QR-Code generator crashes because of Unicode.

[Steve]

Inkscape 0.48+devel r10948
Windows 7 Pro 64bit

Example: "Żółty"

Full message:

Traceback (most recent call last):
  File "render_barcode_qrcode.py", line 1068, in <module>
    e.affect()
  File "C:\Users\User\AppData\Local\Temp\Rar$EX00.713\inkscape\share\extensions\inkex.py", line 227, in affect
    self.effect()
  File "render_barcode_qrcode.py", line 1058, in effect
    grp = inkex.etree.SubElement(self.current_layer, 'g', grp_attribs) #the group to put everything in
  File "lxml.etree.pyx", line 2437, in lxml.etree.SubElement (src/lxml/lxml.etree.c:47345)
  File "apihelpers.pxi", line 210, in lxml.etree._makeSubElement (src/lxml/lxml.etree.c:11391)
  File "apihelpers.pxi", line 205, in lxml.etree._makeSubElement (src/lxml/lxml.etree.c:11327)
  File "apihelpers.pxi", line 283, in lxml.etree._initNodeAttributes (src/lxml/lxml.etree.c:12162)
  File "apihelpers.pxi", line 1242, in lxml.etree._utf8 (src/lxml/lxml.etree.c:19848)
ValueError: All strings must be XML compatible: Unicode or ASCII, no NULL bytes

[su_v (suv-lp)]

Reproduced with Inkscape 0.48+devel r10951 on OS X Lion, Python 2.7.2

Related earlier reports (same error, already fixed):
Bug #320496 “Calendar extension doesn't work”
Bug #366227 “impossible to open / import DXF in linux”:

[Alvin Penner (apenner)]

attached are modified versions of the files for render_barcode_qrcode. This implements character encoding the same as for dxf input.py. These files need to be copied into the share\extensions directory.

[Alvin Penner (apenner)]

also the .inx file

[Alvin Penner (apenner)]

committed to rev 10970

[Steve (stefan-padlo)]

Inkscape 0.48+devel r10990 still doesn't fix this problem.

Testing phrase: późną nocą grań olśniła księżyca jasność

Decoding : ZXing Decoder Online ; Barcode Scanner @ Android

Latin1, CP 1250, CP 1252 : code generated, but polish special signs are encoded in wrong way
UTF8 : crash with message:

Traceback (most recent call last):
  File "render_barcode_qrcode.py", line 1072, in <module>
    e.affect()
  File "C:\Program Files (x86)\Inkscape\share\extensions\inkex.py", line 230, in affect
    self.effect()
  File "render_barcode_qrcode.py", line 1056, in effect
    so.TEXT = unicode(so.TEXT, so.input_encode)
  File "C:\Program Files (x86)\Inkscape\python\Lib\encodings\utf_8.py", line 16, in decode
    return codecs.utf_8_decode(input, errors, True)
UnicodeDecodeError: 'utf8' codec can't decode bytes in position 1-4: invalid data

[su_v (suv-lp)]

> Inkscape 0.48+devel r10990 still doesn't fix this problem.
>
> Testing phrase: późną nocą grań olśniła księżyca jasność
(…)
> UTF8 : crash with message:

Python crash not reproduced with provided sample string and encoding set to 'UTF-8',
using Inkscape 0.48+devel r10992 on OS X 10.7.2 with Python 2.7.2

(produced QR-Code not verified)

[Alvin Penner (apenner)]

unfortunately, I am having the same problem on Windows XP, Inkscape rev 10966.
<offtopic> the crash with utf-8 encoding on Windows is not necessarily a problem, it just means that some other encoding should have been chosen. For example, if you use the calendar extension and use some non-ascii text and use utf-8 encoding you will (on Windows) get a somewhat more graceful message saying 'You must select a correct system encoding', which means the same as the above crash</offtopic>.

 However, the conversion should work properly with some other encoding like Latin 1. With some non-ascii text, like '¼¾èê' the conversion appears to work with Latin 1. This was confirmed by opening the XML editor and checking the inkscape:label property which contains the text. Other text like 'późną' does not appear to be correct. This was confirmed by running the extension Render->Calendar and confirming that this text was not correct. The Calendar extension has a much wider variety of encodings to choose from, but I was not able to get any of them to convert this text correctly. (I didn't try them all, but I tried about half of them)

[Alvin Penner (apenner)]

AImed, just for curiosity, what code page are you using on your computer? Mine is 850. I go to DOS and type in the command chcp and get:

C:\>chcp
Active code page: 850

[Steve (stefan-padlo)]

Aktywna strona kodowa: 852

[Alvin Penner (apenner)]

thanks, I tried using the encoding IBM852 in the Calendar extension to see if that would help, but that also does not work properly with this text, as far as I can tell.

[Steve (stefan-padlo)]

With dev version R11265 the problem still exists.

[Parcly Taxel (parclytaxel-deactivatedaccount)]

Not reproduced on Inkscape 0.91+devel r on Trisquel 7.0. Both "Żółty" and "Zolty" work fine.

[Parcly Taxel (parclytaxel-deactivatedaccount)]

I hasten to add! I was using r13779.

[jazzynico (jazzynico)]

Tested again on Windows XP, Inkscape trunk rev. 13791. I can't reproduce the original issue, but confirm the one reported comment #5.

[Qantas94Heavy (qantas94heavy)]

I can't reproduce this any more with Inkscape 1.0alpha (d666cd1a10, 2019-05-16), both Latin-1 and UTF-8 work correctly.

With Inkscape 0.92.4 can still reproduce incorrect encoding with UTF-8 in comment 5, but does not show any error message. Latin-1 works correctly.

Closed by: https://gitlab.com/Qantas94Heavy