict

NOT-A-BUG (refactoring tracker/RFE). Branch dgp-adopt-tcl9 has 2 check-ins (980977e8c6 Tcl_FreeInternalRep, 4cad747bae Tcl_GetNumberFromObj; 2025-09-03). Neither is merged; main still uses neither API.

  • Platform generic
  • Tested n/a
  • Cause n/a
  • Prototype none (the work is in branch dgp-adopt-tcl9, which needs review and a merge)
  • Suggested ticket comment Branch [dgp-adopt-tcl9] (2 check-ins) is still unmerged. Needs review before merging to core-9-0-branch/main, or the ticket could be retyped as RFE.
⬜beaa8ecd002024Inconsistent reporting of child geometry changes to grid containerNOT-A-BUG (the events in the "width & height" case are synthetic, sent on purpose; <Configure> reports the container's own geometry, not its childr…

Open ticket beaa8ecd00 · (unused) · filed 2024-08-30

Full verdict

NOT-A-BUG (the events in the "width & height" case are synthetic, sent on purpose; <Configure> reports the container's own geometry, not its children's)

  • Platform generic
  • Tested main, 8.6 tip, 8.6.9, 8.5 (Xvfb, no WM). main/8.6: counts 2/1/1/2 as reported. 8.6.9 and 8.5: last case 0, which is consistent.
  • Cause generic/tkPlace.c:PlaceRequestProc. When place fixes both the width and the height of the content, its geometry request is rejected and TkDoConfigureNotify() sends a synthetic ConfigureNotify with the size unchanged ("to indicate that the size change request was rejected"). This was added in d5989e7ea9 (2019-02-07, Fix 58665b91dd, unixEmbed tests, 8.6.10). If only one dimension is fixed, the placer resizes the other one, and a real Configure comes only when the size actually changes. So the extra events are the odd case, not the missing ones. gdb shows no X configure call on the container in that case.
  • Prototype none. For the reporter's use case, bind <Configure> on the children or check winfo reqwidth/reqheight of the container. Removing the synthetic event would mean re-checking the unixEmbed tests.
  • Suggested ticket comment <Configure> on the container reports only changes to its own geometry. The extra events when both width and height (or relwidth/relheight) are fixed are synthetic events. PlaceRequestProc has sent them since 8.6.10 (d5989e7ea9) to signal a rejected size request. Bind <Configure> on the content windows instead.
💬4be479d0852023tkUnixRFont.c: nothing defines XFT_HAS_FIXED_ROTATED_PLACEMENTpatchNOT-A-BUG (dead code, still present in main, 9.0 and 8.6)

Open ticket 4be479d085 · 46. Unix Fonts · filed 2023-11-23

💬 Since triage (updated 2026-11-11)

Status now: Open

  • 💬 2026-11-07 chrstphrchvz: Not asking for something to be done about this ticket anytime soon. Just pointing out the comment above TkDrawAngledChars(), which I finally bothered to read and probably explains the macro: > This would be simple code, except Xft has bugs with cumulative errors in character positioning which are caused by trying to perform all calculations internally with integers. So we have to do the work ourselves with float…
Full verdict

NOT-A-BUG (dead code, still present in main, 9.0 and 8.6)

  • Platform X11 (code cleanup)
  • Tested grep over main, 9.0 and 8.6: the macro is only tested in TkDrawAngledChars and nothing defines it. main builds without it, and font.test, canvText.test and unixFont.test pass after removing the dead branch.
  • Cause unix/tkUnixRFont.c:TkDrawAngledChars has an #ifdef XFT_HAS_FIXED_ROTATED_PLACEMENT branch from the TIP 119 patch (a2e77573). The branch was never compiled.
  • Prototype 4be479d085.patch (118 lines, deletions only, no test). It removes the dead #ifdef branch and keeps the #else code that is actually used.
  • Suggested ticket comment The macro was never defined anywhere, so only the <tt>#else</tt> branch has ever been compiled. Proposed: remove the dead branch (pure deletion).
Prototype verified 2026-10-24, Linux
  • Status VERIFIED (dead-code removal)
  • Applies to main clean (offset 9 lines)
  • Repro before: XFT_HAS_FIXED_ROTATED_PLACEMENT is used only in unix/tkUnixRFont.c and defined nowhere / after: gone, the #else code is kept and it builds without warnings
  • Test none (build-only cleanup, no behaviour change)
  • Test file run font.test 304/0, canvText.test 100/0, unixFont.test 50/0
  • Patch changes none
⬜ac07794b8a2023CI workflows: use CPPFLAGS rather than CFLAGS where appropriateNOT-A-BUG (the proposed change would be harmful for the unix builds)

Open ticket ac07794b8a · 81. Portability · filed 2023-11-17

Full verdict

NOT-A-BUG (the proposed change would be harmful for the unix builds)

  • Platform build (CI)
  • Tested main and 9.0 still have "CFLAGS=-DTK_NO_DEPRECATED=1" in linux-build.yml and linux-with-tcl9(1)-build.yml. 8.6 has no workflows any more, and .travis.yml was removed in a8515bb7.
  • Cause unix/Makefile.in (and Tcl's) never references @CPPFLAGS@. With CPPFLAGS=-DTK_NO_DEPRECATED=1 the define would reach only the configure checks, not the compile, so the "no deprecated" CI job would silently stop testing anything. CFLAGS=-arch in onefiledist.yml is a compiler flag and is correct as is.
  • Prototype none. A patch would first need CPPFLAGS added to CC_SWITCHES in unix/Makefile.in, which is a build-system change.
  • Suggested ticket comment unix/Makefile.in doesn't use <tt>@CPPFLAGS@</tt>, so moving <tt>-DTK_NO_DEPRECATED=1</tt> to CPPFLAGS would drop it from the actual build. Keeping CFLAGS; .travis.yml is already gone. Suggest closing.
⬜bec431fc682023NSWindow drag regions should only be invalidated on the Main ThreadNOT-A-BUG (Tk/Tkinter used from a non-main thread under gradio). CANNOT-TEST (macOS).

Open ticket bec431fc68 · 83. Mac OS X Build · filed 2023-09-02

Full verdict

NOT-A-BUG (Tk/Tkinter used from a non-main thread under gradio). CANNOT-TEST (macOS).

  • Platform Aqua
  • Tested none
  • Cause AppKit requires the main thread. The user's callback creates tk.Tk() in a gradio worker thread, so Tk_Init reaches TkMacOSXMakeRealWindowExist off the main thread.
  • Prototype none
  • Suggested ticket comment Tk Aqua (AppKit) can only be used from the process's main thread; creating a Tk interpreter from a gradio worker thread is unsupported. Closing as not a Tk bug.
⬜2b1ce317802023Generalise scaling variables for multi-display systemsNOT-A-BUG (RFE; the limitation is still present on main)

Open ticket 2b1ce31780 · 46. Unix Fonts · filed 2023-08-23

Full verdict

NOT-A-BUG (RFE; the limitation is still present on main)

  • Platform X11 (design / RFE)
  • Tested main tip. ::tk::scalingPct/::tk::svgFmt are still single global variables that tkCmds.c keeps in sync with [tk scaling] of "." only. The Xvfb test couldn't show a difference because two screens of the same display share one TkDisplay scaling value. Branch scaling-multidisplay (79bdc75153) was last merged in 2023-09 and never went into trunk.
  • Cause the design follows only the main window's display. The proposed fix adds new public commands (tk_scalingPct/tk_svgFormat -displayof), which needs a TIP (fvogel).
  • Prototype none (new API)
  • Suggested ticket comment Still valid for multi-display apps, but the fix adds public commands and needs a TIP; changing type to RFE. Branch scaling-multidisplay is stale since 2023-09.
⬜6bb7fd9ea92023Generalise tk systray for multi-display systemsNOT-A-BUG (RFE; still single-display on main)

Open ticket 6bb7fd9ea9 · (unused) · filed 2023-08-19

Full verdict

NOT-A-BUG (RFE; still single-display on main)

  • Platform X11 (RFE)
  • Tested main source. tk systray has no -displayof option. unix/tkUnixSysTray.c creates the tray icon on Tk_MainWindow(interp), and library/systray.tcl keeps one global icon.
  • Cause by design there is one tray icon per interpreter.
  • Prototype none (new option/API; would need a TIP)
  • Suggested ticket comment This asks for a new <tt>-displayof</tt> option, so it's a feature request (TIP) rather than a bug. Changing type to RFE.
✅da23032e6a2023ttk::combobox: Returned index is {} instead of expected -1NOT-A-BUG (intended TIP 577 change in 9.0; docs already updated in 5471e09e40, 2023-12-26)

