|
2026-10-13
| ||
| 19:54 | • Closed ticket [ad458a3a]: Sections consistency assertion fails when switching -elide in the revised text widget plus 4 other changes artifact: accecd43 user: nab | |
| 19:54 | Merge elide0-insert-in-elided-range and elide-walk-skipped-lines (Serhiy Storchaka): fix [05411ee533] and [ad458a3a72] (tests btree-14.121 and 14.120). The tests of [bbaa3c1e5f] btree-14.120 and 14.121 become btree-14.125 and 14.126. check-in: 07716463 user: nab tags: revised_text, tip-466 | |
|
2026-10-10
| ||
| 11:39 | • New ticket [05411ee5] Revised text: inserting text with an -elide 0 tag into an elided range fails an assertion. artifact: c1a932b1 user: serhiy.storchaka | |
| 11:38 | • Closed ticket [194c0b86]: revised_text branch / tableList demo crash plus 5 other changes artifact: 1916d62b user: serhiy.storchaka | |
| 11:36 | • Ticket [73d8d949] tablelist widget displaying a tree structure crashes when run with revised_text status still Open with 3 other changes artifact: b7588b1c user: serhiy.storchaka | |
| 11:36 | • Ticket [ad458a3a] Sections consistency assertion fails when switching -elide in the revised text widget status still Open with 3 other changes artifact: 3c009019 user: serhiy.storchaka | |
| 11:35 | Fix [ad458a3a72]: removing an elide tag from the end of a range elided together with an adjacent range of another tag left lines of that range elided without switches. When updating the elide information, skip a line without the changed tag only if its elide state is the current state. Test btree-14.120. closed check-in: 8d0c6903 user: serhiy.storchaka tags: elide-walk-skipped-lines | |
|
2026-07-08
| ||
| 13:53 | Merge nab-elide-missing-branch: fix [ad458a3a72] - dangling branch/link connections when toggling -elide across interleaved tag ranges check-in: dadebc84 user: nab tags: revised_text, tip-466 | |
|
2026-06-12
| ||
| 15:46 | Fix [ad458a3a72], [73d8d94916], [194c0b8650]: corrupted branch/link structure when toggling -elide across interleaved tag ranges. UpdateElideInfo now reconnects surviving branch/link partners of removed counterparts, and extends its working range across adjacent ranges of the same tag so that intermediate boundaries are revised in the same pass. Also fix an endless loop in FindNextLink (section cursor was never advanced). With regression tests, including RTL content. check-in: f80a5b2d user: nab tags: nab-elide-missing-branch | |
|
2025-02-22
| ||
| 10:48 | • Ticket [194c0b86] revised_text branch / tableList demo crash status still Open with 3 other changes artifact: 1a7ec170 user: fvogel | |
|
2025-02-03
| ||
| 22:59 | • Ticket [ad458a3a] Sections consistency assertion fails when switching -elide in the revised text widget status still Open with 3 other changes artifact: 38b95895 user: fvogel | |
|
2025-02-02
| ||
| 21:19 | • Ticket [ad458a3a]: 3 changes artifact: 77008ff4 user: fvogel | |
| 08:29 | • Ticket [ad458a3a]: 4 changes artifact: 8b571b81 user: fvogel | |
|
2025-02-01
| ||
| 22:51 | • Ticket [ad458a3a]: 3 changes artifact: 4a6e4f96 user: nalingupta | |
| 20:39 | • New ticket [ad458a3a]. artifact: 62b315ed user: nalingupta | |
| Ticket UUID: | ad458a3a72112080ed073c3aa0cc0e1704eea1d3 | |||
| Title: | Sections consistency assertion fails when switching -elide in the revised text widget | |||
| Type: | Bug | Version: | revised_text | |
| Submitter: | nalingupta | Created on: | 2025-02-01 20:39:16 | |
| Subsystem: | 18. [text] | Assigned To: | fvogel | |
| Priority: | 5 Medium | Severity: | Severe | |
| Status: | Closed | Last Modified: | 2026-10-13 19:54:57 | |
| Resolution: | Fixed | Closed By: | nobody | |
| Closed on: | ||||
| Description: |
TIP 466: Revised Implementation of the Text Widget
Sections consistency assertion fails "CheckSections: too few segments in section".
Steps to reproduce bug:
(1) Include fix from ticket bdd1f64523 over code from https://github.com/tcltk/tk/tree/revised_text/generic
(2) Using Debug-Build, run following code
pack [text .t]
.t tag configure tag_ellipse -elide 0
.t tag configure tag_hidden -elide 1
.t insert end "text_ellipse" tag_ellipse
.t insert end "text_hidden" tag_hidden
.t insert end "text_ellipse" tag_ellipse
.t insert end "text_hidden" tag_hidden
.t tag configure tag_ellipse -elide 1
.t tag configure tag_hidden -elide 0
--> assert failure in tkTextBTree.c CheckSections()
Tcl_Panic("CheckSections: too few segments in section");
Analysis (code analysis also below)
--------
Issue: Missing Branch
|
___________ ___________ \ /__
| | | | ' |
seg1--seg2--seg3--seg4--seg5--seg6--seg7--seg8--seg9--seg10--seg11--seg12--seg13
mark brch char link char brch char link char link mark mark char
| | | | | | | | | | | | |
\ / | \ | / | \ / \ \ / /
\ / | \ | / | \ / \ \ / /
section1 section2 section3 section4 section5 section6
(len=2) (len=1) (len=3) (len=1) (len=2) (len=4)
- rootPtr 0x000001567deb80d0 Node *
...
