that have overlapping ranges in the editor; also got rid of the
id-set stuff
Getting rid of the id-set information that was computed means that now
the mouse-over green bubbles, the "jump to next binding occurrence"/
"jump to bound occurrence" keybindings/menu items, and the renaming
are all being computed from the arrows information as needed, instead
of building up sets as check syntax collects information. This may
change the way Check Syntax behaves in some cases; so far the only
example I've found has been strange and are arguably for the
better. Specifically, this program
(define-syntax-rule (m x) (λ (x) x)) (m z)
no longer draws a green bubble when you mouse over the "z", since
there are no arrows (the only arrow that might have been drawn is
discarded since its start and end points are the same place).
This speeds up the "analyze the expanded code" phase of check syntax,
making it approximately 1.6x faster than before (going from about 31
seconds to about 19 seconds for this phase for the
drracket/private/unit.rkt file (on my machine)). Also, the replay
phase is probably a bit faster now, tho, too: there were 1.07x fewer
elements to process in the trace that comes back from online check
syntax now for that same file (33063 to 30842)
Note that this is only that one phase: this doesn't count the time to
actually expand the program (the dark blue bubble phase) nor the time
to send the results between places, nor the time to replay the
collected information (the light purple bubble phase).
Adds `--from-dir' and `--from-install' flags to select the interpretation
of the argument as a directory or as the name of an installed package.
Relevant to PR 13669
Adds `--as-is' (the default), `--source', and `--binary' flags to
select a pruning mode.
The `raco setup' tool recognizes a `rendered-scribblings' specification
in "info.rkt" to trigger moving rendered documentation into place,
registering its tags in the cross-reference database, and fixing up
references to "local-redirect.js"; the presence of a "synced.rktd"
indicates when those fixups have been performed (since, if the package
is installed in a user-specific scope, the documentation doesn't actually
move anywhere). Finally, "out<n>.sxref" needs to report paths relative to
the documentation's directory, and then the relative-directory references
need to be suitably resolved at derserialization; some support for such
relative paths was in place, but it wasn't quite general enough before.
Even though "info.rkt" files are in collections, `raco setup'
treats them in a directory-specific way --- and that's necessary
for specifying things like `compile-omit-files' in a collection
splice.
When starting a thread, the thread was created and partially
initialized before trying to get a name for the thread from the given
thunk, but getting a name for the thunk could trigger scheduler
descisions, which could get confused by the partially initialized
thread.
There are many cases where you just want to use some other tool like
`#:break' to stop the iteration, so no need to make up a bogus stop
value and no need to spend time checking it.
The `#%module-begin' of `racket/base' and `scheme/base' now introduces
a suitable `configure-runtime' submodule, instead of using the
`module->language-info' path.
A submodule is a lot easier to work with, as illustrated by the
removal of the `racket/private/lang' and `scheme/private/lang'
languages.
Also, add `#%printing-module-begin', which is the old `#%module-begin'
(i.e., the one that doesn't introduce a `configure-runtime' submodule).
A language can now introduce a `configure-runtime' submodule that
is `dynamic-require'd before the enclosing module.
This new submodule protocol provides a more general and
easier-to-understand way of configuring the run-time environment for
a module's language, as compared to the `module->language-info'
path (through a `get-info' function, via a 'configure-runtime value,
and finally loading the specified module).
The `module->language-info' path remains in place, and it is
checked after a `configure-runtime' submodule is run, since
that order is likely to be the most backward compatible.
If a function is bound by a `letrec' (or internal definition)
that cannot be simplified to `let' or lifted to a constant or
top-/module-level function, and if the `letrec' binding is used in
a non-application position, and if the function has in its closure
a typed binding (i.e., boxed, fixnum, flonum, or extflonum),
then the validator was incorrectly rejecting the function's
bytecode --- because the validator didn't distinguish between typed
arguments and typed closure content.
File under "surprised that we didn't hit this one earlier".
Stress mode forces a GC on every N allocation attempts, and it makes
JIT-generated code always take a slow path.
This mode uncovered only a bad test case and some boring start-up
bugs (boring because start-up is deterministic enough that they
never happen), so far.
