General fixes shaken out running clojure.core.cache (66 -> 198 of its assertions):
- Map destructuring applied an :or default only for :keys/:strs/:syms, not a
direct {x :x} binding — so {x :x :or {x 9}} (and the & {…} kwargs form) ignored
the default. Apply it for the direct binding too.
- fn didn't implement :pre/:post: a leading conditions map was evaluated as a body
literal (so % was unbound and (.q %) blew up). Recognize it and assert pre
before the body, bind % to the result, assert post, return %.
- (.q inst) on a deftype field with no matching method reads the field, like the
JVM (was "No method q").
- A deftype implementing the clojure.lang collection interfaces now dispatches
dissoc (without), contains? (containsKey), peek/pop (IPersistentStack), and
keys/vals (via its Seqable seq) through its methods — they were field-only, so
core.cache's caches and data.priority-map didn't behave as maps.
JVM-certified corpus rows for each. make test + shakesmoke green.
input-bytes handled in-stream/bytevector/byte-array sources but not a host
byte-input-stream table (:jolt/input-stream — http-client's ByteArrayInputStream),
so io/copy fell through to rendering it as text (#[chez-htable …]) instead of its
bytes. Drain it like slurp does. Fixes http-client's response-body handling
(jolt-bjbi): its suite goes 100/16-fail -> 116/0, and the ring adapter's.
io/copy handled file/stream/writer targets but not a host byte-output-stream
table (jolt-lang/http-client's ByteArrayOutputStream shim, :jolt/output-stream),
erroring 'don't know how to write to'. Dispatch through the shim's .write method,
byte-exact for a byte source — the JVM's io/copy writes to any OutputStream.
Shaken out getting ring-defaults (and its ring-core/anti-forgery/session stack)
to load and serve static resources on jolt. All general fixes, all runtime:
- Class/forName throws a catchable ClassNotFoundException for a class jolt can't
back (it returned a broken truthy value for any name, and crashed on use). Lets
the common (try (Class/forName "optional.Dep") (catch ...)) probe libraries use
to detect an absent dependency work — e.g. ring's joda-time check.
- deps: reconcile native libs (and source roots) in one step, deduped by library
identity, instead of the ad-hoc distinct at each call site. An app pulling two
libs that declare the same shared object (libcrypto via both jolt-crypto and
http-client) now includes and loads it once.
- io: a File answers getProtocol ("file") / getFile so resource-serving
middleware that expects io/resource to hand back a file: URL works; the
classloader gains getResources (every source root holding the resource).
- clojure.string/replace accepts a char match/replacement, like the JVM.
JVM-certified corpus rows for the Class/forName and string/replace behavior.
Finishes core.match — its full test suite (115/115) now passes, including the
two patterns the earlier work left out:
- Regex-literal patterns. A #"…" now reads as a regex VALUE (Clojure parity: the
reader constructs the Pattern, so a macro receives a regex, not jolt's tagged
form), and the analyzer compiles a regex value to the same :regex IR leaf via
its source. emit-quoted handles a quoted regex; a regex value carries the
java.util.regex.Pattern host tag so extend-protocol/instance? dispatch on it.
- Primitive-array patterns. A ^Type hint's :tag is now the SYMBOL (e.g. `ints`),
matching the JVM, so core.match's array-tag lookup engages the array
specialization (alength/aget). jolt's :tag consumers already tolerate a symbol
(hc-cell-num-ret normalizes; tag->nkind/def-meta handle both).
Also: a library-conformance directive in CLAUDE.md, and the supported-libraries
list (docs + site) simplified to one-line entries — a listed library is assumed
to work fully, so no tallies or feature enumerations. core.match + transit-jolt
added to the list.
Seed change (reader/backend/30-macros) -> re-minted; the rest runtime. JVM-
certified corpus rows; the stale `symbol hint -> :tag` divergence is dropped from
the allowlist (jolt now matches the JVM). make test + shakesmoke green.