Open ticket da23032e6a · 88. Themed Tk · filed 2023-07-21

✅ Since triage (updated 2026-11-11)

Status now: Closed (Wont Fix)

  • 💬 2026-09-30 serhiy.storchaka: This is the intended change of TIP 577 in Tk 9.0. The documentation has been updated in 5471e09e40: current returns an empty string if the current value is not in the list of -values. As noted above, [.c current] works with both 8.6 and 9.0.
  • 💬 2026-09-30 stu: For the record, I still don't agree with this. The comparison "" < 0 is at minimum wrong-looking. "Actually, checking whether [.c current] equals -1 has never been the recommended style." [Citation Needed]
  • 💬 2026-09-30 serhiy.storchaka: Reopened, but I think it is too late to revert this now. The change was intentional and has been in Tk since 9.0, and programs which were broken by it have probably already been adapted, so reverting it could break them again.
  • 💬 2026-10-01 stu: I don't think this needs re-opening; I doubt it's going to change, and I accept that. I just wanted to register my thoughts on the subject. Thanks.
  • 💬 2026-10-01 nemethi: Closing the ticket again.
Full verdict

NOT-A-BUG (intended TIP 577 change in 9.0; docs already updated in 5471e09e40, 2023-12-26)

  • Platform generic
  • Tested ttk::combobox .c; .c current → {} on main, 9.1b1, 9.0.5; -1 on 8.7b1, 8.6.18, 8.5.19. doc/ttk_combobox.n says {} in main/9.0 and -1 in 8.7/8.6, so each branch matches its docs.
  • Cause TIP 577 (9.0 returns an empty string for "no index"). Portable check: [.c current] < 0 works on all versions (nemethi's comment).
  • Prototype none (not needed)
  • Suggested ticket comment Intended by TIP 577. The man page was updated in 5471e09e40 and now documents {} for 9.0+. Use [.c current] < 0 for code that must also run on 8.x. Closing.
⬜6b49149b4e2023Tkinter problem with text getNOT-A-BUG in Tk; fixed on the Tkinter side. Tcl 8.7 registers "utf32string" (tclObj.c:394), and CPython gh-103194 (3.13, backported) makes `_tkinte…

Open ticket 6b49149b4e · 18. [text] · filed 2023-04-16

Full verdict

NOT-A-BUG in Tk; fixed on the Tkinter side. Tcl 8.7 registers "utf32string" (tclObj.c:394), and CPython gh-103194 (3.13, backported) makes _tkinter.FromObj() recognise UTF32StringType, so text get now returns a Python str. The workaround branch bug-6b49149b4e (571b0fd902, 2023-04-16) was never merged and is no longer needed.

  • Platform generic (Tcl 8.7 + Tkinter)
  • Tested source inspection of the Tcl 8.7 tree and CPython 3.13 Modules/_tkinter.c; no Python-with-8.7 build to run.
  • Cause Tkinter did not recognise Tcl 8.7's "utf32string" object type.
  • Prototype none
  • Suggested ticket comment Fixed in Tkinter by python/cpython gh-103194 (it handles the "utf32string" type registered by Tcl 8.7). The Tk-side workaround branch bug-6b49149b4e is not needed. Closing.
⬜f285ddcd232022GIF format not handled correctlyNOT-A-BUG (RFE). The metadata part is done by TIP 632 (832cdddb) in 8.7/9.0/main. What is left, automatic playback of animated GIFs, is a feature…

Open ticket f285ddcd23 · 42. Photo Image|GIF · filed 2022-06-03

Full verdict

NOT-A-BUG (RFE). The metadata part is done by TIP 632 (832cdddb) in 8.7/9.0/main. What is left, automatic playback of animated GIFs, is a feature request.

  • Platform generic
  • Tested made a 2-frame GIF with a partial second frame (PIL) and ran it on main, 9.0 and 8.6. -metadata gives {update region} {5 5 5 5} {delay time} 10 {disposal method} {do not dispose} on main and 9.0 (not available in 8.6). Reading -index 1 into an existing photo keeps the pixels outside the update region, so a script can do correct compositing.
  • Cause n/a
  • Prototype none (RFE)
  • Suggested ticket comment Frame delay, disposal method and update region are available through -metadata since TIP 632 (Tk 8.7/9.0), and reading a frame into an existing photo keeps the pixels outside the update region. Automatic animation would be a separate RFE/TIP. Suggest closing this ticket or turning it into an RFE.
⬜0fc1fd8ec32022tcllib des package missing procNOT-A-BUG (wrong tracker: tcllib, not Tk; reporter admits it; already forwarded as tcllib ticket a1dec8173c)

Open ticket 0fc1fd8ec3 · 23. Option Parsing · filed 2022-04-09

Full verdict

NOT-A-BUG (wrong tracker: tcllib, not Tk; reporter admits it; already forwarded as tcllib ticket a1dec8173c)

⬜a296cf46c62022[pack propagate $w 0] doesn't restore the size of the once shrunk widgetpatchNOT-A-BUG (long-standing documented behaviour: disabling propagation only means the geometry manager stops setting the requested size; it never res…

Open ticket a296cf46c6 · 48. Geometry Management · filed 2022-04-09

Full verdict

NOT-A-BUG (long-standing documented behaviour: disabling propagation only means the geometry manager stops setting the requested size; it never restores a previous one) — DOCS clarification proposed

  • Platform generic
  • Tested main, 8.6, 8.4.20 on X11: all keep the shrunk reqsize (e.g. 63x34) after pack propagate .f 0, pack forget, destroy; .f configure -width [.f cget -width] restores 300x200. grid behaves the same.
  • Cause tkPack.c/tkGrid.c only call Tk_GeometryRequest while propagation is on; the frame's -width/-height are applied only at configure time (same root cause as withdrawn TIP #454).
  • Prototype a296cf46c6.patch (docs only: pack.n + grid.n, 11 lines; also fixes pack.n typo "requested width and height of the packer" → "of the container")
  • Suggested ticket comment This is by design: when propagation is turned off, pack/grid simply stop changing the container's requested size, they do not restore an earlier one (re-apply -width/-height for that). Documented this in pack.n and grid.n.
⬜80f2e1ad792022combobox autocomplete in readonly modeNOT-A-BUG (feature request; needs TIP, as fvogel noted)

Open ticket 80f2e1ad79 · 88. Themed Tk · filed 2022-01-29

Full verdict

NOT-A-BUG (feature request; needs TIP, as fvogel noted)

  • Platform generic
  • Tested n/a
  • Cause n/a
  • Prototype none (RFE)
  • Suggested ticket comment This is a feature request; changing type to RFE. Needs a TIP and an implementation (see wiki code on ttk::combobox).
⬜2058b607302021tip#611 proposed implementationNOT-A-BUG (TIP 611 implementation: photo rotate/reflect subcommands. It is a feature, not a bug, and not in main.)

Open ticket 2058b60730 · 41. Photo Images · filed 2021-09-03

Full verdict

NOT-A-BUG (TIP 611 implementation: photo rotate/reflect subcommands. It is a feature, not a bug, and not in main.)

  • Platform generic
  • Tested the attached tip611.patch (against core-8-6-branch, 2021) does not apply to main (3 of 4 hunks fail). main has no photo rotate/reflect.
  • Cause n/a
  • Prototype none (feature; needs TIP vote and rebase)
  • Suggested ticket comment This is a TIP #611 implementation rather than a bug. It should go to a tip-611 branch (the patch needs a rebase onto main), so this ticket can be closed or changed to a Feature Request.
⬜497e7621b42021Text widget does not honor font option from Tk's databaseNOT-A-BUG (by design: every -font value is a complete font description, and unspecified attributes get defaults. The previous value, e.g. one from …

Open ticket 497e7621b4 · 75. wish · filed 2021-06-06

Full verdict

NOT-A-BUG (by design: every -font value is a complete font description, and unspecified attributes get defaults. The previous value, e.g. one from the option database, is not merged in. Also, TkFixedFont is a named font, not a family, so -family TkFixedFont falls back to the default family.)

