| Summary: | Okular crashes when trying to modify a form-pdf | ||
|---|---|---|---|
| Product: | [Applications] okular | Reporter: | Emanuele Spirito <emaster987> |
| Component: | general | Assignee: | Okular developers <okular-devel> |
| Status: | RESOLVED UPSTREAM | ||
| Severity: | crash | CC: | aacid, yurchor |
| Priority: | NOR | Keywords: | drkonqi |
| Version First Reported In: | 1.11.1 | ||
| Target Milestone: | --- | ||
| Platform: | Fedora RPMs | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
| Attachments: | The pdf-form written in the bug description | ||
|
Description
Emanuele Spirito
2021-04-11 18:22:03 UTC
can we have that file? Created attachment 137525 [details]
The pdf-form written in the bug description
To reproduce the bug: open the PDF-form, click on See forms, then you will see in the first page, above on the right, the section "Death saves, success and failure" which becomes editable, so you can select and unselect the squares. Furthermore, all the sections of the sheet are text boxes forms, but they are not editable with okular.
Of course. I forgot it. So, if you click in the squares following what described here above, okular will crash Doesn't crash here :/ 2 things: a) Which poppler version are you running? b) can you run in a terminal valgrind okular thefile.pdf and try to make it crash? Even if it doesn't crash, can you attach the valgrind output? (In reply to Albert Astals Cid from comment #4) > Doesn't crash here :/ I can reproduce this with poppler 20.12.1, so it must have been fixed somewhere in between. To Albert: (0) Even without reproducing that specific bug, can you see all the form-fields? As descripted in the description of the attachment mostly of the pdf is editable (1) Poppler version I'm using: Poppler 0.90.0 (2) I reproduced the bug and the outcome was the same. Here it is the valgrind output: [espirito@localhost Scrivania]$ valgrind okular Presticulo_47775440.pdf ==3701== Memcheck, a memory error detector ==3701== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al. ==3701== Using Valgrind-3.16.1 and LibVEX; rerun with -h for copyright info ==3701== Command: okular Presticulo_47775440.pdf ==3701== ==3701== Conditional jump or move depends on uninitialised value(s) ==3701== at 0x21A40C61: ??? ==3701== by 0x1DEDA5A7: ??? ==3701== ==3701== Conditional jump or move depends on uninitialised value(s) ==3701== at 0x21A40C61: ??? ==3701== by 0x1DE8EA27: ??? ==3701== ==3701== Conditional jump or move depends on uninitialised value(s) ==3701== at 0x1C28316A: ??? (in /usr/lib64/qt5/plugins/okularpart.so) ==3701== by 0x5133573: QWidget::event(QEvent*) (in /usr/lib64/libQt5Widgets.so.5.15.2) ==3701== by 0x51DC6F1: QFrame::event(QEvent*) (in /usr/lib64/libQt5Widgets.so.5.15.2) ==3701== by 0x5F53994: QCoreApplicationPrivate::sendThroughObjectEventFilters(QObject*, QEvent*) (in /usr/lib64/libQt5Core.so.5.15.2) ==3701== by 0x50F1EB1: QApplicationPrivate::notify_helper(QObject*, QEvent*) (in /usr/lib64/libQt5Widgets.so.5.15.2) ==3701== by 0x5F53BD7: QCoreApplication::notifyInternal2(QObject*, QEvent*) (in /usr/lib64/libQt5Core.so.5.15.2) ==3701== by 0x512985F: QWidgetPrivate::setGeometry_sys(int, int, int, int, bool) (in /usr/lib64/libQt5Widgets.so.5.15.2) ==3701== by 0x512A6DF: QWidget::setGeometry(QRect const&) (in /usr/lib64/libQt5Widgets.so.5.15.2) ==3701== by 0x51DDB76: QAbstractScrollAreaPrivate::layoutChildren_helper(bool*, bool*) (in /usr/lib64/libQt5Widgets.so.5.15.2) ==3701== by 0x51DE01E: QAbstractScrollAreaPrivate::layoutChildren() (in /usr/lib64/libQt5Widgets.so.5.15.2) ==3701== by 0x51DF633: QAbstractScrollArea::event(QEvent*) (in /usr/lib64/libQt5Widgets.so.5.15.2) ==3701== by 0x50F1EC2: QApplicationPrivate::notify_helper(QObject*, QEvent*) (in /usr/lib64/libQt5Widgets.so.5.15.2) ==3701== ==3701== Thread 12 Okular::PixmapGe: ==3701== Use of uninitialised value of size 8 ==3701== at 0x2B2343C8: GfxImageColorMap::getRGBXLine(unsigned char*, unsigned char*, int) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B2DA1E5: SplashOutputDev::imageSrc(void*, unsigned char*, unsigned char*) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B2EFF21: Splash::scaleImageYdXd(bool (*)(void*, unsigned char*, unsigned char*), void*, SplashColorMode, int, bool, int, int, int, int, SplashBitmap*) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B2F46D3: Splash::scaleImage(bool (*)(void*, unsigned char*, unsigned char*), void*, SplashColorMode, int, bool, int, int, int, int, bool, bool) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B2F95E0: Splash::drawImage(bool (*)(void*, unsigned char*, unsigned char*), void (*)(void*, SplashBitmap*), void*, SplashColorMode, bool, int, int, double*, bool, bool) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B2E0757: