Bug in category file import

[rkg82]

If the first line in a plain text file used to import categories starts with a string enclosed with double quotes because of a literal comma in the string, as specified in v. 8 documentation, the double quotes are not recognized, and the import incorrectly splits the string into two separate strings, and text string for the third field is dropped.

Subsequent lines in the import file whose first field is enclosed with double quotes, are unaffected, and are imported correctly.

Content of the import file:

"Block diagrams, QITC verification",*.vsd,Visio files
"Bugzilla, QITC",/Excerpts/Bugzilla/;\Excerpts\Bugzilla\,Parses local Bugzilla bug list report file
"Class reference, Doxygen, QNoC 4.6",/doxygen/nttp4.6-dev/,TWT IIS server

This bug may affect earlier versions as well.

[Ray]

I tested this out in V8 build 1007 (latest as of post) and I could not produce the problem you described.

As this code has not changed for probably 10 years, it's unlikely to be due to older versions or builds.

As I simply copied and pasted the example from your forum post, it's possible your import text file is actually different. One common problem is when people use a word processor application such as Microsoft Word, which may have utilised "smart" or "curly" quotes, and your text editor may have carried this over when you copied and pasted to the text file. When it was posted into the forum, it's possible the forum converted it back to normal (not curly) quotes and thus it worked just fine.

So make sure your quote characters are actually the standard ones on your keyboard (sometimes this can be obscured by your system font choice). If you still have trouble, e-mail us a copy of the text file in question and we can take a closer look.

[rkg82]

Attached is the test file I used for imports. It's a UTF-8 with BOM file generated and edited in my text editor, EditPad Pro. The characters appear to me to be "straight" quotes. Display font is the standard Windows 10 "Consolas". Also attached is a screen shot of the result in the Zoom search GUI and dialog box, where the issue is apparent.

[Ray]

It's the UTF-8 BOM that's causing trouble. The Import categories feature isn't handling the UTF-8 BOM.

When I resaved the text file with the BOM removed, it imported in fine.

Will add that to the list of things todo.

[rkg82]

Thanks for the workaround. Looking forward to the fix. In the interim, perhaps you could add a note to the documentation about this?

Also, would it be possible to confirm if any of the other file import features might be affected? I haven't noticed any issues, but I haven't tested this either.

[Ray]

The Recommended Links import feature and Synonyms import feature also have this same behaviour.
It will be fixed in the next public release (V8.0 build 1009).

[rkg82]

Good news, thanks Ray and the development team.