- linePtr 0x000001567debfff0 TkTextLine *
...
+ nextPtr 0x000001567dea6700 TkTextLine *
+ prevPtr 0x0000000000000000 TkTextLine *
- segPtr 0x000001567ded37c0 TkTextSegment *
+ typePtr 0x00007ff923401c30 tkTextLeftMarkType "mark" Tk_SegType *
- nextPtr 0x000001567dec5a80 TkTextSegment *
+ typePtr 0x00007ff9233f9e78 tkTextBranchType "branch" Tk_SegType *
- nextPtr 0x000001567deb58c0 TkTextSegment *
+ typePtr 0x00007ff9233f9e40 tkTextCharType "character" Tk_SegType *
- nextPtr 0x000001567dec6f80 TkTextSegment *
+ typePtr 0x00007ff9233f9eb0 tkTextLinkType "connection" Tk_SegType *
- nextPtr 0x000001567deb1bb0 TkTextSegment *
+ typePtr 0x00007ff9233f9e40 tkTextCharType "character" Tk_SegType *
- nextPtr 0x000001567dee0470 TkTextSegment *
+ typePtr 0x00007ff9233f9e78 tkTextBranchType "branch" Tk_SegType *
- nextPtr 0x000001567deb8250 TkTextSegment *
+ typePtr 0x00007ff9233f9e40 tkTextCharType "character" Tk_SegType *
- nextPtr 0x000001567dea3d00 TkTextSegment *
+ typePtr 0x00007ff9233f9eb0 tkTextLinkType "connection"Tk_SegType *
- nextPtr 0x000001567dec5e80 TkTextSegment *
+ typePtr 0x00007ff9233f9e40 tkTextCharType "character" Tk_SegType *
- nextPtr 0x000001567ded3dc0 TkTextSegment *
+ typePtr 0x00007ff9233f9eb0 tkTextLinkType "connection" Tk_SegType *
- nextPtr 0x000001567dea8510 TkTextSegment *
+ typePtr 0x00007ff923401c68 tkTextRightMarkType "mark" Tk_SegType *
- nextPtr 0x000001567dec1cf0 TkTextSegment *
+ typePtr 0x00007ff923401c68 tkTextRightMarkType "mark" Tk_SegType *
- nextPtr 0x000001567debff70 TkTextSegment *
+ typePtr 0x00007ff9233f9e40 tkTextCharType "character" Tk_SegType *
+ nextPtr 0x0000000000000000 TkTextSegment *
Code analysis
-------------
TkBTreeUpdateElideInfo()
{
...
while () {
...
UpdateElideInfo( firstSegPtr=first-instance-of-"hidden_text", search.segPtr=second-instance-of-"ellipse_text"
UpdateElideInfo()
{
...
if (!(endSegPtr = endSegPtr->nextPtr)) { -->moves endSegPtr to second-instance-of-"hidden_text" line 8673
Root-cause
----------
Moving endSegPtr to second-instance-of-"hidden_text" causes code following this statement to generate inconsistent segment-chain which has a "branch" missing.
| |||
| User Comments: |
nab added on 2026-10-13 19:54:57:
Merged into revised_text with [07716463c0|07716463c0], as you asked. Thank you. serhiy.storchaka added on 2026-10-10 11:36:33: The scripts from this ticket pass on revised_text since [f80a5b2df7], but removing a tag from the end of a range elided together with an adjacent range of another tag still left lines elided without switches:
pack [text .t]; .t debug 1
.t tag configure A -elide 1; .t tag configure B -elide 1
for {set i 1} {$i <= 9} {incr i} {.t insert end "row $i\n"}
.t tag add B 2.0 4.0; .t tag add A 4.0 8.0
.t tag remove B 3.0 6.0
UpdateElideInfo() skipped lines 4 and 5 because they do not carry B, although they are elided by A. Proposed fix in branch elide-walk-skipped-lines, [8d0c690381], test btree-14.120.
