https://bugs.freedesktop.org/show_bug.cgi?id=80149
Priority: medium
Bug ID: 80149
Assignee: [email protected]
Summary: FILEOPEN: Control characters in imported CSV files
cause problems once saved as xlsx
Severity: normal
Classification: Unclassified
OS: Windows (All)
Reporter: [email protected]
Hardware: Other
Whiteboard: BSA
Status: UNCONFIRMED
Version: 4.2.4.2 release
Component: Spreadsheet
Product: LibreOffice
Created attachment 101245
--> https://bugs.freedesktop.org/attachment.cgi?id=101245&action=edit
minimal csv file containing a 0x16 character
Problem description:
Steps to reproduce:
1. Import a CSV file that includes a low-value control character (e.g. 0x16)
2. Save resulting spreadsheet in xlsx format
3. Attempt to re-open spreadsheet in LibreOffice or Excel
Current behavior:
The resulting .xlsx file is treated as corrupt (invalid UTF8) by Excel.
LibreOffice truncates columns after the corrupted cell when the xlsx file is
reloaded.
Expected behavior:
CSV import filter could reject the file; could strip out control characters or
XLSX export could use an encoding that coped with the corrupt characters.
Although the initial csv file is obviously malformed, such corrupt data exists
in the wild and is hard to detect. LibreOffice's silent truncation of columns
after the corrupt cell is problematic as it can easily be overlooked.
(in my original example rows after the corrupt cell stayed but with right-hand
columns missing - no error was reported by LibreOffice)
Operating System: Windows 7
Version: 4.2.4.2 release
--
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
Libreoffice-bugs mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs
