Steps to Reproduce: 1. New Calc with two sheets ("Sheet1" and "Sheet2"). 2. Put cursor in Sheet2:B3, then click the "Freeze Columns and Rows" toolbar icon. 3. Put cursor in Sheet2:C4, then try to reference to any cells in Sheet1. (i.e, type in "=", click "Sheet1", then click any cells in Sheet1). Expected Behaviour: In step 3, the formula bar should show the cell reference address when you click any cells in Sheet1. Current Behaviour: Formula bar is always showing "=" without the cell reference. It is noted that it works OK when the columns and rows are not frozen in Sheet2. Version: 5.4.0.0.alpha0+ Build ID: a6c4552e792f6ab20fae83d354b369defd3bb345 CPU Threads: 4; OS Version: Linux 4.4; UI Render: default; VCL: gtk2; Locale: zh-CN (zh_CN.UTF-8); Calc: group Ubuntu 16.04 LTS X64.
I can't reproduce it in Version: 5.4.0.0.alpha0+ Build ID: eb7b03b052ffe8c2c577b2349987653db6c53f76 CPU threads: 1; OS: Windows 6.1; UI render: default; TinderBox: Win-x86@62-merge-TDF, Branch:MASTER, Time: 2017-02-26_22:34:18 Locale: en-GB (es_ES); Calc: group nor Version: 5.4.0.0.alpha0+ Build ID: 880033edde516fc30225005245253293a6a58ba4 CPU Threads: 4; OS Version: Linux 4.8; UI Render: default; VCL: gtk2; Locale: ca-ES (ca_ES.UTF-8); Calc: group To be certain the reported issue is not related to corruption in the user profile, could you please reset your Libreoffice profile ( https://wiki.documentfoundation.org/UserProfile ) and re-test?
Created attachment 131503 [details] backtrace I can also reproduce this with both my own build, and the offical lo-linux-dbgutil-daily bibsect repo. Attached is the backtrace from lo-linux-dbgutil-daily.
(In reply to Kevin Suo from comment #2) The version used in comment 1 is: Version: 5.4.0.0.alpha0+ Build ID: af871d02914c9fc6a08079f67b4af71a198e166a CPU threads: 4; OS: Linux 4.4; UI render: default; VCL: gtk3; Locale: zh-CN (zh_CN.UTF-8); Calc: group Ubuntu 16.04 lts x64.
I run "make clean" and then re-build using the following autogen.input, I still encounter the same problem. Note that the following major lines are the same as the official daily build options: --without-help --without-java --with-system-zlib --without-system-poppler --without-system-openssl --without-system-libpng --without-system-libxml --without-system-jpeg --without-system-jars --without-system-postgresql --without-junit --with-linker-hash-style=both --with-external-thes-dir=/usr/share/mythes --with-external-hyph-dir=/usr/share/hyphen --with-external-dict-dir=/usr/share/hunspell --enable-dbus --enable-extension-integration --enable-odk --enable-kde4 --disable-gtk3 --enable-gio --enable-scripting-beanshell --enable-scripting-javascript --enable-ext-wiki-publisher --enable-ext-nlpsolver --enable-epm --enable-python=internal --disable-online-update --enable-breakpad --disable-dconf --with-vendor=KevinSuo --with-package-format=archive --disable-dependency-tracking --with-export-validation --disable-mergelibs --with-lang=zh-CN zh-TW --disable-firebird-sdbc However I do not encounter the freeze when using the daily dbg-util bibsect repo. Please let me know if more info is needed.
No crash. Kevin: are you saying it crashed with dbgutil build from Feb 27th, but not with a dbgutil build from 28th? Win 7 Pro 64-bit Version: 5.4.0.0.alpha0+ Build ID: eb7b03b052ffe8c2c577b2349987653db6c53f76 CPU threads: 4; OS: Windows 6.1; UI render: default; TinderBox: Win-x86@62-merge-TDF, Branch:MASTER, Time: 2017-02-26_22:34:18 Locale: fi-FI (fi_FI); Calc: CL
Fix title sumnary. Build ID a6c4552e792f6ab20fae83d354b369defd3bb345 is the first build I noticed the bug. But tge actual bad commit may be two says older as I normally rebuild the source every two days. Will try to bisect.