fvogel added on 2025-02-03 22:59:14: Code analysis
-------------
When processing the last statement, i.e. '.t tag configure tag_ellipse -elide 1', at the call to UpdateElideInfo():
1. UpdateElideInfo() is first called with tagPtr{name=tag_ellipse, elide=1) firstSegPtr=seg2 ("text_1") lastSegPtr=seg4 ("text_2") reason=ELISION_HAS_BEEN_CHANGED
2. UpdateElideInfo() is then called with tagPtr{name=tag_ellipse, elide=1) firstSegPtr=seg6 ("text_3") lastSegPtr=seg8 ("text_4") reason=ELISION_HAS_BEEN_CHANGED
After execution of the first call to UpdateElideInfo(), the line architecture is correct (there is a newly inserted branch segment: seg13):
__________________ ___________
| | | |
seg1--seg13--seg2--seg4--seg5--seg6--seg7--seg8--seg9--seg10--seg11--seg12
mark brch char char link char brch char link mark mark char
| | | | | | | | | | | |
\ / \ / \ \ / | \ \ / /
\ / \ / \ | / | \ \ / /
section1 section2 section3 section4 section5
(len=2) (len=2) (len=3) (len=1) (len=4)
(text_1text2) (text_3) (text_4) (\n)
(elided) (elided)
After execution of the second call to UpdateElideInfo(), we get:
_____________<???>____________
| |
seg1--seg13--seg2--seg4--seg6--seg8--seg9--seg10--seg11--seg12
mark brch char char char char link mark mark char
| | | | | | | | | |
\ / \ | | / \ \ / /
\ / \ | | / \ | | /
section1 section2 section3
(len=2) (len=4) (len=4)
(text_1text2text_3text_4) (\n)
(elided)
The connection between the branch and the link is missing!
The seg13 branch points to something that is not a link (typePtr is NULL)
The link points to something that is not a branch (typePtr is NULL)
--> KABOOM!
We're narrowing down. Stay tuned to our next issue!
fvogel added on 2025-02-02 21:19:55: An analysis start (adopting the formalism you have used):
I'm using a slightly modified script (this eases debugging):
package require Tk
pack [text .t]
.t tag configure tag_hidden -elide 1
.t insert end "text_1" tag_ellipse
.t insert end "text_2" tag_hidden
.t insert end "text_3" tag_ellipse
.t insert end "text_4" tag_hidden
.t tag configure tag_ellipse -elide 1 ; # assertion fails "BranchCheckProc: connection is not linked" WHEN tkBTreeDebug = 1 ONLY
Snapshot before last statement
------------------------------
linePtr = 0x00000209288e0550
linePtr->numBranches = 2
linePtr->numLinks = 2
linePtr->segPtr = seg1Ptr = 0x00000209288c1560
linePtr->segPtr->sectionPtr 0x0000020928885570
linePtr->lastPtr = 0x00000209288f24a0
addr type value
seg1Ptr->prevPtr=NULL
seg1: 0x00000209288c1560 tkTextLeftMarkType
seg2: 0x00000209289a7be0 tkTextCharType "text_1"
seg3: 0x00000209289a7b60 tkTextBranchType
seg4: 0x00000209289a7ce0 tkTextCharType "text_2" "tag_hidden"->elide=1
seg5: 0x00000209289a7960 tkTextLinkType
seg6: 0x00000209289982a0 tkTextCharType "text_3"
seg7: 0x0000020928958bd0 tkTextBranchType
seg8: 0x00000209288cca00 tkTextCharType "text_4" "tag_hidden"->elide=1
seg9: 0x000002092aa29040 tkTextLinkType
seg10: 0x000002092831c020 tkTextRightMarkType
seg11: 0x000002092831b0a0 tkTextRightMarkType
seg12: 0x00000209288f24a0 tkTextCharType "\n"
seg12->nextPtr=NULL
sec1Ptr->prevPtr=NULL
sec1: 0x0000020928885570 size=6 len=3 linePtr=0x00000209288e0550 seg1+seg2+seg3
sec2: 0x00000209288294c0 size=6 len=1 linePtr=0x00000209288e0550 seg4
sec3: 0x00000209288bbbc0 size=6 len=3 linePtr=0x00000209288e0550 seg5+seg6+seg7
sec4: 0x000002092889fb10 size=6 len=1 linePtr=0x00000209288e0550 seg8
sec5: 0x000002092888dc90 size=1 len=4 linePtr=0x00000209288e0550 seg9+seg10+seg11+seg12
sec5->nextPtr=NULL
Each branch points to its corresponding link, and each link points to its corresponding branch.
Representing the situation graphically, it is easy to see that everything is correct at this point:
___________ ___________
| | | |
seg1--seg2--seg3--seg4--seg5--seg6--seg7--seg8--seg9--seg10--seg11--seg12
mark char brch char link char brch char link mark mark char
| | | | | | | | | | | |
\ | / | \ | / | \ \ / /
\ | / | \ | / | \ \ / /
section1 section2 section3 section4 section5
(len=3) (len=1) (len=3) (len=1) (len=4)
(elided) (elided)
Next step is code analysis when processing the last statement, i.e. '.t tag configure tag_ellipse -elide 1', at the call to UpdateElideInfo(). Stay tuned.
fvogel added on 2025-02-02 08:29:56: Thank you for the short script, this is so useful! It makes a world of difference when trying to solve the problem. One useful trick: When you build the revised text widget with int tkBTreeDebug = 1;in tkTextBtree.c:135, then quite some more consistency checks are exercised at runtime. In the present case, when running your provided script one gets a panic in BranchCheckProc(): "BranchCheckProc: connection is not linked". This happens before the reported CheckSections panic triggered by the second tag configure command from your script. It happens when running the first tag configure command, so it is very interesting in narrowing down the issue. And this time it looks very much like [73d8d94916] (same panic) for which we didn't yet have a short reproducible script. Great! | |||
