Fix#131:
- Raise a syntax error when bogus mode arguments are given.
- Fix a bug where `extend-final` and `augment-final` are not recognized
- Document the `public` mode which was already supported but not
documented.
- Clarify the default mode when `maybe-mode` is not given.
The result of `load-xref` with an on-demand function only made sense
for a single use context, such as a single rendering request. Add an
on-demand callback that can work right for multiple uses.
This works, provided that:
- the package is on the catalog at pkgs.racket-lang.org
- the package is hosted on GitHub or GitLab
Restriction 2 could be lifted for other known sites or packages
hosted as directories.
Restriction 1 would be harder to lift.
This only links to the _top level_ file that defines the overall
document, not the file defining the particular section. Fixing that
would require the Scribble renderer to provide more detailed
information, although it's certainly a possible extension.
Closes#223Closes#208Closesracket/racket#874
Related to #76, #228
- fix#216 by coercing author-element data to a string, add tests
- doc edits:
- import `scribble/core` to fix link to `content?`
- fix typo in `dissertation-location`
- replace unbound reference to `name/c` with a real contract
* scribble/srcdoc: add `class*-doc` and `class-doc` provide forms.
* Add missing require of racket/class which contains the class? predicate.
* Add @history stanza to class*-doc and class-doc scribblings and bump version number in scribble-lib/info.rkt to 1.30.
The indentation for multi-line typesetting is weird, because it still
uses the table-basd layout that lines up all arguments independent of
nesting. But at least the parentheses are not broken.
Relevant to #211
I couldn't locate gif.rkt in the `file' collection. See a complete working example was helpful to get started. I think it will be as useful for
other Racketeers .(especially beginners like myself).
If gif.rkt is indeed gone due to refactorings, I guess the reference to it should removed.
1. Focus on the search box for either "s" or "S"
... accepting only "s" makes sense to me, but the comment said it
accepted "S" and well why not
2. Look for a "keyup" event instead of key press, so that pressing "s"
ONLY focuses on the box and does not focus-and-write-the-letter-"s"
"keyup" events apparently don't have a useful `charCode` field so this
PR looks at the `keyCode` field instead
