Inkscape

Extension Text > Extract cannot deal with non-ASCII characters

Bug #1810626 reported by Hachmann on 2019-01-05

This bug affects 1 person

Affects		Status	Importance	Assigned to	Milestone
	Inkscape	Invalid	Low	Qantas94Heavy	Inkscape 0.92.5 "0.92.5"

Bug Description

Tested on 0.92.x as of Dec. 23rd 2018

1. Select the text that contains an ö in the attached SVG file#
2. Apply Extensions > Text > Extract...
3. Get error message:

alkjg aghalh ajgajhajhgha ahgah ahgp
Traceback (most recent call last):
  File "text_extract.py", line 164, in <module>
    e.affect()
  File "/opt/inkscape_0.92.x/share/inkscape/extensions/inkex.py", line 283, in affect
    self.effect()
  File "text_extract.py", line 148, in effect
    self.recurse(deepcopy(self.selected[item[1]]))
  File "text_extract.py", line 156, in recurse
    inkex.errormsg(inkex.etree.tostring(node, method='text').strip())
  File "src/lxml/etree.pyx", line 3342, in lxml.etree.tostring
  File "src/lxml/serializer.pxi", line 103, in lxml.etree._tostring
  File "src/lxml/serializer.pxi", line 75, in lxml.etree._textToString
UnicodeEncodeError: 'ascii' codec can't encode character u'\xf6' in position 0: ordinal not in range(128)

Tags:

Revision history for this message

Hachmann (marenhachmann) wrote on 2019-01-05:

text-ext.svg Edit (6.9 KiB, image/svg+xml)

tags:

added: extensions-plugins

Revision history for this message

Firoz Taverbi (firoztaverbi) wrote on 2019-05-01:

Closing because Inkscape 1.0 has a different bug with Text > Extract, currently being investigated here: https://gitlab.com/inkscape/inbox/issues/409

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

Changed in inkscape:
status:	New → Invalid

Revision history for this message

Hachmann (marenhachmann) wrote on 2019-05-01:

Thanks for looking at this.

I disagree that the other issue makes this one moot, though. I added a comment to the other issue, please verify that it works with öüä etc. also, when that issue is fixed.

Changed in inkscape:
status:	Invalid → New

Revision history for this message

Firoz Taverbi (firoztaverbi) wrote on 2019-05-02:

Hi Hachmann,

As far as I'm aware as Inkscape 1.0 is in active development bugs in 0.92 are no longer being looked into, hence the migration. We've been instructed to close bugs that aren't reproducible in Inkscape 1.0. You can find more info here: http://alpha.inkscape.org/bug-migration/index.html

Changed in inkscape:
status:	New → Invalid

Revision history for this message

Hachmann (marenhachmann) wrote on 2019-05-02:

Hi Firoz, I'm a long-term Inkscape project member, and I will retest this when the extension works again. I will leave this open. Thanks for your understanding.

It is untrue that bugs in 0.92.x are no longer being fixed, but what I am saying is this needs checking again with 1.0 WHEN THE EXTENSION WORKS.

Changed in inkscape:
status:	Invalid → New

Revision history for this message

Hachmann (marenhachmann) wrote on 2019-05-02:

So, to be more clear: This issue isn't currently reproducible because another issue is making it untestable. That is not the same as saying that this issue is fixed now.

Revision history for this message

Hachmann (marenhachmann) wrote on 2019-05-02:

Only issues that are FIXED should be marked as such without migration.

Qantas94Heavy (qantas94heavy) on 2019-05-04

Changed in inkscape:
status:	New → Triaged
importance:	Undecided → Low
milestone:	none → 0.92.5

Qantas94Heavy (qantas94heavy) on 2019-05-04

Changed in inkscape:
assignee:	nobody → Qantas94Heavy (qantas94heavy)

Revision history for this message

Hachmann (marenhachmann) wrote on 2019-05-04:

(I assume this might be fixed already, due to the unicode migration that was done with extensions, but needs testing)

Revision history for this message

Qantas94Heavy (qantas94heavy) wrote on 2019-05-04:

I just tested this with 1.0alpha (56f1b1b843, 2019-04-29) and this was not fixed.

There's a few parts to it with Python 2:

1. etree.tostring fails because encoding=utf8 is not specified.
2. If you fix that, inkex.errormsg attempts to convert this utf8-encoded byte string into a Unicode string, but unicode(msg) fails because the encoding is not specified.
3. If you fix that, you quickly realise that was sort of useless, as sys.stderr.write expects a byte string and not a Unicode string. (BUT you only see this error when testing if you set LC_ALL=C!)
4. You then realise that some extension files have unicode_literals set but not others, which means whether you receive a byte string or Unicode string becomes a minefield.

All in all, quite a fun (cough) bug! Now I've got to figure out how to fix this properly :)

Revision history for this message

Hachmann (marenhachmann) wrote on 2019-05-04:

#10

Arghs... I had high hopes for this one to 'just work'...

Thank you for investigating and working on a fix, Quantas94Heavy!

(how the heck did you figure out the LC_ALL thing...? You must be super patient and methodical.)

Revision history for this message

Qantas94Heavy (qantas94heavy) wrote on 2019-05-15:

#11

Hi there, I'm going to move this conversation over to our new bug tracker on GitLab. This will make it easier for us to discuss a solution. Please check there for future updates. Thank you!

Moved to: https://gitlab.com/inkscape/extensions/issues/76
Closed by: https://gitlab.com/Qantas94Heavy