saved SVG is not well-formed

[anatoly techtonik]

Attached SVG file was saved by Inkscape 0.46, but it can't be read neither by Inkscape nor FF or other programs. Because it contains invalid symbol at Line Number 229, Column 69. I guess that this symbol got there as a result of copy-paste operation, but was not escaped during save operation.

[sas (sas-sas)]

Confirmed in SVN revision 19107.

The invalid character is U+001A. The bug can be reproduced by copying and pasting this character into a text object in Inkscape, and then saving.

The character needs to be removed rather than escaped (as it would still be invalid even if escaped).

[Luke V (lvercimak)]

Confirmed in SVN revision 19574

The XML standard defines the following as valid characters
char ::= #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]
(Extensible Markup Language (XML) 1.0 (Fourth Edition) paragraph 2.2)
http://www.w3.org/TR/REC-xml/#charsets

The following patch checks for control characters (0x01-0x19 excluding 0x09, 0x0a, and 0x0d) on the pasted text and removes them if necessary. It can be expanded to check for the rest of the gaps in character ranges above if they become a problem (they were left out in favor of speed). By preventing the control characters from comming in off of the paste buffer, files like the above attachment won't get generated.

[Luke V (lvercimak)]

See attached patch

[Buliabyak-users (buliabyak-users)]

thanks for the patch, committed