  • Platform generic
  • Tested main, 8.6 (Xvfb, no WM). {-family TkFixedFont -size 48} → actual Noto Sans 48. Then {-overstrike 1} → Noto Sans 12 (the default size), overstrike 1.
  • Cause font.n, form [5] "option value ...": missing attributes take their defaults (size 0 means the platform default). Nothing in the -font option inherits.
  • Prototype none (not a bug)
  • Suggested ticket comment A font description in option-value form is complete: omitted attributes take their defaults, not the values of the widget's previous font. To derive a font, use a named font (font create F {*}[font actual TkFixedFont] -size 48; font configure F -overstrike 1). Note that -family TkFixedFont isn't valid either: TkFixedFont is a named font, not a family. Closing as not a bug.
⬜e5ccf542542020Binding to toplevels fails for the root windowNOT-A-BUG (documented behaviour: "." gets the application class, and its bindtags are ". <Class> all")

Open ticket e5ccf54254 · 01. Bindings · filed 2020-12-14

Full verdict

NOT-A-BUG (documented behaviour: "." gets the application class, and its bindtags are ". <Class> all")

  • Platform generic
  • Tested main, 8.6: winfo class . = app class, bindtags . = {. App all}, bindtags .t = {.t Toplevel all}
  • Cause tkWindow.c/tkFrame.c create "." as a toplevel with -class set to the application class. This is the same as toplevel .t -class X.
  • Prototype none (changing it would be an incompatible behaviour change; that would need a TIP)
  • Suggested ticket comment Works as documented (wish(1), bindtags(n)): "." is a toplevel whose class is the application class. Use bindtags . [linsert [bindtags .] 2 Toplevel] if needed. Suggest closing as "Works as designed", or turning this into a TIP for 9.x.
⬜4dc9cdf8f12020font sizing not correct for bold font on windows 10NOT-A-BUG (documented behaviour; RFE)

Open ticket 4dc9cdf8f1 · 47. Win Fonts · filed 2020-06-23

Full verdict

NOT-A-BUG (documented behaviour; RFE)

  • Platform generic
  • Tested X11 main and 8.6. font create B -weight bold gives -size 0, which resolves to 12 while TkDefaultFont is 9/10. This is the same behaviour on every platform.
  • Cause font.n: "If size is unspecified or zero, a platform-dependent default size is chosen". New fonts do not inherit from TkDefaultFont.
  • Prototype none (inheriting from TkDefaultFont would be a behaviour change and needs a TIP)
  • Suggested ticket comment Documented behaviour: a size of 0 means the platform default size, not the TkDefaultFont size. Use font create B {*}[font actual TkDefaultFont] -weight bold. Inheriting defaults from TkDefaultFont would be an RFE/TIP.
⬜9cb6a7300320198.6.10 testingNOT-A-BUG (obsolete). This is a log of test results for the 8.6.10 release candidate on several systems, not a single bug. Most of the listed failu…

Open ticket 9cb6a73003 · 86. Test Tools · filed 2019-10-01

Full verdict

NOT-A-BUG (obsolete). This is a log of test results for the 8.6.10 release candidate on several systems, not a single bug. Most of the listed failures depend on the environment (no WM/Wayland, fonts, macOS drawing).

  • Platform generic (test-suite report)
  • Tested main in the wt build, Xvfb with no WM: listbox, scrollbar and select tests pass (listbox-21.9/21.10 from the comments pass). focus-2.6/2.7/2.9/5.1 fail only because no WM was running.
  • Cause n/a
  • Prototype none
  • Suggested ticket comment This is a release-testing log for 8.6.10. Current failures should be reported as separate tickets. Closing as out of date.
⬜67384bce7d2019Side effects of destroying an embedded window depend on the OSNOT-A-BUG (intentional platform difference; the tests were adapted in 02ca3773, 2019-05)

Open ticket 67384bce7d · 65. Generic Window Operations · filed 2019-01-24

Full verdict

NOT-A-BUG (intentional platform difference; the tests were adapted in 02ca3773, 2019-05)

  • Platform generic (design question)
  • Tested main, 8.6, 8.5, 8.4 on X11. Destroying the embedded toplevel also destroys the container on every version (unix/tkUnixEmbed.c EmbedStructureProc → Tk_DestroyWindow on DestroyNotify). winfo-13.2 expects this on unix and passes.
  • Cause unix/tkUnixEmbed.c and macosx destroy the container on purpose; win/tkWinEmbed.c does not.
  • Prototype none (would be a behaviour decision). The behaviour is not in the docs (frame.n and toplevel.n say nothing about it). Documenting it in frame.n under -container would be the cheap resolution.
  • Suggested ticket comment This has been the documented-by-tests behaviour on X11 since 8.4, and winfo-13.2 is now platform-aware. Unless someone wants Windows changed, I suggest noting in frame.n (-container) that on X11/macOS the container is destroyed with the embedded window, then closing.
⬜8634f9d5372018namespace path errors in ::oo::InfoObjectNOT-A-BUG (still behaves this way in Tcl 9.1b1, 8.6.18 and 8.6.8, and it is expected)

Open ticket 8634f9d537 · (unused) · filed 2018-12-19

Full verdict

NOT-A-BUG (still behaves this way in Tcl 9.1b1, 8.6.18 and 8.6.8, and it is expected)

  • Platform generic (Tcl, not Tk)
  • Tested tclsh 9.1b1, 8.6.18, 8.6.8
  • Cause info object namespace is implemented as the command ::oo::InfoObject::namespace. Inside namespace eval ::oo::InfoObject, the word namespace resolves to that command, which treats "path" as an object name. Using ::namespace path works.
  • Prototype none
  • Suggested ticket comment Not a bug, and a Tcl matter anyway. ::oo::InfoObject contains a command named namespace (the implementation of info object namespace), and it shadows the global one inside that namespace. Use ::namespace path. Closing.
✅65c7412f492018make distclean errors out the second timeNOT-A-BUG (still behaves this way on main, and it is the standard autoconf/GNU behaviour)

Open ticket 65c7412f49 · 84. Unix Build · filed 2018-11-10

✅ Since triage (updated 2026-11-11)

Status now: Closed (Invalid)

  • 💬 2026-10-20 serhiy.storchaka: This is expected: make distclean removes all files created by configure, including Makefile, as required by the GNU Coding Standards and as in Tcl and Automake-based packages. After that there is no Makefile, so make cannot know any target. Run configure again, or use make clean if the Makefile should be kept.
Full verdict

NOT-A-BUG (still behaves this way on main, and it is the standard autoconf/GNU behaviour)

  • Platform build
  • Tested read unix/Makefile.in in main (distclean: rm -rf Makefile config.status ...). Tcl's unix/Makefile.in does the same.
  • Cause the GNU Coding Standards require distclean to remove everything configure created, including Makefile. After that, make has no Makefile to read.
  • Prototype none
  • Suggested ticket comment This is expected. distclean removes all files created by configure, including the Makefile (GNU Coding Standards; Tcl behaves the same), so a second make distclean has no Makefile to run. Closing.
⬜e01a250dc82017Tk radiobutton indicator does not change size with font sizeNOT-A-BUG (design since TIP #109). The indicator is still a fixed size relative to the font.

Open ticket e01a250dc8 · 03. [*button] and [label] · filed 2017-09-07

Full verdict

NOT-A-BUG (design since TIP #109). The indicator is still a fixed size relative to the font.

  • Platform X11
  • Tested main 9.1b1, 8.6.18, 8.4.20 on Xvfb, fonts Helvetica 6 and 32. The indicator space grows with the font in all versions (8.6: 21 to 77 px). The drawn indicator stays fixed: 8.6 uses a hardcoded bitmap, and main uses a 16x16 SVG scaled only by TkScalingLevel2 (tk scaling), not by the font size.
  • Cause unix/tkUnixButton.c TkpDrawCheckIndicator uses fixed RADIO_BUTTON_DIM, scaled by display scaling only. This was intentional in TIP #109 (Windows-like look).
  • Prototype none (would be a behaviour change and needs a TIP or at least a discussion)
  • Suggested ticket comment This is by design since TIP #109. In 9.x the indicator scales with [tk scaling], not with the font size. For font-proportional indicators use -image/-selectimage, or ttk::radiobutton with a style -indicatordiameter. Suggest closing as Won't Fix.
⬜c2290bd04b2017Iwidgets incompatible with auto_load - produces command error.NOT-A-BUG (in Tk). Without Itcl the script works on main 9.1 and 8.6 (prints "alpha beta Charlie"). Itcl/Iwidgets is not installed here, so I could…

Open ticket c2290bd04b · 99. Other · filed 2017-01-29

Full verdict

NOT-A-BUG (in Tk). Without Itcl the script works on main 9.1 and 8.6 (prints "alpha beta Charlie"). Itcl/Iwidgets is not installed here, so I could not run the failing case.

