I've forked and been making some changes to the xslt2 stylesheets for a
personal project, and I'm trying to get some of those changes into a
shape where they might be pull-requestable. A few of them are very
small, but I have some questions about the intended functionality.
1. The `olink.doctitle` parameter is a numeric type by default in the
stylesheet params definition, but the stylesheets compare it as a string
in one place
<https://github.com/docbook/xslt20-stylesheets/blob/master/xslt/base/common/olink.xsl#L1023>,
and a numeric in another
<https://github.com/docbook/xslt20-stylesheets/blob/c1b89c3aff3c3ce2e5b8dafbcd376eb148afa944/xslt/base/common/gentext.xsl#L868>.
Which is supposed to be correct? Numeric makes more sense to me personally.
2. The Docbook schema says that only one `glossterm` may exist as a
direct child of a `glossentry`. But the stylesheets have contingency for
multiple `glossterm`
<https://github.com/docbook/xslt20-stylesheets/blob/master/xslt/base/html/glossary.xsl#L108>
(which is actually a very useful feature even though it's "wrong").
Unfortunately, other code
<https://github.com/docbook/xslt20-stylesheets/blob/master/xslt/base/html/inlines.xsl#L692>
will crash if multiple `glossterm` exist (node-set will get passed into
normalize-space which will fail). Is it preferable to remove the
contingency templates, or to "correct" the crashing code to handle the
"bad" case?
Thank you,
Winslow Dalpe