Created attachment 127023 [details] crash document reproduce 1. open attached file 2. select sheet "Monat" 3. select column A and press ⌘1 (menubar Format > Cells) Sadly I am not presented with a crash log. The recovery dialog automatically shows up after the crash, recovery works, repeat and enjoy endless loop.
Version: 5.3.0.0.alpha0+ Build ID: 3e6052ce1b9a24f82358e574de32919fa2b6499f CPU Threads: 4; OS Version: Mac OS X 10.11.6; UI Render: default; TinderBox: MacOSX-x86_64@49-TDF, Branch:master, Time: 2016-08-23_01:56:05 Locale: de-DE (de.UTF-8); Calc: group and persisting with nightly from today
No crash.. maybe I should update to the latest. Win 7 Pro 64-bit Version: 5.3.0.0.alpha0+ (x64) Build ID: 264466bd604d7613aa0a5e4b34c149ad66414580 CPU Threads: 4; OS Version: Windows 6.1; UI Render: default; TinderBox: Win-x86_64@62-TDF, Branch:MASTER, Time: 2016-08-21_21:59:38 Locale: fi-FI (fi_FI); Calc: CL
Might be MacOs only bug, let's specify hardware for the moment.
No crash. Win 7 Pro 64-bit Version: 5.3.0.0.alpha0+ Build ID: 1e84b23839d96068c862e746c9162db79d2c8c62 CPU Threads: 4; OS Version: Windows 6.1; UI Render: default; TinderBox: Win-x86@39, Branch:master, Time: 2016-08-25_05:49:05 Locale: fi-FI (fi_FI); Calc: CL
Comment on attachment 127023 [details] crash document crash not related to any specific document
Created attachment 127031 [details] screencast
Not related to specific document. Screencast added showing instant crash on selecting Format > Cells with empty sheet. @Alex could you try to repor this on your macOS machine?
Version: 5.3.0.0.alpha0+ Build ID: 03f576466c9f110fbef8923604908654d59484a6 CPU Threads: 2; OS Version: Mac OS X 10.11.6; UI Render: default; Locale: fr-FR (fr.UTF-8); Calc: group Confirming, spectacular ! Will try and get a bt. it makes no difference whether the shortcut or the menu Format > Cell is used, both produce a crash.
Created attachment 127032 [details] full bt debug build osx 10.11.6
This assertion fails: include/svl/itemset.hxx: class SVL_DLLPUBLIC SfxItemSet ... template<class T> const T* GetItem(sal_uInt16 nWhich, bool bSearchInParent = true) const { const SfxPoolItem* pItem = GetItem(nWhich, bSearchInParent); const T* pCastedItem = dynamic_cast<const T*>(pItem); assert(!pItem || pCastedItem); // if it exists, must have the correct type return pCastedItem; } Kendy - any thoughts on that ? =)
My guess would be that it is down to commit 200e336d1d5c036b5ecd308d93e4131274b1efb8 (only the change in sc/source/ui/inc/attrdlg.hxx), but of course bibisect / bisect would be most appreciated.
Then again, the commit is a year old (was confused that it was in August too, but did not check the year). Really, need a bibisect here I fear.
Reverting status change as bibisectRequest is not a valid reason for NEEDINFO (and will cause the report to be ignored by bibisecters most likely).
works in Version: 5.3.0.0.alpha0+ Build ID: 40fc2c1a0d2ebdf47131651045107c9d5abb850d CPU Threads: 4; OS Version: Mac OS X 10.12; UI Render: default; TinderBox: MacOSX-x86_64@49-TDF, Branch:master, Time: 2016-10-20_06:13:56 Locale: de-DE (de_DE.UTF-8); Calc: group