  • Platform generic (Itcl, not Tk)
  • Tested tclsh9.1 and tclsh8.6 with the ticket's src.tcl plus auto_mkindex and dst.tcl.
  • Cause itcl library/itcl.tcl (as of itcl.fossil trunk 2017) redefines ::auto_import with an Itcl-3-era version that creates ::itcl::import::stub stubs instead of sourcing the auto_index entry. It also lacks Tcl's check that the qualifiers match. So namespace import ::src::* does not load the procs once Itcl is loaded.
  • Prototype none (belongs to Itcl, duplicate of Tcl ticket ba79c2f0ce)
  • Suggested ticket comment Not a Tk issue. Itcl's library/itcl.tcl overrides ::auto_import with its own stub-based version, which breaks autoloading on namespace import. It is already tracked as Tcl ba79c2f0ce; it should be moved to the Itcl tracker and closed here.
⬜4a063ca0672016bordercolor in ttk buttonNOT-A-BUG (theme design; RFE at most)

Open ticket 4a063ca067 · 88. Themed Tk · filed 2016-11-19

Full verdict

NOT-A-BUG (theme design; RFE at most)

  • Platform generic
  • Tested main and 8.6, ttk::style element options for Button.border in each theme. -bordercolor is supported by clam, and by alt only for the default-button ring. default and classic draw 3-D bevels derived from -background. Native themes (vista, xpnative, aqua) draw with the OS.
  • Cause the Button.border element of the default and classic themes (ttkElements.c) has no -bordercolor option by design.
  • Prototype none (would be a feature: a new element option in the default theme)
  • Suggested ticket comment -bordercolor is a theme-specific styling option. It is honored by clam (and by alt for the default ring), not by the 3-D bevel themes or the native ones. Use clam, or create a custom element with ttk::style element create. Suggest closing as "Works as designed" or turning it into an RFE.
⬜1dde2144bd2016ttk style configure overwrites Tk option add Button but not LabelNOT-A-BUG (ttk::label has a -font widget option; ttk::button does not)

Open ticket 1dde2144bd · 88. Themed Tk · filed 2016-09-29

Full verdict

NOT-A-BUG (ttk::label has a -font widget option; ttk::button does not)

  • Platform generic
  • Tested main, 8.6, 8.5: label gets {sans-serif 12} from the option db (h=25), button gets courier 40 from the style (h=68).
  • Cause ttk::label has a -font widget option (documented in ttk_label.n), which the option database initializes, and widget options override style settings. ttk::button has no -font widget option, so only the style applies.
  • Prototype none
  • Suggested ticket comment Works as designed: ttk::label has a -font widget option, which option add *Font initializes and which overrides the style. ttk::button has no such option. Use option add *TLabel.font ... or a more specific pattern, or set the font with ttk::style only.
⏳985607f2402016Tk reports wrong physical screen sizeNOT-A-BUG (winfo screenmm* reports the nominal X screen size, which tk scaling rescales; per-monitor physical size would be a new feature)

Open ticket 985607f240 · 67. Unix Window Operations · filed 2016-01-16

⏳ Since triage (updated 2026-11-11)

Status now: Pending

    Full verdict

    NOT-A-BUG (winfo screenmm* reports the nominal X screen size, which tk scaling rescales; per-monitor physical size would be a new feature)

    • Platform X11
    • Tested main on Xvfb 1280x1024. With default scaling 1.389 it reports 325x260 mm; after tk scaling 1.0 it reports 452x361 mm. So the value follows Tk's scaling, not EDID data.
    • Cause generic/tkCmds.c (tk scaling) writes the mm dimensions back into the Screen structure. For two monitors as one X screen, the server's own value is the logical size at an assumed dpi.
    • Prototype none (a per-monitor query needs a TIP).
    • Suggested ticket comment Works as designed: the millimetre size is the logical X screen size and follows tk scaling. Reporting the physical size of each monitor would need a TIP (see the related multi-monitor tickets listed in c95d4691ff). I suggest closing, or converting to an RFE.
    ⬜7fcb7ca5862014exec on a script which execute the 'rsh' commandNOT-A-BUG (standard Unix job control)

    Open ticket 7fcb7ca586 · 75. wish · filed 2014-10-08

    Full verdict

    NOT-A-BUG (standard Unix job control)

    • Platform generic (Unix)
    • Tested main on Xvfb, with wish started as a background job under an interactive bash in a pty, using a child script that reads stdin.
    • Cause exec cmd & without redirection makes the child inherit wish's stdin, the controlling tty. wish was started with &, so its job is a background process group. When rsh reads the tty, the kernel sends SIGTTIN to the whole process group, which includes wish, and everything stops (ps shows wish "Tl", job "Stopped (tty input)"). With </dev/null wish keeps running. tclsh would behave the same way.
    • Prototype none
    • Suggested ticket comment This is job control, not Tk. The background job reads from the terminal, so SIGTTIN stops the whole process group, including wish. Use exec ./script.sh </dev/null & (or rsh -n).
    ✅0e7185dc4c2013ttk combobox selection issueNOT-A-BUG (by design; the dropdown case is NOT-REPRODUCED)

    Open ticket 0e7185dc4c · 88. Themed Tk · filed 2013-07-12

    ✅ Since triage (updated 2026-11-11)

    Status now: Closed (Out of Date)

    • 💬 2026-09-30 serhiy.storchaka: Not reproducible with 8.6 and 9.1 on Linux/X11 and Windows: after choosing an item from the drop-down list, the whole new value is selected and exported as the PRIMARY selection. The report was for the separate Tile 0.8.6 package, not the ttk built into Tk. Changing the -textvariable directly still keeps the selection range, as jenglish explained; entry, ttk::entry and ttk::combobox all behave the same way.
    Full verdict

    NOT-A-BUG (by design; the dropdown case is NOT-REPRODUCED)

    • Platform generic
    • Tested main (9.1b1), 8.6.18, 8.5.19 on Xvfb with no WM. The ticket's steps were simulated: select the entry text, post the dropdown, pick "Europe" with LBSelected.
    • Cause ttk::combobox::SelectEntry runs selection range 0 end, so picking an item from the list selects the whole value ("Europe" is 0-6, and selection get returns Europe) on all versions. Changing the -textvariable leaves the sel.first/sel.last indices as they were and only clamps them to the new value. For example, 0-end on "Europe" followed by set combo USA gives 0-3. The core entry and ttk::entry do the same. jenglish says in the ticket that this is intended.
    • Prototype none (whether to clear the selection when the value changes is a design decision)
    • Suggested ticket comment Could not reproduce the dropdown case with 8.5.19, 8.6.18 or 9.1b1: picking an item always selects the whole new value. When only the -textvariable changes, the selection indices are kept, as in entry and ttk::entry, so this works as designed. Proposing to close.
    ⬜3613977fff2013Tk radiobutton missing an optionNOT-A-BUG (feature request)

    Open ticket 3613977fff · 03. [*button] and [label] · filed 2013-05-27

    Full verdict

    NOT-A-BUG (feature request)

    • Platform generic
    • Tested code check on main. selectBorder (-selectcolor) is the only selected-state colour in tkButton.c and tkUnixButton.c. There is still no selected-state foreground option.
    • Cause n/a. Adding a new option (e.g. -selectforeground for radiobutton/checkbutton with -indicatoron 0) needs a TIP.
    • Prototype none (new option = API change)
    • Suggested ticket comment This is a feature request (a new option such as -selectforeground), not a bug. It would need a TIP. ttk::radiobutton with a style map on -foreground for the selected state works today.
    ⬜3603396fff2013ttk::sizegrip doesn't use native bindingsNOT-A-BUG (RFE: native Windows modal sizing loop with outline and Escape to cancel)

    Open ticket 3603396fff · 88. Themed Tk · filed 2013-02-05

    Full verdict

    NOT-A-BUG (RFE: native Windows modal sizing loop with outline and Escape to cancel)

    • Platform Windows
    • Tested none
    • Cause sizegrip.tcl resizes the window live by setting wm geometry. Native behaviour would need a C helper, e.g. posting WM_SYSCOMMAND SC_SIZE|WMSZ_BOTTOMRIGHT to the toplevel.
    • Prototype none (feature, needs new C-level hook)
    • Suggested ticket comment This is a feature request. It could be done on win32 by starting the system size loop (WM_SYSCOMMAND SC_SIZE+WMSZ_BOTTOMRIGHT) from the sizegrip binding. Re-categorize as RFE.
    ⬜3598814fff2012Invalid kern table error for 'font metrics'NOT-A-BUG (broken kern table in the third-party font "Bauhaus 93"; CoreText writes a log message and Tk raises no error). CANNOT-TEST (macOS)

