GNOME Bugzilla – Bug 66808
Text file import breaks on unknown characters
Last modified: 2004-12-22 21:47:04 UTC
The text file import will ignore any data in a file after a NULL byte is encountered, plus if any character not in the current locale is found it will refuse to load the file. In the latter case an option to ignore the issue and continue would be ideal, and indicating the line number would help a little.
We are currently in a hard freeze with a 1.0 release pending in the next couple of weeks. There are several known issues like this in the importer that we'll address after porting to gnome2 and switching to utf8 internals.
What is the usefulness of being able to import files that have null bytes in them. I understand the need to be able to import files that have been created in any encoding, but are you suggesting that gnumeric ought to import some arbitrary byte files as strings?
The case in point were CSV files with characters not in the current locale, and in one case with a null byte.
- CSV files with characters not in the current locale I am in the process of fixing this in cvs head. (In fact it has already been fixed for csv/tsv non-configurable import: you can choose the encoding of the imported file.) - in one case with a null byte Since I am not aware of any encoding with characters with a leading null byte, this should be a case of multi byte characters with a second or later byte being null. In this case it will be or has been fixed as part of the first item.
text import in cvs head now provides the possibility of selecting the encoding of the to be imported file. This should fix this problem.