This is related to the closed issue #536 .
It would be very useful if the user tries to run "pandoc input.odt [-o output.txt]" to specify that
"Pandoc can convert to ODT, but not from ODT; to convert from ODT, we recommend using LibreOffice to convert the file to HTML; then use pandoc on the result."
The previous error message "hGetContents: invalid argument (invalid byte sequence)" and the current error message (as per changelog) "invalid UTF-8 stream" is not practical for end users to diagnose the issue when they encounter it.
Thanks
Edit: mention of changelog item for c80ffe9#diff-ba7984d3c29de05a63bd95889f1bac2f and its new error message
This is related to the closed issue #536 .
It would be very useful if the user tries to run "pandoc input.odt [-o output.txt]" to specify that
"Pandoc can convert to ODT, but not from ODT; to convert from ODT, we recommend using LibreOffice to convert the file to HTML; then use pandoc on the result."
The previous error message "hGetContents: invalid argument (invalid byte sequence)" and the current error message (as per changelog) "invalid UTF-8 stream" is not practical for end users to diagnose the issue when they encounter it.
Thanks
Edit: mention of changelog item for c80ffe9#diff-ba7984d3c29de05a63bd95889f1bac2f and its new error message