Running clojure.core.match (a macro-heavy library that builds its compiler out
of deftypes implementing clojure.lang interfaces) shook out a cluster of general
gaps. Its own suite goes from not-loading to 111/115 assertions.
- deftype/defrecord implementing a clojure.lang collection interface now drives
the core fns: Indexed -> nth, Counted -> count, Associative -> assoc, ILookup
-> get/valAt (non-field keys only, so a method's own field bindings don't
recurse), ISeq -> seq/first/rest, IPersistentCollection -> conj, IFn -> the
value is callable. A jrec is still a map of fields by default; the interface
method wins when declared.
- Multi-arity inline methods are grouped into one fn (a type with (nth [_ i]) and
(nth [_ i x]) kept only the last before). Built as data, not a nested
syntax-quote, so a `(= ~ocr ~l) method body keeps its unquotes.
- instance?/satisfies? recognize a protocol a type implements, including a MARKER
protocol with no methods (core.match's IPseudoPattern) — deftype/defrecord now
record protocol satisfaction even with zero methods. Added ILookup/Indexed/
Counted to the instance? taxonomy for the built-in collections.
- Syntax-quote: a fully-qualified class name (clojure.lang.ILookup) stays bare
instead of being namespace-qualified; (unquote x) is detected in a lazy seq
(a macro that builds its template with map, e.g. deftype's rewrite-set).
- clojure.set union/intersection/difference are variadic (& sets) + union 0-arity.
- java map view methods: keySet/values/entrySet/size/isEmpty.
- deprecated java.util.Date getters (getYear/getMonth/...) + the multi-arg
(Date. year-1900 month0 date hrs min) constructor.
Seed change (deftype/defrecord macros + clojure.set) -> re-minted; the rest are
runtime. 11 JVM-certified corpus rows; make test + shakesmoke green.
clojure.edn's __read-tagged seam called (empty-pmap) — applying the empty-pmap
VALUE as a procedure — so an unknown tag (e.g. #object[...] from a JVM-printed
object, or any unregistered #foo) crashed the Chez VM with "attempt to apply
non-procedure" and surfaced a malformed condition (class :object, nil message)
instead of a catchable error.
Throw a clean ex-info naming the tag, matching the JVM's "No reader function
for tag <tag>". A reader port over edn (transit-jolt's read-conformance skip
path, aero, etc.) now catches a real exception instead of aborting.
clojure.core/read-string stays lenient (returns the tagged form) so clojure.edn
can apply :readers / :default before falling through to this throw.
Expand java.io so libraries that touch the filesystem work unchanged.
File: the full method surface — length, lastModified, can{Read,Write,Execute},
isHidden, list, mkdir(s), delete, createNewFile, renameTo, getParentFile,
get{Absolute,Canonical}File, compareTo/equals/hashCode — plus the statics
separator / pathSeparator / createTempFile / listRoots. A File now keeps the
path as given (new File("rel").getPath() is "rel", .isAbsolute false); a
relative path resolves against JOLT_PWD only when the filesystem is touched,
matching the JVM. slurp/spit and the dir helpers go through the same
resolution, fixing a spit-vs-slurp inconsistency.
Streams (host/chez/io-streams.ss) — each a jhost wrapping a Chez port, so
buffering, EOF and binary<->char transcoding come from Chez:
- FileInputStream / FileOutputStream / ByteArrayInputStream /
ByteArrayOutputStream / BufferedInputStream / BufferedOutputStream
- FileReader / FileWriter / InputStreamReader / OutputStreamWriter /
BufferedReader / BufferedWriter
Buffered* return the wrapped stream (Chez ports are already buffered).
clojure.java.io: input-stream/output-stream now yield real byte streams (were
aliased to the char reader/writer); added copy (byte-exact for byte sources),
make-parents, delete-file. with-open also closes file-writer/port-writer/
print-writer (a pre-existing gap).