git bisect from source: 1c3a784d38fc5986c78bab8d283c4661347baf43 is the first bad commit commit 1c3a784d38fc5986c78bab8d283c4661347baf43 Author: Justin Luth <justin_luth@sil.org> Date: Mon Feb 20 10:35:25 2017 +0300 tdf#105256 gtk findfocus return only if focused A patch related to tdf#96685 (commit 598e6a024163f1510d076000788b7745625f5ed5) caused the initial regression that Japanese-mozc would freeze(tdf#100903). That patch only modified what happened when the text was marked as focused. However, the regression fix for Japanese bug 100903 (commit 1c81af2c1814e8bd12701f85e09cebf5fe206647) was not placed inside of the "if focused" code, and so it applies to more situations than were broken by the first regression, thus causing new bug 105256. So, applying fix 100903 ONLY when focused solves both problems. Change-Id: I1d8381f2d8cf793d481287ae975ce8a485b6c573 Reviewed-on: https://gerrit.libreoffice.org/34453 Reviewed-by: Justin Luth <justin_luth@sil.org> Tested-by: Justin Luth <justin_luth@sil.org> Reviewed-by: Takeshi Abe <tabe@fixedpoint.jp> :040000 040000 9c2b39204acd2b1fae31c64823c4856154381753 ab6a3f50dbc600872b11d81b91eef7cd350b65d3 M vcl Adding Takeshi Abe to cc as the author Justin Luth is not on bugzilla. Would you please take a look, thanks. git bisect log: # bad: [1324b7f1ce136cb3de617ee72b30468c8717aed9] s/AUTOLAYOUT_TITLE_VERTICAL_OUTLINE_CLIPART/AUTOLAYOUT_TITLE_2VTEXT # good: [872a2d389588e26e0f1f8bbd7a95e5273cd8263f] coverity#1401345 Uninitialized scalar field git bisect start '1324b7f1ce136cb3de617ee72b30468c8717aed9' '872a2d389588e26e0f1f8bbd7a95e5273cd8263f' # good: [b9dd9bbdf248c0b5da981533c5411460bc354fd0] remove unused External from ConvErr enum git bisect good b9dd9bbdf248c0b5da981533c5411460bc354fd0 # bad: [917d5b8b26a7428f7b7dd495a8db14a3ce16aa55] Resolves: tdf#97731 allow status bar to adapt to its own preferred size git bisect bad 917d5b8b26a7428f7b7dd495a8db14a3ce16aa55 # good: [65cfa0eed38ceeb4e96c97b76772bb191129fd83] desktop: unit test to for sd ViewAnnotations command git bisect good 65cfa0eed38ceeb4e96c97b76772bb191129fd83 # good: [b02416ea4742873153227b6eaca00a15e7335987] sd lok: Don't rely on model notify events git bisect good b02416ea4742873153227b6eaca00a15e7335987 # good: [6fbce0cb340448dc0074ff2462babc657da46056] gbuild-to-ideNS: stop printing multi-target warning, again git bisect good 6fbce0cb340448dc0074ff2462babc657da46056 # bad: [3ee1c41ac72aaae9f364708b1c4df58d7643a341] tdf#105931 Add function ROUNDSIG to Calc. git bisect bad 3ee1c41ac72aaae9f364708b1c4df58d7643a341 # bad: [1c3a784d38fc5986c78bab8d283c4661347baf43] tdf#105256 gtk findfocus return only if focused git bisect bad 1c3a784d38fc5986c78bab8d283c4661347baf43 # first bad commit: [1c3a784d38fc5986c78bab8d283c4661347baf43] tdf#105256 gtk findfocus return only if focused
Adding Justin L.
The phrase "calc is freezing" is unfortunate because that suggests that LO has become completely unresponsive (i.e. crashed). Instead, in this case OP simply means that a click is not "registering" and auto-filling the cell reference in the formula bar. I can reproduce this in LO 5.0 (bibisect50max). I can reproduce this under windows with LO 5.1.6.2. Because it also fails in Windows, it cannot have any direct relation to the Unix GTK code change as suggested in comment 7. I'll do some more intensive bibisecting later...
Kevin: could you create a new report for your crashing problem? You can reference to the backtrace as attachment 131503 [details]
(In reply to Justin L from comment #9) The original problem I reported was related to the un-responsive (freeze) of Calc, which is only reproducible in builds after the commit 1c3a784d38fc5986c78bab8d283c4661347baf43 which is on Feb 20 2017, which may not affect other versions such as 5.0 and 5.1. The issue you observed may be another problem. Should I report another bug, or should we stick on this one and wait?
(In reply to Kevin Suo from comment #11) > Should I report another bug, or should we stick on this one and wait? Yes, please create another bug for the crashing problem, since the description of this bug doesn't indicate anything about LO becoming unresponsive. Also, before creating the new bug, please test this against similar 5.1 builds. The chain of changes made relating to the commit referenced in comment 7 all happened starting in 5.2, so first we should confirm that no crash was happening before that time. This code is also related to "accessibility", so state what accessibility program you are using if you know that (perhaps keyboarding software?). I am unable to reproduce the crashing problem. Ubuntu 12.04, with Feb28 2017 master compiled. Accessibility component: ibus / kmfl keyboarding software.
(In reply to Justin L from comment #12) Thanks. Will report new bug. Is keylocks indicator one of the keyboard software you mentioned? I have it installed in my system. See http://packages.ubuntu.com/xenial/indicator-keylock
Confirmed =Sheet bug all the way back to bibisect43all oldest. Steps to confirm: 1.) open calc 2.) click on the plus to create Sheet 2. 3.) click on cell B3 and freeze row/column 4.) still in cell B3, type '=', click on sheet1, and click any cell. Notice that the formula bar does not insert Sheet1.<cell> when the sheet2 cell is in the non-frozen range. (It works for cells in the frozen range, and works when freeze row/column is not active.) Resetting the fields to indicate the description's bug is likely inherited from AOO.
duplicate of #47349 ? *** This bug has been marked as a duplicate of bug 47349 ***