| matmarex | |
| Feb 4 2015, 11:00 PM |
| F37325: pasted_file | |
| Feb 4 2015, 11:00 PM |
| F37327: pasted_file | |
| Feb 4 2015, 11:00 PM |
This terribly malformed table is parsed differently than in PHP parser:
{| class=wikitable
random text 1
|-
random text 2
| table cell 1
|-
random text 3
| table cell 2
|}| PHP parse | Parsoid parse |
|---|---|
| https://test.wikipedia.org/wiki/Misnesting_test | http://parsoid-lb.eqiad.wikimedia.org/testwiki/Misnesting_test |
If this is intentional/acceptable difference, please state so and decline. I just found the difference interesting enough to report.
This is a known issue ... I could have sworn I had a ticket for it.
But, the general problem is that Parsoid introduces an implicit <td> when bare content is found after a "|-" construct.
If you look in parserTests.txt, I've added comments about implicit <td>s causing diffs between PHP parser and Parsoid. @GWicke and I probably discussed why we ever added that support in the first place, but I cannot remember now. Maybe he remembers and clarify. I imagine it is a side effect of having to support templates after "|-" rows. But, we also have bugs like T71123
But, we should consider removing this support from Parsoid .. or leave it behind as a feature .. not sure what is better.
[subbu@earth:~/work/wmf/parsoid] php ../mediawiki/maintenance/parse.php < /tmp/wt parse.php: warning: reading wikitext from STDIN. Press CTRL+D to parse. random text 1 random text 2 random text 3 <table class="wikitable"><tbody><tr> <td>table cell 1 </td></tr><tr> <td>table cell 2 </td></tr></tbody></table> [subbu@earth:~/work/wmf/parsoid] php bin/parse.php --normalize < /tmp/wt [warn/dsr/inconsistent] DSR inconsistency: cs/s mismatch for node: tr s:67 ; cs:82 [warn/dsr/inconsistent] DSR inconsistency: cs/s mismatch for node: tr s:35 ; cs:50 [warn/dsr/inconsistent] DSR inconsistency: cs/s mismatch for node: tbody s:18 ; cs:33 <p>random text 1 random text 2 random text 3</p> <table> <tbody> <tr> <td>table cell 1</td> </tr> <tr> <td>table cell 2</td> </tr> </tbody> </table>