    Open ticket 3598814fff · 45. Aqua Fonts · filed 2012-12-29

    Full verdict

    NOT-A-BUG (broken kern table in the third-party font "Bauhaus 93"; CoreText writes a log message and Tk raises no error). CANNOT-TEST (macOS)

    • Platform Aqua
    • Tested none
    • Cause The message is logged by CoreText on stderr/console for a malformed font; other applications print it too (wordtech's comment).
    • Prototype none
    • Suggested ticket comment The message comes from CoreText about a malformed font and is not a Tk error. Closing as third-party font issue.
    🔧3565268fff2012bind<ButtonRelease-1> doesnt have same behavior on windowsNOT-A-BUG (by design) / NEEDS-WINDOWS to confirm

    Open ticket 3565268fff · 01. Bindings · filed 2012-09-06

    🔧 Since triage (updated 2026-11-11)

    Status now: Open

    • 🔧 2026-10-08 check-in 57bfd88713 on x11-check-toggle-on-release, not merged: Fix [3565268]: On X11, classic checkbuttons and radiobuttons are now invoked when mouse button 1 is released over them, not when it is pressed, as on Windows an
    • 💬 2026-10-08 serhiy.storchaka: On X11, the classic checkbutton and radiobutton were invoked when mouse button 1 was pressed. On Windows and macOS they are invoked when it is released, in the class binding, which runs after the widget binding. All other toolkits on X11 also toggle on release: Motif, Athena, GTK 3, GTK 4 and Qt 6. So do Tk's ttk checkbutton and radiobutton on all platforms. Proposed fix in branch 57bfd88713: on X11 the classic …
    Full verdict

    NOT-A-BUG (by design) / NEEDS-WINDOWS to confirm

    • Platform Windows (also Aqua)
    • Tested X11 main and 8.5: a widget <ButtonRelease-1> binding sees the new value. Repro: win/3565268fff.tcl
    • Cause library/button.tcl. On x11 the class binding toggles on <Button-1> (tk::CheckInvoke / tk::CheckRadioInvoke). On win32 and aqua it toggles in the class <ButtonRelease-1> binding (tk::ButtonUp), which follows the native look and feel. Widget bindings run before class bindings, so on Windows the widget's release binding still sees the old value.
    • Prototype none (intentional platform behaviour)
    • Suggested ticket comment This is by design. On Windows and macOS the check/radiobutton toggles on button release in its class binding, and that runs after the widget binding. Use -command, a variable trace, or a bindtag placed after the class tag.
    ⏳3543420fff2012Toggle.TButton style doesn’t work on WindowsNOT-A-BUG (not run: the Windows VM was powered off, VMState=poweroff since 2026-10-03 21:23, so no 8.6.18 or trunk run)

    Open ticket 3543420fff · 88. Themed Tk · filed 2012-07-13

    ⏳ Since triage (updated 2026-11-11)

    Status now: Pending (Remind)

      Full verdict

      NOT-A-BUG (not run: the Windows VM was powered off, VMState=poweroff since 2026-10-03 21:23, so no 8.6.18 or trunk run)

      • Platform Windows (vista/xpnative themes)
      • Tested Read the attached toggle.pyw. It maps -relief sunken/raised on "Toggle.TButton" but keeps the default TButton layout. Its second button already uses the Toolbutton style. I relied on jenglish's 2012 analysis in the ticket. A local read of library/ttk/vistaTheme.tcl was blocked by the permission classifier, so I did not re-check the theme files myself.
      • Cause On the native Windows themes, TButton's border element is drawn by the visual-styles API (Button.button) and has no -relief option, so style map ... -relief has no effect. Only the default/alt/clam/classic elements honour -relief.
      • Prototype none (works as designed)
      • Suggested ticket comment Works as designed: native Windows button elements ignore -relief. For a toggle button, use ttk::checkbutton -style Toolbutton, which draws a pushed-in state when selected on every theme. This has been Pending since 2016 with nothing new, so it can be closed.
      ⬜3517803fff2012Modality not properly workingNOT-A-BUG (documented grab semantics)

      Open ticket 3517803fff · 56. [grab] · filed 2012-04-14

      Full verdict

      NOT-A-BUG (documented grab semantics)

      • Platform generic
      • Tested X11 main and 8.5, no WM: with a grab on .app.fsb, clicks in the sibling toplevel .help go to the grab window (.help.b is not invoked). Descendant toplevels of the grab window still work.
      • Cause generic/tkGrab.c TkGrabState. The mainPtr check separates applications, not toplevel trees. grab.n says a local grab affects the whole application: windows outside the grab subtree "will be insensitive until the grab is released". Only toplevels under the grab window take part in it.
      • Prototype none. The proposed change would redefine what local grabs mean.
      • Suggested ticket comment This is how grab is documented: a local grab makes every window of the application outside the grab subtree insensitive. To keep a help window usable during the modal dialog, make it a descendant of the grabbed window (for example .application.fileSelBox.help).
      ⬜3476090fff2012-help does not work in tk 8.5.11 on MacOspatchNOT-A-BUG (Tk_ParseArgv called before Tk is initialized) / DOCS

      Open ticket 3476090fff · 73. Argv Parsing · filed 2012-01-19

      Full verdict

      NOT-A-BUG (Tk_ParseArgv called before Tk is initialized) / DOCS

      • Platform generic (C API)
      • Tested own C program calling Tk_ParseArgv(interp, NULL, ...) with -help, without Tk_Init: segfault on Linux main and 8.6.18, works on 8.5.19. With Tcl_Init+Tk_Init first: prints the usage on main and 8.6.
      • Cause libtk is built with Tcl stubs (on Unix since 8.6, probably already in the 8.5 macOS framework). The Tk-side tclStubsPtr is only set up by Tk_Init. Plain option parsing makes no Tcl calls, but PrintUsage calls Tcl_NewStringObj through the stub table, which is NULL at that point, so it crashes. The attached program is not available; this is the most likely explanation.
      • Prototype 3476090fff.patch (4 lines, doc/ParseArgv.3 only: says that Tk must be initialized in interp before calling Tk_ParseArgv, even if tkwin is NULL)
      • Suggested ticket comment The crash happens because Tk_ParseArgv is called before Tk_Init. Tk calls Tcl through its stub table, and Tk_Init is what sets that table up. Call Tcl_Init/Tk_Init first. The manual page will say this explicitly.
      ✅3414673fff2011Improper look of ttk::checkbutton -style ToolbuttonNOT-A-BUG (standard sunken relief)

      Open ticket 3414673fff · 88. Themed Tk · filed 2011-09-27

      ✅ Since triage (updated 2026-11-11)

      Status now: Closed (Fixed)

      • 🔧 2026-09-30 check-in a0094e41af on toolbutton-pressed-look, in main, 9.0: Fix [3414673]: Improper look of ttk::checkbutton -style Toolbutton. The contents of a pressed or checked Toolbutton are shifted as for a pressed TButton in the
      • 🔧 2026-10-02 check-in bfa641b9a2 on trunk, in main, 9.0, 8.6: Fix for 3414673fff: Improper look of ttk::checkbutton -style Toolbutton.
      • 🔧 2026-10-02 check-in fe64661016 on core-9-0-branch, in main, 9.0: Fix for 3414673fff: Improper look of ttk::checkbutton -style Toolbutton.
      • 🔧 2026-10-02 check-in 7d19595beb on core-8-6-branch, in main, 9.0, 8.6: Fix for 3414673fff: Improper look of ttk::checkbutton -style Toolbutton.
      • 💬 2026-09-30 serhiy.storchaka: The colors of the sunken border are the usual ones (dark top and left, light bottom and right) and the same as for a pressed TButton. What is missing is the shift of the contents: TButton sets -shiftrelief 1 in the default, alt and winnative themes, but Toolbutton does not (only classic sets it), so a pressed or checked Toolbutton gets the sunken border without moving its contents. In the clam theme, whose border …
      • 💬 2026-10-02 nemethi: Many thanks, Serhiy! Merged into main, core-9-0-branch, and core-8-6-branch by commits bfa641b9, fe646610, and 7d19595b.
      Full verdict

      NOT-A-BUG (standard sunken relief)

