diff options
| author | Stephan Bergmann <sbergman@redhat.com> | 2019-05-31 15:37:09 +0200 |
|---|---|---|
| committer | Stephan Bergmann <sbergman@redhat.com> | 2019-05-31 22:27:33 +0200 |
| commit | de4d296619b978ec303f1d7b1e2c78e13fa7a512 (patch) | |
| tree | d62248b1fe005bde45500c75f292f66706b5adc5 /sc | |
| parent | a7f8047f917c0ff45c7bbe76b9f129fbf0d53f10 (diff) | |
Avoid overflow in ScColumn::GetOptimalColWidth
With -fsanitize=float-cast-overflow, opening csv/fdo40680-1.csv as obtained by
bin/get-bugzilla-attachments-by-mimetype (i.e., the attachment at
<https://bugs.documentfoundation.org/show_bug.cgi?id=40680#c0>) fails with
> sc/source/core/data/column2.cxx:741:53: runtime error: 833163 is outside the range of representable values of type 'unsigned short'
> #0 in ScColumn::GetOptimalColWidth(OutputDevice*, double, double, Fraction const&, Fraction const&, bool, unsigned short, ScMarkData const*, ScColWidthParam const*) const at sc/source/core/data/column2.cxx:741:53 (instdir/program/../program/libsclo.so +0xc9374aa)
> #1 in ScTable::GetOptimalColWidth(short, OutputDevice*, double, double, Fraction const&, Fraction const&, bool, ScMarkData const*, ScColWidthParam const*) at sc/source/core/data/table1.cxx:427:23 (instdir/program/../program/libsclo.so +0xe02deee)
> #2 in ScDocument::GetOptimalColWidth(short, short, OutputDevice*, double, double, Fraction const&, Fraction const&, bool, ScMarkData const*, ScColWidthParam const*) at sc/source/core/data/document.cxx:4242:30 (instdir/program/../program/libsclo.so +0xd18b457)
> #3 in ScDocShell::ConvertFrom(SfxMedium&) at sc/source/ui/docshell/docsh.cxx:1614:53 (instdir/program/../program/libsclo.so +0x110b9ad4)
> #4 in SfxObjectShell::DoLoad(SfxMedium*) at sfx2/source/doc/objstor.cxx:768:23 (instdir/program/libsfxlo.so +0x49d934a)
[...]
where nWidth = 5336 and nPPTX = 0.0647688. Given csv/fdo40680-1.csv has rather
much text in the 17th column, these values do not look completely implausible
---which of course begs the question whether sal_uInt16 is an appropriate data
type here.
But assuming sal_uInt16 is a useful choice, just clamp the calculated width
accordingly.
Change-Id: I9629b45183354d148d572a0850314c9a00548311
Reviewed-on: https://gerrit.libreoffice.org/73273
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins
Diffstat (limited to 'sc')
| -rw-r--r-- | sc/source/core/data/column2.cxx | 4 |
1 files changed, 3 insertions, 1 deletions
diff --git a/sc/source/core/data/column2.cxx b/sc/source/core/data/column2.cxx index a4b52eef7061..a29505584a27 100644 --- a/sc/source/core/data/column2.cxx +++ b/sc/source/core/data/column2.cxx @@ -60,6 +60,7 @@ #include <formula/vectortoken.hxx> #include <algorithm> +#include <limits> #include <memory> #include <numeric> @@ -738,7 +739,8 @@ sal_uInt16 ScColumn::GetOptimalColWidth( if (bFound) { nWidth += 2; - sal_uInt16 nTwips = static_cast<sal_uInt16>(nWidth / nPPTX); + sal_uInt16 nTwips = static_cast<sal_uInt16>( + std::min(nWidth / nPPTX, double(std::numeric_limits<sal_uInt16>::max()))); return nTwips; } else |