So again a tracker bug for all bugzilla documents that crash calc in the automatic import test. First one: Bug 45266 in Pivot table code in dbgutil build
Created attachment 71885 [details] fdo41391
Created attachment 71886 [details] fdo44758
Created attachment 71887 [details] fdo45266
Created attachment 71888 [details] fdo45268
Created attachment 71889 [details] fdo45276
Created attachment 71890 [details] fdo54638
Created attachment 71891 [details] fdo55022
Created attachment 71892 [details] fdo55022-2
Created attachment 71893 [details] fdo55174
Created attachment 71894 [details] gnome519788
Created attachment 71895 [details] gnome627150
Created attachment 71896 [details] gnome627420
Created attachment 71897 [details] kde245919
Created attachment 71898 [details] ooo74264
Initial list of all crasher docs. We'll update the list as we go fixing. * fdo45266 * fdo41391 * fdo44758 * fdo45266 * fdo45268 * fdo45276 * fdo54638 * fdo55022 * fdo55022-2 * fdo55174 * gnome519788 * gnome627150 * gnome627420 * kde245919 * ooo74264
I hope someone can provide a stack trace on fdo#45266 as I don't use dbgutil build. Else I'll have to set that aside for now.
Fixed fdo#41391, plus removed duplicated fdo#45266 entry (there was two of them). * fdo41391 * fdo44758 * fdo45266 * fdo45268 * fdo45276 * fdo54638 * fdo55022 * fdo55022-2 * fdo55174 * gnome519788 * gnome627150 * gnome627420 * kde245919 * ooo74264
Ugh. Remove fdo41391 for real. * fdo44758 * fdo45266 * fdo45268 * fdo45276 * fdo54638 * fdo55022 * fdo55022-2 * fdo55174 * gnome519788 * gnome627150 * gnome627420 * kde245919 * ooo74264
Kohei Yoshida committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=79f1ef44e77074d8f5a1d32e0447118e5b9c4e70 fdo#58539: Resizing matrix should also resize the flag storage too. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Crasher with the fdo45266 doc should be fixed now. Confirmation appreciated. * fdo44758 * fdo45268 * fdo45276 * fdo54638 * fdo55022 * fdo55022-2 * fdo55174 * gnome519788 * gnome627150 * gnome627420 * kde245919 * ooo74264
Kohei Yoshida committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=c27e334661b62897a81a450ea9188f80e053261e fdo#58539: Avoid throwing exception not specified in the signature. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Markus Mohrhard committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=fa1b00095011def26cc291e74948817632e3545d don't overflow column number during import, fdo#58539 The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Markus Mohrhard committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=5693934456f917861af7cb02734578b049a46940 limit the number of imported digits, fdo#58539 The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
* fdo44758 * fdo45268 * fdo45276 * fdo54638 * fdo55022 * fdo55022-2 * fdo55174 * gnome519788 * gnome627150 * ooo74264 kde245919 and gnome627420 should be fixed now.
Kohei Yoshida committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=30285360e5d1fbb14bb6bf54e55a3a9f9b7619e7&h=libreoffice-4-0 fdo#58539: Resizing matrix should also resize the flag storage too. It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Kohei Yoshida committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=b060b43f093dce23222fd99375b1c6bd433703d9&h=libreoffice-4-0 fdo#58539: Avoid throwing exception not specified in the signature. It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
I'm working on the fdo44758 document at the moment.
Created attachment 72417 [details] reduced version of fdo44758 I've narrowed what causes the fdo44758 document to crash down to a single cell, in this reduced version. I still haven't figured out how to properly fix this since it's a very convoluted formula... Perhaps Eike can take a look at this?
The fdo45268 document (.xlsx) doesn't crash for me. Perhaps we can take this off the list (?)
fdo45276 doesn't crash for me either.
fdo54638 is an open performance issue which has yet to be fixed. Maybe we should take it off the list as well and handle this document in Bug 54638 instead.
These two: fdo55022, fdo55022-2 don't crash on me. Hmm...
Kohei Yoshida committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=7a8a3e46978b3397ca8aba38917bd5dbc808c9b8 fdo#58539: Properly initialize a member pointer value. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
fdo55174 is now fixed. * fdo44758 * fdo45268 (?) * fdo45276 (?) * fdo54638 (open perf issue) * fdo55022 (?) * fdo55022-2 (?) * gnome519788 * gnome627150 * ooo74264
Hmm. ooo74264 seems to be a writer document with an .ods extension (!) * fdo44758 * fdo45268 (?) * fdo45276 (?) * fdo54638 (open perf issue) * fdo55022 (?) * fdo55022-2 (?) * gnome519788 * gnome627150 * ooo74264 (writer doc)
Kohei Yoshida committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=eee87edf551f7a9f8c4da4267e5710edafd1725d&h=libreoffice-4-0 fdo#58539: Properly initialize a member pointer value. It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Kohei Yoshida committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=f204cd9b52ec7af3e0d20f5c88f09737b38b9e7d&h=libreoffice-4-0 fdo#58539: Check for mxValueLink being NULL before dereferencing. It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Kohei Yoshida committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=329f3cd02546dfe58bf00f6ad3b71bc84a8d4320 fdo#58539: Check for mxValueLink being NULL before dereferencing. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
gnome519788 is now fixed. * fdo44758 * fdo45268 (?) * fdo45276 (?) * fdo54638 (open perf issue) * fdo55022 (?) * fdo55022-2 (?) * gnome627150 * ooo74264 (writer doc)
Markus Mohrhard committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=74de75156f1ac6de427b62f000c6dcd248f914f6&h=libreoffice-4-0 don't overflow column number during import, fdo#58539 It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Markus Mohrhard committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=f373868e16a1d83ceda531e4965cb00eaa523b8a&h=libreoffice-4-0 limit the number of imported digits, fdo#58539 It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
(In reply to comment #32) > These two: fdo55022, fdo55022-2 don't crash on me. Hmm... At least according to the output in the console this looks like another document that was crashing due to the graphic object lifetime problem that dtardon fixed.