      • Platform generic (X11 themes)
      • Tested main on X11 with the default, clam, alt and classic themes, in screenshots. A selected Toolbutton uses -relief sunken: dark top/left, light bottom/right. This is the same as a classic button -relief sunken.
      • Cause the selected/pressed state is mapped to relief sunken in library/ttk/defaults.tcl, clamTheme.tcl, altTheme.tcl and classicTheme.tcl. The reporter wants a different pressed look (thin border, shifted label), which is an RFE.
      • Prototype none (design preference)
      • Suggested ticket comment A selected Toolbutton is drawn with the standard sunken relief, the same as a Tk button with -relief sunken. A different look can be set with ttk::style map Toolbutton. Suggest closing as not a bug (or RFE).
      ⬜3403387fff2011Canvas Text ClippingNOT-A-BUG (RFE: a height/max-lines limit for canvas text items)

      Open ticket 3403387fff · 05. Canvas Items · filed 2011-09-02

      Full verdict

      NOT-A-BUG (RFE: a height/max-lines limit for canvas text items)

      • Platform generic
      • Tested main source. Canvas text items still have only -width for wrapping and no -height, -maxlines or clipping option.
      • Cause n/a (feature request)
      • Prototype none (new option means a TIP-level API change)
      • Suggested ticket comment This asks for a new feature (clipping or a line limit for canvas text items), not a bug. Needs a TIP; changing the type to RFE.
      ⬜3310586fff2011Doubleclicks generated at any speedNOT-A-BUG (documented behaviour; -time current added in 8.6.10 as the fix, TIP #532 merge 84fbb57ab3, 2019-08-28)

      Open ticket 3310586fff · 69. Events · filed 2011-06-02

      Full verdict

      NOT-A-BUG (documented behaviour; -time current added in 8.6.10 as the fix, TIP #532 merge 84fbb57ab3, 2019-08-28)

      • Platform generic
      • Tested main, 9.0, 8.6 tip, 8.6.9, 8.5, 8.4 on Xvfb. Without -time, all tick counts are 0, so a second generated click always counts as <Double-1> (all versions). On 8.6.10+ event generate ... -time current gives single/single/single, and a real double only when <10 ms apart. Explicit -time values (100/1100/1200) also work.
      • Cause generic/tkBind.c HandleEventGenerate: fields that are not given default to 0, so all generated button events have time 0 and fall inside the double-click interval.
      • Prototype none. Defaulting -time to "current" would change long-standing behaviour that tests rely on. Side note: -time [clock milliseconds] fails with "integer value too large" because Tcl_GetIntFromObj is used for an X Time (unsigned long). That is a separate small issue.
      • Suggested ticket comment Generated events get time 0 unless -time is given, so consecutive generated clicks count as multi-clicks. Since 8.6.10 (TIP #532) use event generate .c <Button-1> -time current (or explicit -time values). Closing as works-as-documented.
      ⬜3301481fff2011ttk::spinbox on windows doesnt handle -width 0NOT-A-BUG on X11 / NEEDS-WINDOWS to confirm vista

      Open ticket 3301481fff · 88. Themed Tk · filed 2011-05-13

      Full verdict

      NOT-A-BUG on X11 / NEEDS-WINDOWS to confirm vista

      • Platform Windows (X11 checked)
      • Tested main, 8.6 on X11, all themes. For ttk::entry, ttk::spinbox and ttk::combobox, -width 0 gives the same request as -width 1, and spinbox and combobox scale with -width exactly like entry. Windows script: win/3301481fff.tcl (prints VERDICT and checks the vista/winnative themes too).
      • Cause generic/ttk/ttkEntry.c TextareaElementSize treats width <= 0 as 1 char for all ttk entry-like widgets (since 2006). Unlike the classic entry, ttk has no "fit contents" meaning for 0 (doc/ttk_entry.n marks this "Not in ttk"). Shrinking in a narrow parent works through geometry management, not -width.
      • Prototype none
      • Suggested ticket comment In ttk, -width 0 is the same as -width 1 for entry, spinbox and combobox on all platforms (ttk has no "fit the contents" meaning for 0). Use a small -width and let the geometry manager shrink or grow the widget. Unless the attached script shows a difference with the vista theme, this can be closed.
      ✅2935309fff2010No multi-column menus in Tk CocoaNOT-A-BUG (documented limitation: menu.n says -columnbreak "is ignored on Aqua/macOS, where menus are always a single column"; NSMenu has no such s…

      Open ticket 2935309fff · 11. Aqua Menus · filed 2010-01-19

      ✅ Since triage (updated 2026-11-11)

      Status now: Closed (Wont Fix)

      • 💬 2026-10-08 serhiy.storchaka: Menus on macOS are native NSMenus, which have no multi-column layout. This is now documented in the menu manual page: -columnbreak is ignored on Aqua/macOS, where menus are always a single column.
      Full verdict

      NOT-A-BUG (documented limitation: menu.n says -columnbreak "is ignored on Aqua/macOS, where menus are always a single column"; NSMenu has no such support)

      • Platform Aqua
      • Tested none (source/doc check only; no -columnbreak handling in macosx/tkMacOSXMenu.c)
      • Cause native NSMenu does not support multiple columns; would need a custom menu view (RFE).
      • Prototype none (RFE)
      • Suggested ticket comment This is a documented platform limitation (see -columnbreak in menu(n)): native macOS menus are single-column. Closing; an implementation would be a feature request.
      ⬜2891877fff200932k pixel limit when viewing a scrolled frame on a canvasNOT-A-BUG (window positions are 16-bit on X11 ConfigureWindow and on Windows; see the 2026-09-22 comment)

      Open ticket 2891877fff · 05. Canvas Items · filed 2009-11-04

      Full verdict

      NOT-A-BUG (window positions are 16-bit on X11 ConfigureWindow and on Windows; see the 2026-09-22 comment)

      • Platform X11, Windows (platform limitation)
      • Tested not re-tested; the 2026-09-22 analysis already covers 9.1 on X11 and Windows
      • Cause Tk_CanvasWindowCoords() stores the position in a short, but removing that clamp makes no difference because the windowing system limits child window positions to 16 bits.
      • Prototype none (platform limit; workaround is to scroll the embedded widget itself)
      • Suggested ticket comment This is a windowing system limit, not a Tk bug. Please scroll the content inside the embedded widget, or draw it with canvas items. Closing.
      ✅2874745fff2009TEA 3.6 incompatible TK_INCLUDE_SPEC defineNOT-A-BUG (TK_INCLUDES is not meant to come from tkConfig.sh)

      Open ticket 2874745fff · 84. Unix Build · filed 2009-10-08

      ✅ Since triage (updated 2026-11-11)

      Status now: Closed (Invalid)

      • 💬 2026-10-20 serhiy.storchaka: tkConfig.sh does not need to define TK_INCLUDES: in TEA it is computed and substituted by TEA_PUBLIC_TK_HEADERS (and TEA_PRIVATE_TK_HEADERS), which have used TK_INCLUDE_SPEC from tkConfig.sh to find tk.h since tclconfig 8803659fc8 (March 2009). An extension which uses @TK_INCLUDES@ in its Makefile.in must call TEA_PUBLIC_TK_HEADERS in its configure.ac.
      Full verdict

      NOT-A-BUG (TK_INCLUDES is not meant to come from tkConfig.sh)

      • Platform build
      • Tested code inspection. unix/tkConfig.sh.in still defines only TK_INCLUDE_SPEC; TEA's tcl.m4 computes and AC_SUBSTs TK_INCLUDES itself in TEA_PUBLIC_TK_HEADERS / TEA_PRIVATE_TK_HEADERS (e.g. Tix8.4.3/tclconfig/tcl.m4:3534-3567). An extension using @TK_INCLUDES@ must call TEA_PUBLIC_TK_HEADERS in its configure.ac.
      • Cause extension's configure (itclwidgets) not calling TEA_PUBLIC_TK_HEADERS.
      • Prototype none
      • Suggested ticket comment TK_INCLUDES is set by TEA itself (TEA_PUBLIC_TK_HEADERS / TEA_PRIVATE_TK_HEADERS in tcl.m4), not by tkConfig.sh. Extensions using @TK_INCLUDES@ need to call one of these macros in configure.ac. Closing as not a Tk bug.
      ⬜2820123fff2009Canvas item bindings don't always work.NOT-A-BUG (the cursor lines drawn at the pointer become the topmost "current" item, so the rectangle gets <Leave>)

      Open ticket 2820123fff · 05. Canvas Items · filed 2009-07-11

      Full verdict

      NOT-A-BUG (the cursor lines drawn at the pointer become the topmost "current" item, so the rectangle gets <Leave>)

      • Platform generic
      • Tested main, 8.6 (event generate + -warp): with d=0 the current item is a cursor line and the rectangle toggles Enter/Leave. With the cursor items created -state disabled the rectangle stays current.
      • Cause generic/tkCanvas.c PickCurrentItem picks the topmost item under the pointer, which is the freshly created line through the hot spot. This is by design.
      • Prototype none (not a bug)
      • Suggested ticket comment Works as designed: items drawn under the pointer become the ''current'' item, so the rectangle gets <Leave>. Create the cursor items with <tt>-state disabled</tt> (disabled items are ignored when picking) and the bindings work as expected.
      ⏳2543460fff2009Enter/Leave cnv events not reported while mouse btn is downpatchNOT-A-BUG (the canvas imitates the implicit pointer grab). The doc check-in 30be58dda3 is tagged "mistake" and is not on main; doc/canvas.n on ma…

      Open ticket 2543460fff · 04. Canvas Basics · filed 2009-01-28

      ⏳ Since triage (updated 2026-11-11)

      Status now: Pending (Wont Fix)

        Full verdict

        NOT-A-BUG (the canvas imitates the implicit pointer grab). The doc check-in 30be58dda3 is tagged "mistake" and is not on main; doc/canvas.n on main does not mention this behaviour.