All runtime shims, no re-mint. 15 JVM-certified corpus rows; make test +
shakesmoke green.
Shaking out tick's api and alpha.interval suites (api 353->359, interval
0->103 passing) cleared a set of general gaps:
- Named-zone DST. Zones resolved to a fixed representative offset, so
America/New_York in August read -05:00 not -04:00. Add US/EU DST rules
(compact transition-date math) and make instant<->zoned, the zone rules'
getOffset, and the zoned equality arm DST-aware.
- Nanosecond zoned/offset times. Instant is nanos but atZone/atOffset/
toInstant/withZoneSameInstant and Instant/parse went through epoch-ms,
truncating sub-ms. Route them through nanos.
- Locale month/day names. A formatter dropped its Locale; carry it and add
French names so MMM under Locale/FRENCH renders "mai".
- Callable records. A defrecord implementing clojure.lang.IFn (tick's
GeneralRelation) is now invokable: jolt-invoke dispatches to its inline
invoke method. Also give collections the Iterable host tag so a protocol
extended to Iterable matches vectors/seqs.
- Imported class short names. (:import [java.time ZonedDateTime]) then
(. ZonedDateTime parse s) resolved to nil; an otherwise-unresolved bare
Capitalized name that's a registered host class now resolves as a class.
- Data readers. A project's data_readers.{clj,cljc} is loaded into
*data-readers* (reader namespaces required eagerly); registered #tag
literals in source rewrite to (reader-fn 'form). clojure.core/read-string
now applies #inst/#uuid/#"regex" and *data-readers* like Clojure.
- Duration/between accepts zoned/offset date-times.
All runtime shims, no re-mint. docs/libraries.md: tick full pass + aero.
Four general gaps, shaken out by loading clojure.spec.alpha:
- Special forms were shadowable by a same-named macro. analyze-list
macroexpanded before checking special forms, so a ns that redefs def/and/or
(spec excludes them via :refer-clojure :exclude) made a bare def resolve to
the macro instead of the special form, breaking every defn after. Now a head
in the special-form set is never macroexpanded, matching the reference
macroexpand1 isSpecial check.
- reify dropped all but the last arity of a multi-arity protocol method (spec
reifies (specize* [s]) and (specize* [s _])). The macro keyed methods by name
and overwrote; now it groups arities into one multi-arity fn.
- reify instances did not implement IObj: with-meta threw and (instance?
clojure.lang.IObj r) was false. Every Clojure reify carries metadata. with-meta
now copies the reify to a fresh identity (shared method table) and keys its
meta; instance? IObj/IMeta is true for any reify. This was the registry bug —
spec's with-name returned nil for specs, so get-spec missed.
- (set! (. Class field) val) was rejected. spec toggles
clojure.lang.RT/checkSpecAsserts this way; the analyzer now lowers it to a
jolt.host/set-static-field! call over a mutable-statics table, and a plain
Class/field read consults that table.
Also: .name/.getName on a Namespace and .ns/.sym on a Var (spec's ns-qualify /
->sym). analyzer + reify are seed sources (re-minted). spec.alpha now does
valid?/conform/cat/keys/explain-str/check-asserts. tick.alpha.interval-test still
needs time-literals data readers (separate).
The Instant jhost stored epoch-ms, so plusNanos/getNano rounded to the
millisecond and two instants a nanosecond apart compared =. Store epoch-nanos
instead: mk-instant still takes ms (scales to nanos) for the ms-based zone/Date
callers, mk-instant-nanos/inst-nanos own the nano arithmetic, and inst-ms floors
nanos back to ms. plus/minus/getNano/truncatedTo/compare/equals and the ISO
renderer all run on nanos; toString shows the fraction in 3/6/9-digit groups like
ISO_INSTANT. Fixes tick's interval coincidence test (shift end by one nano).