SplashOutputDev::drawSoftMaskedImage(GfxState*, Object*, Stream*, int, int, GfxImageColorMap*, bool, Stream*, int, int, GfxImageColorMap*, bool) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B2174C3: Gfx::doImage(Object*, Stream*, bool) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B211012: Gfx::opXObject(Object*, int) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B206BD1: Gfx::go(bool) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B207593: Gfx::display(Object*, bool) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B26B362: Page::displaySlice(OutputDev*, double, double, int, bool, bool, int, int, int, int, bool, bool (*)(void*), void*, bool (*)(Annot*, void*), void*, bool) (in /usr/lib64/libpoppler.so.101.0.0) ==3701== by 0x2B0591CB: Poppler::Page::renderToImage(double, double, int, int, int, int, Poppler::Page::Rotation, void (*)(QImage const&, QVariant const&), bool (*)(QVariant const&), bool (*)(QVariant const&), QVariant const&) const (in /usr/lib64/libpoppler-qt5.so.1.25.0) ==3701== QSocketNotifier: Invalid socket 8 and type 'Read', disabling... QSocketNotifier: Invalid socket 9 and type 'Read', disabling... ==3627== ==3627== HEAP SUMMARY: ==3627== in use at exit: 52,897,422 bytes in 115,284 blocks ==3627== total heap usage: 2,143,872 allocs, 2,028,588 frees, 1,459,454,060 bytes allocated ==3627== ==3627== LEAK SUMMARY: ==3627== definitely lost: 8 bytes in 1 blocks ==3627== indirectly lost: 32 bytes in 1 blocks ==3627== possibly lost: 1,245,720 bytes in 4,401 blocks ==3627== still reachable: 51,651,662 bytes in 110,881 blocks ==3627== of which reachable via heuristic: ==3627== newarray : 456 bytes in 13 blocks ==3627== multipleinheritance: 142,456 bytes in 173 blocks ==3627== suppressed: 0 bytes in 0 blocks ==3627== Rerun with --leak-check=full to see details of leaked memory ==3627== ==3627== Use --track-origins=yes to see where uninitialised values come from ==3627== For lists of detected and suppressed errors, rerun with: -s ==3627== ERROR SUMMARY: 3479134 errors from 5 contexts (suppressed: 0 from 0) ==3701== Thread 1: ==3701== Invalid read of size 8 ==3701== at 0x2B04CBF1: Poppler::Document::formCalculateOrder() const (in /usr/lib64/libpoppler-qt5.so.1.25.0) ==3701== by 0x2AFB8D4E: ??? (in /usr/lib64/qt5/plugins/okular/generators/okularGenerator_poppler.so) ==3701== Address 0x50 is not stack'd, malloc'd or (recently) free'd ==3701== KCrash: Application 'okular' crashing... KCrash: Attempting to start /usr/libexec/drkonqi [1] Uscita 253 valgrind okular Presticulo_47775440.pdf [2]+ Fermato valgrind okular Presticulo_47775440.pdf To Albert and Yuri: After this valgrind output of the bug: do you think it is just due to my old version of Poppler? If so, how can I update it to a version which is supported and stable? (In reply to Emanuele Spirito from comment #7) > To Albert and Yuri: > After this valgrind output of the bug: do you think it is just due to my old > version of Poppler? If so, how can I update it to a version which is > supported and stable? I can confirm that the file does not crash Okular with poppler 21.04 (the current version from git). I think that the most constructive way to fix it would be filing bug report in Fedora Bugzilla with the hash of the commit that fixes the bug (Albert knows better as the fix was probably made by him). The wrong way which breaks Inkscape, Okular and other packages, as they should be recompiled, is building it from scratch: http://www.linuxfromscratch.org/blfs/view/svn/general/poppler.html (In reply to Yuri Chornoivan from comment #8) > I can confirm that the file does not crash Okular with poppler 21.04 (the > current version from git). I think that the most constructive way to fix it > would be filing bug report in Fedora Bugzilla with the hash of the commit > that fixes the bug (Albert knows better as the fix was probably made by him). Ok, so, I wait for Albert reply, then I open a bug report in bugzilla for the poppler component with: - description of the issue of opening that pdf (ecc...); - the current bug report; - I'll say that the latest git poppler version doesn't have that problem; - I put the commit of the fix you cited above. Do I have your permission to put you two in cc list for that bug in bugzilla? > The wrong way which breaks Inkscape, Okular and other packages, as they > should be recompiled, is building it from scratch: > > http://www.linuxfromscratch.org/blfs/view/svn/general/poppler.html Ok, I won't do that! I don't really have the time to figure out which of the commits in poppler fixed this issue, but since both me and Yuri can confirm that with modern versions of it it doesn't crash i'm closing the bug. I understand that's not super helpful for you that still have a crash issue, you may want to try https://flathub.org/apps/details/org.kde.okular that should provide a recent enough okular+poppler combination. |