When an evaluation result from a sandboxed computation is convertible
(so that it should be preserved for rendering) and serializable, then
serialize the value from transfer from the sandboxed environment to
the rendering environment.
This change complements the one that makes `pict`s serializable.
Serializing a pict result avoids retaining the sandboxed environment,
including its instance of the `pict` and `racket/draw` modules,
when a sandbox-created pict is part of the generated document.
For example, the memory use of the result of "math.scrbl" is about
20% smaller with this change.
original commit: c9a11c4ee4dff8e099e30d8c5040b087dae4a42b
The case where (car a) is a string wasn't handled correctly. The error was provoked by sicp-manual.scrbl.
original commit: 9a4f6b1a194cda0b841bac36df18fa4b822bf469
The `tt` and `racketfont` functions treat their arguments strangely,
applying styles only to immediate strings and symbols.
Add `racketplainfont`, which is a non-strange version of `racketfont`.
original commit: 737330deb63dcfa01ebf304c3e5c06276a283c81
When text is literally a quote (ASCII 39) or backquote (ASCII 96),
then render it in a way that prevents LaTeX from converting to
curly variants.
original commit: 69d64e2893cca5e143609494547b64b7f90b7228
The `at-exp` reader now delays picking up the current readtable until
`read`/`read-syntax` is called. Also, it uses the new 'dynamic configuration
of readers for the command and datum parts of an @-form, which delays a
decision of readtable on each part until reading the part.
Thanks to Alexander Knauth for sorting out pieces of the puzzle.
original commit: a087aea3e58560427980b28d9ccb64815edbff01
Now that some entries have changes over multiple versions, a forward
sort order seems better than reverse to me.
original commit: a67ecb96ab18a54dff75f06a737c492ba92943a4
Use 'pdf-bytes+bounds8, 'png-bytes+bounds8, etc. to convert images
for Latex/PDF and HTML output. The new conversions particularly
help with picts that draw text, since the baseline of the text
can be handled properly in HTML output.
The new conversion also helps hide the difference between a bounding
box and the "inked" area of a pict, since picts normally ink only
slightly outside the bounding box, and the default for pict conversion
pads the image with a few pixels/units in each direction.
Also, add a `++convert <fmt>` command-line option to select a
preferred target for image conversions. (This preference can be
overridden by the existing `render-convertible-as` style property
for HTML rendering.)
original commit: 05760a12f690b7efcd58e65ee2b723a05133e229
This adds a new back-end pipeline for generating pdf to
scribble, with the hope that included picts (e.g., those
generated by Redex) will look better when viewed with
on-screen pdf viewers
Current problem: the descent adjustment is messed up,
as can be seen by this example program's output:
#lang scribble/base
@(require redex/pict
pict
redex/reduction-semantics)
@title{}
@(define-language L
(e ::= (e e) x (λ (x) e)))
@(render-language L)
ff@(text "f" '(italic . roman) 20)ff
x@(text "x")y@(text "y")
original commit: 5280395f8802d21764ef074540421441cb8ac999
Normalize strings to composed form before splitting a
string into characters to handle individually.
original commit: d970d5aaead2bfbae0651b583f79e6ac719c7d39
This changes the behavior of make-cite and authors, but it would
probably be better to change the behavior of generate-bibliography
original commit: 74831b41cc2589af94a4e7cc6ffefa1d8809226c
Modules are not supposed to move to different packages, anymore, so
I'm breaking a rule. The problem is that having `rackunit/docs-complete`
in "scribble-lib" made "scribble-lib" depend on "racket-index", and
that's an undesirable dependency. (Installing Scribble shouldn't
force you to have Racket documentation.) I should have noticed
the problem when I moved `setup/xref` to "racket-index", but I missed
it.
As far as I can tell, there's only one package on pkgs.racket-lang.org
that isn't in the main repository and uses `rackunit/docs-index`. The
package doesn't correctly declare its dependencies, anyway.
original commit: d33c2252e10dbeb22e49c85f03707345e2756ce5
... instead of putting the figure's caption at the top of the
browser window.
I can't decide whether I like the technique here, which is to use
JavaScript to move a figure caption's <a name=...> when the page is
loaded. On the one hand, fixing up HTML via load-time JavaScript code
seems ugly compared to generating the right HTML in the first place.
On the other hand, it simplifies the generatation of a
backend-independent Scribble document (i.e., don't try to generate
different target locations for HTML and Latex), and HTML-specific
behavior seems properly handled at the HTML/JavaScript level.
original commit: 2b33b8d9668988b67f6b68b9eb012eb6d8f417d5
Use Javascript's decodeURIComponent() instead of unescape().
This repair makes searching for Unicode characters work from DrRacket
(e.g., via F1).
original commit: d1525edf62a1db2ade3c64dd17eb81bff87a4112
This change avoids the problem of a bad "index.html" on a doucment
rendering error, where the presense of an "index.html" would count
as successes on a retry.
original commit: e4e8a69e06863b423f4b7b716b369b002c2deea1
This configuration option is a better way to redirect documentation
access (at a place like NEU) than patching "search.rkt".
original commit: a8105dc0e34e5571ec9ecdb47635e409e29882c5
At mbutterick's recommendation:
it's causing a strange display problem in TOC listings when the
cursor is hovering over the link (namely, the underlined space
dangling from the right edge)
original commit: cc26a86cb1975bdbb1ed3f816b6a50decfdec510