        • Platform generic
        • Tested main 9.1.1 and 8.6.18 on X11 with event generate (ButtonPress-1 outside the item, Motion with B1 held over it, then release). Result on both: during the drag the current tag is empty and nothing fires; at release <Enter> fires and the item becomes current.
        • Cause generic/tkCanvas.c PickCurrentItem() does not change the current item while a button is down. This is intended and matches the window-system grab.
        • Prototype 2543460fff.patch (19 lines, doc only; the text of 30be58dda3, re-applied on main)
        • Suggested ticket comment This is by design: like the implicit grab for windows, the current item does not change while a button is pressed, so <Enter>/<Leave> and "current" update at the release. Documented in doc/canvas.n. Tracking the item during a drag would need a TIP.
        ⬜2378039fff2008treeview custom layout problemsNOT-A-BUG. -border 1 in a layout marks the element as a border element (children are placed inside its padding); it is not the border width. With…

        Open ticket 2378039fff · 88. Themed Tk · filed 2008-12-02

        Full verdict

        NOT-A-BUG. -border 1 in a layout marks the element as a border element (children are placed inside its padding); it is not the border width. Without Treeview.field, the empty area is filled with Treeview -background, as jenglish explained.

        • Platform Aqua (originally); generic on X11
        • Tested main, 8.6, 8.5 on X11 with a custom layout that has no field, checked by sampling pixels. Items and the empty area take -background; with the field, the empty area takes -fieldbackground.
        • Cause design. The field element draws the border and the field background.
        • Prototype none.
        • Suggested ticket comment Works as designed (see jenglish's comment). Use ttk::style configure Treeview -borderwidth 0, or remove the field element and set -background. We propose closing.
        ✅2128972fff2008wish 8.5.4 core dumpNOT-A-BUG (bug in the Oratcl exit handler) / CANNOT-TEST (Solaris, Oratcl)

        Open ticket 2128972fff · 75. wish · filed 2008-09-25

        ✅ Since triage (updated 2026-11-11)

        Status now: Closed (Invalid)

        • 💬 2026-10-08 serhiy.storchaka: The crash is in Oratcl: its exit handler (Oratcl_Exit → Oratcl_Clean) uses the hash table of an interpreter which was already deleted. When the last window is closed, wish deletes the interpreter and then calls Tcl_Exit, so the exit handlers run after the interpreter deletion callbacks. tclsh evaluates exit while the interpreter still exists, which is why the script did not crash there. Oratcl should handle the de…
        Full verdict

        NOT-A-BUG (bug in the Oratcl exit handler) / CANNOT-TEST (Solaris, Oratcl)

        • Platform generic (Solaris, Oratcl extension)
        • Tested none (third-party extension)
        • Cause Tk_MainEx calls Tcl_DeleteInterp(interp) and then Tcl_Exit(0) (tkMain.c, unchanged on main at lines 373/376). Oratcl's exit handler then walks a hash table that its interp-delete callback has already freed (Tcl_FirstHashEntry(NULL)). Under tclsh+Tk the process exits via [exit] without deleting the interp first, which is why it does not crash there.
        • Prototype none
        • Suggested ticket comment The crash is in Oratcl_Exit, which uses per-interp state after the interp was deleted. Tk_MainEx deletes the interp before Tcl_Exit (still true on main). Extensions must not use interp data in exit handlers. Suggest closing as third-party.
        ⬜2026017fff2008Event state bit mask is platform dependentpatchNOT-A-BUG (by design; the modifier mapping is platform and server specific). The missing documentation is a DOCS issue.

        Open ticket 2026017fff · 69. Events · filed 2008-07-23

        Full verdict

        NOT-A-BUG (by design; the modifier mapping is platform and server specific). The missing documentation is a DOCS issue.

        • Platform generic
        • Tested Source inspection on main. On X11, Alt is whatever modifier Alt_L is mapped to (Mod1=8 on Xvfb, per xmodmap). macOS uses Mod2Mask (16) for Option (tkMacOSXKeyEvent.c). Windows uses ALT_MASK = AnyModifier<<2 = 131072 (generic/tkInt.h, win/tkWinX.c, tkWinPointer.c).
        • Cause There is no portable Alt bit in X11. Tk passes the native state through.
        • Prototype 2026017fff.patch (24 lines, docs only: the bind.n %s entry lists the bit values and the platform-dependent Alt bit).
        • Suggested ticket comment The %s value is the raw modifier mask, and the Alt bit is necessarily platform dependent (X11: modifier holding Alt_L, usually Mod1=8; macOS Mod2=16; Windows 131072). Proposed resolution: document it in bind(n) (patch available). Portable code should use <Alt-...> bindings.
        ⬜1248241fff2005DllMain calls TkFinalize (arbitrary side effects)NOT-A-BUG (design limitation. The reported crash is mitigated. A real fix needs a public Tk_Finalize, i.e. a TIP)

        Open ticket 1248241fff · 74. Application Embedding · filed 2005-07-31

        Full verdict

        NOT-A-BUG (design limitation. The reported crash is mitigated. A real fix needs a public Tk_Finalize, i.e. a TIP)

        • Platform Windows
        • Tested source inspection only (main, 8.6)
        • Cause win/tkWin32Dll.c DllMain still calls TkFinalize(NULL) on DLL_PROCESS_DETACH. The SEH wrapper was removed in 71ee5f9d57 (2025-11-24). The crash when exit() runs before Tcl/Tk init was fixed by a tclStubsPtr check in TkFinalize, ab7f10ab98 (2016-02-11, "Fix crash in TkFinalize() if Tk_Init() is never called").
        • Prototype none (the change would affect the API or behaviour)
        • Suggested ticket comment The early-exit crash was fixed in ab7f10ab98 (TkFinalize returns if Tcl stubs were never initialized). Calling exit handlers from DllMain still works as designed. Removing that call needs a public Tk_Finalize (TIP). Suggest closing.
        ⏳1245475fff2005Using text and image on menu entry - extra space added.NOT-A-BUG (with -compound center the entry takes the larger of text and image; button padding comes from -padx/-pady)

        Open ticket 1245475fff · 10. Generic Menus · filed 2005-07-26

        ⏳ Since triage (updated 2026-11-11)

        Status now: Pending (Works for Me)

          Full verdict

          NOT-A-BUG (with -compound center the entry takes the larger of text and image; button padding comes from -padx/-pady)

          • Platform generic
          • Tested not re-tested (8.4.11, 8.6, 9.1 on X11 and Windows, per the 2026-09-22 comment)
          • Cause documented behaviour
          • Prototype none
          • Suggested ticket comment Pending without response; works as documented. Closing.
          ⬜1038760fff2004Tcl_DoOneEvent doesn't return when modal dialog is open.NOT-A-BUG (nested event loop by design)

          Open ticket 1038760fff · 70. Event Loop · filed 2004-10-01

          Full verdict

          NOT-A-BUG (nested event loop by design)

          • Platform generic (report is Linux/X11)
          • Tested code/design analysis only (no C harness needed)
          • Cause the menu -command runs inside the Tcl_DoOneEvent call that dispatched the menu event. tk_getOpenFile (tkfbox.tcl on X11) then waits in its own nested event loop (vwait/tkwait) until the dialog closes. That outer Tcl_DoOneEvent call cannot return while its handler is still running. TCL_DONT_WAIT only means "don't block when there are no events". It does not limit how long a handler may run. The external poll loop starves because its C work runs only between Tcl_DoOneEvent calls.
          • Prototype none (not a bug)
          • Suggested ticket comment Works as designed. The dialog runs a nested event loop inside the handler that the outer Tcl_DoOneEvent dispatched, so that call can't return until the dialog closes. Do the periodic C work from inside the Tcl event loop instead: a Tcl_CreateTimerHandler callback, a file handler, or a custom event source. The nested loop then keeps running it. Closing.
          ⬜1025471fff2004Problem with newline translation when appending to clipboardNOT-A-BUG (really an RFE). The behaviour is unchanged in main.

