Bug 103967 - DOC import: Empty table cell has wrong font size
Summary: DOC import: Empty table cell has wrong font size
Status: NEW
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Writer (show other bugs)
Version:
(earliest affected)
4.4.0.3 release
Hardware: All All
: medium normal
Assignee: Not Assigned
URL:
Whiteboard:
Keywords: bibisected, bisected, filter:doc, regression
: 111802 (view as bug list)
Depends on:
Blocks: 78756 Regressions-705a8c22
  Show dependency treegraph
 
Reported: 2016-11-17 06:53 UTC by Justin L
Modified: 2019-06-13 16:41 UTC (History)
5 users (show)

See Also:
Crash report or crash signature:


Attachments
blank_edit_topTables.pdf: shows regression - rows are too thin / frames are not equal heights (57.68 KB, application/pdf)
2016-11-17 06:53 UTC, Justin L
Details
TwoPageTable.doc: simple unit test - should be two pages long. (24.00 KB, application/msword)
2016-11-18 08:01 UTC, Justin L
Details
Bold and Italic not imported in this file (19.50 KB, application/msword)
2016-11-28 20:19 UTC, Xisco Faulí
Details
Meeting or Workshop Attend sheet.doc: another doc affected by the same commit. (54.50 KB, application/msword)
2017-02-11 09:20 UTC, Justin L
Details
Meeting or Workshop Attend Sheet Word2007.png: layout with table positionng properties shown (107.45 KB, application/pdf)
2017-02-11 09:28 UTC, Justin L
Details
tdf103967_minimal.doc: minimal version of doc from Comment 5 (26.50 KB, application/msword)
2017-02-14 07:01 UTC, Justin L
Details
Floating Font Size Example REDLINED.doc: modified test from bug 135545 (21.00 KB, application/msword)
2017-10-26 14:01 UTC, Justin L
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Justin L 2016-11-17 06:53:46 UTC
Created attachment 128796 [details]
blank_edit_topTables.pdf: shows regression - rows are too thin / frames are not equal heights

In attachment 99125 [details] from bug 78756, the two boxes at the top of the page should be roughly equal. However, due to a regression in 4.4, the right-side box is smaller than it should be.

In the top-right hand table, the second column is blank.  In the first column, the font size was changed from the default style's 14 to 10.  The second column should be using the default size of 14 - thus making each row taller.
(ignore the fact that the table below is pushed over top of these boxes - that is bug 78756.)

Using Linux bibisect-44max, the regression was found to be caused by
https://cgit.freedesktop.org/libreoffice/core/commit/?h=libreoffice-3-5&id=705a8c226aee3e68db492083b7cf8b704335328b
    commit 705a8c226aee3e68db492083b7cf8b704335328b
    Author: Caolán McNamara <caolanm@redhat.com>
    CommitDate: Wed Sep 24 10:04:48 2014 +0100
        WW8PLCFMan::AdjustEnds deeply flawed concept wrt change tracking
Comment 1 Xisco Faulí 2016-11-17 09:21:13 UTC
Setting this bug as NEW as the commit that introduced the regression has been
identified

Adding Cc: to Caolán McNamara
Comment 2 Justin L 2016-11-18 08:01:48 UTC
Created attachment 128836 [details]
TwoPageTable.doc: simple unit test - should be two pages long.
Comment 3 Xisco Faulí 2016-11-28 20:19:25 UTC
Created attachment 129096 [details]
Bold and Italic not imported in this file

Another file affected by the same commit. In this case, the title Johnston Franklin should be italic and bold
Comment 4 Justin L 2017-02-11 09:20:35 UTC
Created attachment 131087 [details]
Meeting or Workshop Attend sheet.doc: another doc affected by the same commit.

The problem here is in the line containing the text: Dates covered by the Meeting/workshop.  The font switches to Comic Sans MS (the default style's font), but it should be Abyssinica SIL.
Comment 5 Justin L 2017-02-11 09:28:08 UTC
Created attachment 131088 [details]
Meeting or Workshop Attend Sheet Word2007.png: layout with table positionng properties shown

(In reply to Justin L from comment #4)
It appears that the floating table was "defined" somewhere in relation to the line with the problematic font. (In LibreOffice, the floating attribute is removed, and the NB line is placed underneath the table instead of above - which is why I suggest that the table is anchored in that line.)

That suggests that Caolan's comment https://bugs.documentfoundation.org/show_bug.cgi?id=102334#c6 probably applies to this document also.
Comment 6 Justin L 2017-02-11 18:43:30 UTC
(In reply to Justin L from comment #5)
> That suggests that Caolan's comment
> https://bugs.documentfoundation.org/show_bug.cgi?id=102334#c6 probably
> applies to this document also.
I think I properly reverted that old 2011 commit in testing, and it did not resolve the font problem.  Reverting the 2014 regression mentioned in the description however DID resolve the problem.
Comment 7 Justin L 2017-02-14 07:01:28 UTC
Created attachment 131200 [details]
tdf103967_minimal.doc: minimal version of doc from Comment 5
Comment 8 Justin L 2017-02-18 18:56:04 UTC
Giving up on this one. I think the "regression" commit is OK, but probably there is table-related code somewhere that fudges the character attribute start position which needs to be updated to match the new code.
Comment 9 Xisco Faulí 2017-09-05 22:35:39 UTC
*** Bug 111802 has been marked as a duplicate of this bug. ***
Comment 10 Aron Budea 2017-09-06 07:25:28 UTC
Justin, bug 111802 regressed at the same commit, and its bugdoc (attachment 135545 [details]) has no tables, maybe that's easier to work with.
Comment 11 Justin L 2017-10-26 14:01:54 UTC
Created attachment 137302 [details]
Floating Font Size Example REDLINED.doc: modified test from bug 135545

(In reply to Aron Budea from comment #10)
> Justin, bug 111802 regressed at the same commit, and its bugdoc (attachment
> 135545 [details]) has no tables, maybe that's easier to work with.

Thanks, but still failed since my redlined version of the test still has an extra large paragraph. But maybe this is better than nothing:

+    else if (&rDesc == m_pChp && rDesc.nEndPos == m_nLineEnd && rDesc.nEndPos > rDesc.nStartPos && !HasCharSprm(NS_sprm::sprmCDttmRMarkDel).pSprm)
+        rDesc.nEndPos--;            // ... then shorten by one character
Comment 12 QA Administrators 2018-12-24 03:44:13 UTC Comment hidden (spam)