Closed jgm closed 4 years ago
There's one issue remaining: citeproc.js gives the result
(<i>Magazine</i> 2012a; <i>Magazine</i> 2012b; <i>Magazine</i> 2012c; <i>Magazine</i> 2012d; <i>Newspaper</i> 2012a; <i>Newspaper</i> 2012b)
while this library collapses and gives
(<i>Magazine</i> 2012a; 2012b; 2012c; 2012d; <i>Newspaper</i> 2012a; 2012b)
Which is correct according to the spec? (@denismaier do you know?)
I think your library is correct since the title is rendered as through
cs:substitute
.
@fbennett, what do you think? Is there a special reason for this test behaviour?
Will wait for other responses, but it looks to me like collapsing there is the right thing to do, aesthetically. An edge-case question would be whether the match for collapse purposes should be based on exact match of rendered field and value, string value, or string value and decorations (italics etc). Would think the first of those three, but that would be good to confirm.
Currently this fails. We get Newspaper 2012e, 2012f instead of a,b.