          Open ticket 1025471fff · 52. [clipboard] · filed 2004-09-09

          Full verdict

          NOT-A-BUG (really an RFE). The behaviour is unchanged in main.

          • Platform Windows
          • Tested code reading of win/tkWinClipboard.c:TkWinClipboardRender in main. It still turns every LF into CRLF and only serves the STRING target, as CF_UNICODETEXT.
          • Cause CRLF is the required line ending for CF_UNICODETEXT. Excel's paste treats LF inside a quoted field as a line break within the cell, and shows the extra CR as a stray character. Custom -format targets are not exported to the Windows clipboard at all, so there is currently no raw path.
          • Prototype none. A real solution would be an RFE: export named -format targets as registered clipboard formats (e.g. "Csv") without translation.
          • Suggested ticket comment The LF->CRLF translation is required for CF_TEXT/CF_UNICODETEXT, so it isn't a bug. Excel reads quoted in-cell newlines as bare LF. Handing raw data to Excel would need Tk to export custom -format targets (e.g. CSV) to the Windows clipboard, which is a feature request.
          ✅776646ffff2003cursor incompatabilitiesNOT-A-BUG (RFE); the concrete Windows problems were fixed in 2003 (hand2→IDC_HAND, size/fleur→IDC_SIZEALL, patthoyts)

          Open ticket 776646ffff · 31. Cursor Names · filed 2003-07-24

          ✅ Since triage (updated 2026-11-11)

          Status now: Closed (Fixed)

          • 💬 2026-09-28 serhiy.storchaka: The bugs with the hand2, size and fleur cursors on Windows were fixed long ago. The Windows-only cursor names (no, size_ns, wait, etc.) are still not available on X11, as documented in cursors.n. Portable cursor names mapped to the native cursors of every platform are provided by ttk::cursor (standard, text, link, busy, forbidden, move, hresize, vresize, etc.), but it is not documented, see 835e5e17ec.
          Full verdict

          NOT-A-BUG (RFE); the concrete Windows problems were fixed in 2003 (hand2→IDC_HAND, size/fleur→IDC_SIZEALL, patthoyts)

          • Platform generic
          • Tested main, 8.6, 8.5 on X11. The Windows-only names no, size, size_ne_sw, size_ns, size_nw_se, size_we, starting, uparrow, wait and ibeam all fail with "bad cursor spec". cursors.n documents them as Windows-only.
          • Cause the remaining request is to alias the Windows cursor names on X11 (and make "arrow" the native pointer). That is a feature, not a bug.
          • Prototype none. It would be a new alias table in unix/tkUnixCursor.c. Mappings: ibeam→xterm, wait→watch, size_ns→sb_v_double_arrow, size_we→sb_h_double_arrow, size→fleur, uparrow→sb_up_arrow. There is no good X font cursor for the diagonal ones.
          • Suggested ticket comment The Windows issues were fixed in 2003. The rest asks for cross-platform cursor aliases, which is an RFE. Suggest closing, or moving to a feature request.
          ⏳678939ffff2003reqwidth and reqheight not resetNOT-A-BUG (deliberate behaviour, as serhiy wrote on 2026-09-23)

          Open ticket 678939ffff · 48. Geometry Management · filed 2003-02-02

          ⏳ Since triage (updated 2026-11-11)

          Status now: Pending (Wont Fix)

            Full verdict

            NOT-A-BUG (deliberate behaviour, as serhiy wrote on 2026-09-23)

            • Platform generic
            • Tested main (wt build and the fixed main wish9.1), 8.6, 8.4.20. All print 1x1, then 82x24 (main) or 71x20 (8.4), and the size stays the same after destroy.
            • Cause generic/tkPack.c ArrangePacking() and tkGrid.c ArrangeGrid() leave the container size as-is when the last content window is removed, so another geometry manager can take over.
            • Prototype none
            • Suggested ticket comment Works as designed: when the last content window goes, pack and grid leave the container's requested size alone. Set -width/-height on the container to get it back. Closing as "Works as designed".
            💬456877ffff2001App name/send/multiple display problemNOT-A-BUG (by design; this is really an RFE)

            Open ticket 456877ffff · 58. [send] · filed 2001-08-30

            💬 Since triage (updated 2026-11-11)

            Status now: Open

            • 💬 2026-09-28 serhiy.storchaka: Still the same in 9.1, but this is not about the same names: an application is registered for send only on the display of its main window. A toplevel on other display does not register it there, so applications on that display cannot send to it with any name, and it is not listed by winfo interps there. It can only be reached with send -displayof from a window on its own display. Supporting this would be a new fea…
            Full verdict

            NOT-A-BUG (by design; this is really an RFE)

            • Platform X11
            • Tested code reading only. Reproducing needs two X displays, and the brief allows only one.
            • Cause unix/tkUnixSend.c: Tk_SetAppName registers the interpreter (RegAddName) only in the registry of its main window's display. Opening a toplevel on a second display does not register the app there, so apps on that display cannot find it, whatever the names are. A tk appname -displayof would be a new feature.
            • Prototype none (new feature).
            • Suggested ticket comment An application is registered for send only on its main display (Tk_SetAppName). Sending back from the other display would need a new registration feature such as tk appname -displayof. Suggest closing or turning this into an RFE.
            ⬜220933ffff2000making tearoff-menus transientNOT-A-BUG (RFE; still as described)

            Open ticket 220933ffff · 10. Generic Menus · filed 2000-11-01

            Full verdict

            NOT-A-BUG (RFE; still as described)

            • Platform generic (X11)
            • Tested code reading, library/tearoff.tcl on main.
            • Cause tk::TearOffMenu still runs -tearoffcommand after $menu post, so a wm transient set there comes after mapping, and some WMs ignore it. On win32 Tk itself makes the torn-off menu transient before posting (Bug 3181181); on X11 it does not.
            • Prototype none. Two options: make X11 do the same wm transient before posting as win32 (a behaviour change), or document the workaround wm withdraw $m; wm transient $m .; wm deiconify $m in the tearoffcommand.
            • Suggested ticket comment Still as described. On Windows Tk already makes torn-off menus transient before posting; the same could be done on X11. Workaround: withdraw/deiconify around wm transient in the -tearoffcommand.
            ✅220931ffff2000option get command responding to invalid inputNOT-A-BUG (X resource semantics, as dkf said in 2000)

            Open ticket 220931ffff · 57. [option] · filed 2000-11-01

            ✅ Since triage (updated 2026-11-11)

            Status now: Closed (Fixed)

            • 🔧 2026-09-28 check-in 0fda9c1b82 on doc-option-get-class, in main, 9.0: Fix [220931]: the documentation of "option get" suggested that an entry must match both the name and the class of the option. An entry matches if the last word
            • 🔧 2026-10-25 check-in 0a7003b0ac on trunk, in main, 9.0: Fix for 220931ffff: option get command responding to invalid input
            • 🔧 2026-10-25 check-in b9ea3c8ad4 on core-9-0-branch, in main, 9.0, 8.6: Fix for 220931ffff: option get command responding to invalid input
            • 🔧 2026-10-25 check-in 5e360f9f11 on core-8-6-branch, in 8.6: Fix for 220931ffff: option get command responding to invalid input
            • 💬 2026-09-28 serhiy.storchaka: The behavior is correct and the same in 9.1: an entry matches if the last word of its pattern is either the name or the class, the other one is not checked. But the documentation said that the entry should match both the name and the class. Proposed documentation fix in branch doc-option-get-class.
            • 💬 2026-10-25 nemethi: Serhiy, many thanks for the documentation fix! Merged into main, core-9-0-branch, and core-8-6-branch by commits 0a7003b0, b9ea3c8a, and 5e360f9f.
            Full verdict

            NOT-A-BUG (X resource semantics, as dkf said in 2000)

            • Platform generic
            • Tested 9.1 main: with *Button.background cyan, option get .b background Width returns cyan and option get .b bogus Background returns "". A pattern that matches the option name matches whatever class argument is given.
            • Cause n/a
            • Prototype none. option.n could say that a name match does not depend on the class argument (docs only).
            • Suggested ticket comment Works as designed: an entry that names the option matches whatever class is given. Suggest closing (maybe a doc note in option.n).