Nokogiri::CSS::SyntaxError: unexpected '|' after '#<Nokogiri::CSS::Node:0x000000012ae10458>'

[shanecav84]

This is an issue that exists with Nokogiri 1.13.1, but not 1.13.0. There was a change in how XPath selectors are handled in CSS rules¹ but I don't know enough to be certain that's the cause.

Failure here:

odf-report/lib/odf-report/table.rb

Line 75 in 16f5b38

doc.at_css("table|table[@table|name='#{@name}']")

Backtrace:

.../nokogiri-1.13.1/lib/nokogiri/css/parser_extras.rb:84:in `on_error'
(eval):3:in `_racc_do_parse_c'
(eval):3:in `do_parse'
.../nokogiri-1.13.1/lib/nokogiri/css/parser_extras.rb:66:in `parse'
.../nokogiri-1.13.1/lib/nokogiri/css/parser_extras.rb:76:in `xpath_for'
.../nokogiri-1.13.1/lib/nokogiri/css.rb:46:in `xpath_for'
.../nokogiri-1.13.1/lib/nokogiri/xml/searchable.rb:246:in `block in xpath_query_from_css_rule'
.../nokogiri-1.13.1/lib/nokogiri/xml/searchable.rb:245:in `map'
.../nokogiri-1.13.1/lib/nokogiri/xml/searchable.rb:245:in `xpath_query_from_css_rule'
.../nokogiri-1.13.1/lib/nokogiri/xml/searchable.rb:237:in `block in css_rules_to_xpath'
.../nokogiri-1.13.1/lib/nokogiri/xml/searchable.rb:237:in `map'
.../nokogiri-1.13.1/lib/nokogiri/xml/searchable.rb:237:in `css_rules_to_xpath'
.../nokogiri-1.13.1/lib/nokogiri/xml/searchable.rb:206:in `css_internal'
.../nokogiri-1.13.1/lib/nokogiri/xml/searchable.rb:129:in `css'
.../nokogiri-1.13.1/lib/nokogiri/xml/searchable.rb:141:in `at_css'
.../odf-report-0.7.2/lib/odf-report/table.rb:75:in `find_table_node'
.../odf-report-0.7.2/lib/odf-report/table.rb:13:in `replace!'
.../odf-report-0.7.2/lib/odf-report/report.rb:61:in `block (3 levels) in generate'
.../odf-report-0.7.2/lib/odf-report/report.rb:61:in `each'
.../odf-report-0.7.2/lib/odf-report/report.rb:61:in `block (2 levels) in generate'
.../odf-report-0.7.2/lib/odf-report/template.rb:83:in `process_entry'
.../odf-report-0.7.2/lib/odf-report/template.rb:35:in `block (2 levels) in update_files'
.../rubyzip-2.3.2/lib/zip/entry.rb:540:in `get_input_stream'
.../odf-report-0.7.2/lib/odf-report/template.rb:30:in `block in update_files'
.../rubyzip-2.3.2/lib/zip/entry_set.rb:38:in `block in each'
.../rubyzip-2.3.2/lib/zip/entry_set.rb:37:in `each'
.../rubyzip-2.3.2/lib/zip/entry_set.rb:37:in `each'
.../rubyzip-2.3.2/lib/zip/central_directory.rb:185:in `each'
.../odf-report-0.7.2/lib/odf-report/template.rb:26:in `update_files'
.../odf-report-0.7.2/lib/odf-report/report.rb:58:in `block in generate'
.../odf-report-0.7.2/lib/odf-report/template.rb:20:in `block in update_content'
.../rubyzip-2.3.2/lib/zip/output_stream.rb:63:in `write_buffer'
.../odf-report-0.7.2/lib/odf-report/template.rb:18:in `update_content'
.../odf-report-0.7.2/lib/odf-report/report.rb:56:in `generate'

Footnotes

1. https://github.com/sparklemotion/nokogiri/issues/2419 ↩

[hoeni]

Same Problem here (odf-report 0.7.2), results in same stack trace. 1.13.0 works fine, 1.13.1 fails with this.

[shanecav84]

@sandrods This is preventing me from upgrading Nokogiri to patch a CVE GHSA-fq42-c5rg-92c2

[sandrods]

Sorry, folks. Gonna take a look at it ASAP

…

On Tue, 22 Feb 2022 at 19:32 Shane Cavanaugh ***@***.***> wrote: @sandrods <https://github.com/sandrods> This is preventing me from upgrading Nokogiri to patch a CVE GHSA-fq42-c5rg-92c2 <https://github.com/sparklemotion/nokogiri/security/advisories/GHSA-fq42-c5rg-92c2> — Reply to this email directly, view it on GitHub <#120 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAABUWJ6RCQXAILNYOKDT5LU4QFIPANCNFSM5OMZDHJQ> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[sandrods]

Well, pipes are part of CSS3 specifications, and are used to query using namespaces

https://www.w3.org/TR/selectors-3/#univnmsp

I'll have to look into it further. It may be a regression bug in Nokogiri, since it worked until 1.13.0.

I may need to translate these queries into xpath. Right now I have no clue.

Will keep you all posted. Any directions would be appreciated

[sandrods]

I think I might have fixed it. My specs are green now.
Could you run a quick test using master to confirm, so I can release a new version?

[shanecav84]

It's working for me. Thanks!

[sandrods]

closed by 1fb39e1

[hoeni]

Works like a charm for me. Thank you so much!