* fdo45268 (?) * fdo45276 (?) Both seem to be another case of the already fixed pivot table problem.
Remaining files: * fdo44758 * fdo54638 (open perf issue) * gnome627150 * ooo74264 (writer doc)
Markus Mohrhard committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=dfdbf14006d4baf822f399821dfb9d5412f8ab9e&h=libreoffice-4-0 make sure the column index is valid, fdo#58539 It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Markus Mohrhard committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=5bc348ae930b57f546238a8d24efebdd0972753c make sure the column index is valid, fdo#58539 The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Markus Mohrhard committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=c12df14aedbfcd9bd2bfd6dcc1e18524322bb939&h=libreoffice-4-0 workaround problems with row import, fdo#58539 It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Markus Mohrhard committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=ec6b0cb9da208d520972b3f305ecd79fdddb9023&h=libreoffice-4-0 don't overflow SCCOL during repeated cell import, fdo#58539 It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Markus Mohrhard committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=3e85d9c62c69ee39dd4c6ae7cd556fdf53f5e01d workaround problems with row import, fdo#58539 The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Markus Mohrhard committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=94f7417b038fad6f2a92d1c7c5f8f00c3e64711b don't overflow SCCOL during repeated cell import, fdo#58539 The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
All three crashes with gnome#627150 should be fixed now. ooo74264 seems to have crashed because I used the filters unit test which does not have the corresponding uno components. The remaining two problems are: * fdo44758 * fdo54638 (open perf issue)
Created attachment 72493 [details] patch for fdo#44758 This patch fixes the crash and looks like a potential bug fix for fdo#44758. @Eike: Could you have a look at this one?
(In reply to comment #52) > Created attachment 72493 [details] [review] > patch for fdo#44758 > > This patch fixes the crash and looks like a potential bug fix for fdo#44758. > > @Eike: Could you have a look at this one? Actually that's what I'd come up with, which should indeed fix the crash itself. But the concern here is that why the matrix become empty? The matrix there should at least contain one element, so neither the column nor row dimension should become zero there (it should be a 1 by 1 matrix, not empty matrix). That's what I meant to say in Comment 28.
> Actually that's what I'd come up with, which should indeed fix the crash > itself. But the concern here is that why the matrix become empty? The > matrix there should at least contain one element, so neither the column nor > row dimension should become zero there (it should be a 1 by 1 matrix, not > empty matrix). > > That's what I meant to say in Comment 28. Ah ok. But in the end we either need this patch or we can remove the if clause and ignore it. Right now every time we reach this PushIllegalArgument() call we crash later.
(In reply to comment #54) > > Actually that's what I'd come up with, which should indeed fix the crash > > itself. But the concern here is that why the matrix become empty? The > > matrix there should at least contain one element, so neither the column nor > > row dimension should become zero there (it should be a 1 by 1 matrix, not > > empty matrix). > > > > That's what I meant to say in Comment 28. > > Ah ok. But in the end we either need this patch or we can remove the if > clause and ignore it. Right now every time we reach this > PushIllegalArgument() call we crash later. Sure. We do need this change. There is no dispute about it. We just also need to at least find out why the matrix is empty since that may mean there is a problem lurking elsewhere..
Markus Mohrhard committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=9ddd869873603f929d7c342d62a8d52d19de83ec Revert "workaround problems with row import, fdo#58539" The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Markus Mohrhard committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=734c469a91fad731715870e30886de9bb420c5c6&h=libreoffice-4-0 Revert "workaround problems with row import, fdo#58539" It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Let's remove fdo54638 off the list. I'll tackle that one as a separate performance bug for 4.1. We can't do anything for it for 4.0, I'm afraid. So, what remains is only this: * fdo44758
(In reply to comment #58) > Let's remove fdo54638 off the list. I'll tackle that one as a separate > performance bug for 4.1. We can't do anything for it for 4.0, I'm afraid. > > So, what remains is only this: > > * fdo44758 I'm still not through with this, but at least I know what's causing the crash. The type of the formula cell result is set to hybrid string, but it has to be set to matrix type for the interpreter to work properly. A quick test hack suggests that if we find a way to correctly load cached matrix values as matrix type (not as hybrids) the crash will go away. Alternatively, if we just skip loading any cached values for matrix range (or just the upper-left cell of the matrix range), the crash will go away too. I'll continue tomorrow what's the best option for this.
Kohei Yoshida committed a patch related to this issue. It has been pushed to "master": http://cgit.freedesktop.org/libreoffice/core/commit/?id=b6fbc8642928893ca608a35b802032cfe42e40bb fdo#58539: Correctly set cached matrix formula result. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
I pushed my last change to gerrit for review for the 4.0 branch. Once that's cleared and the change goes into 4.0 (unless it's objected), we can close this bug.
Kohei Yoshida committed a patch related to this issue. It has been pushed to "libreoffice-4-0": http://cgit.freedesktop.org/libreoffice/core/commit/?id=97a10ade202537a60bf15ad58f2e44cfb785bb3f&h=libreoffice-4-0 fdo#58539: Correctly set cached matrix formula result. It will be available in LibreOffice 4.0. The patch should be included in the daily builds available at http://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: http://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
The last one is in. Let's close this bug.