ZoneOffset/ZoneId (SHORT_IDS, fixed-offset + UTC + system; named zones via a
fixed-offset table), ZonedDateTime/OffsetDateTime/OffsetTime, Clock (fixed/
system, with now [clock] arity), and DateTimeFormatter integration (ofPattern
+ ISO_* constants, .format/.parse over the rich java.time values via the
inst-time.ss pattern engine). systemDefault resolves to UTC to keep the
#inst atZone/toInstant round-trip machine-tz-independent.
tick.core + tick.protocols + tick.locale-en-us load; tick's api_test runs
31 tests / 352 pass / 7 fail / 0 error. The 7 are host gaps: named-zone DST
(no tzdb), French locale month names (no locale DB), nanosecond Instant.
General fixes surfaced by tick: :ns/keys map destructuring ({:tick/keys [..]})
in 00-syntax.clj (re-minted), and extend-protocol to java.time classes
(records.ss host-type-set). 12 corpus rows certified vs JVM. make test +
shakesmoke green, selfhost holds, 0 new divergences, data.json stays 138/139.
Duration (ISO PT.. toString, between, full arithmetic), Period (between with
borrow, P.. toString, normalized), full Month/DayOfWeek enums (named constants,
print as their name — fixes the Phase-1 raw-jhost print), Year, YearMonth
(2020-02 toString, leap, atDay/atEndOfMonth), ChronoUnit (between/getDuration)
and ChronoField. The temporal machinery on the Phase-1 types now works with
ChronoUnit/ChronoField: (.plus t n DAYS), (.until t1 t2 unit), (.get/.getLong
t field), (.with t field v), (.isSupported ..), (.truncatedTo ..).
Analyzer: (. Class method args) with a class target lowers to a static call
(Class/method args) instead of mis-dispatching as an instance call on the arg
— matches JVM; needed by cljc.java-time.year. Seed re-minted; selfhost holds.
The Phase-2 cljc.java-time namespaces load; tick.core advances to a Phase-3
zone gap. 10 corpus rows certified vs JVM. make test + shakesmoke green, 0 new
divergences, data.json stays 138/139.
Core java.time value types as jolt host objects backed by the inst-time.ss
calendar engine (days-from-civil/civil-from-days/inst-fields/format-ms), in a
new host/chez/java-time.ss. tz-free reps: LocalDate=epoch-day,
LocalTime=nano-of-day, LocalDateTime=(epoch-day,nano-of-day); Instant reuses
the ms-based instant jhost (ms granularity; nano is a documented gap). Each
type registers statics (of/parse/now/MIN/MAX/...), instance methods
(plus/minus/with/get/isBefore/until/toString/...), =/hash, compare,
ISO-8601 print, instance?, and value-host-tags for protocol dispatch.
Reconciles the old ms-based local-date/local-dt stubs into the rich types
(LocalDateTime now prints ISO; .toLocalDate/.atZone paths preserved). The
four cljc.java-time namespaces (local-date/local-time/local-date-time/instant)
load. Deep temporal-field/unit machinery (range/get-long/with-field/until/
adjust-into) stubbed for Phase 2.
12 corpus rows certified vs JVM 1.12.5. make test + shakesmoke green, 0 new
divergences, data.json stays 138/139, selfhost holds.
class / .getClass now return a Class value that renders like the JVM —
(str c)/.toString -> "class <name>", pr -> "<name>", .getName/.getSimpleName/
.getCanonicalName work — but stays = and hash equal to its name string, so
(= (class x) String), class-keyed maps, multimethod dispatch on class, and
instance? keep working against the bare class-name tokens. instance? unwraps
a Class passed as the type arg.
clojure.test/class-match? no longer assumes (class e) is a string (a jolt-ism;
on the JVM a Class isn't a string either) — reads the name via .getName.
Matches JVM Class.toString, which libraries surface in error messages
(clojure.data.json DJSON-54 expects "...of class java.net.URI"). data.json
suite 139/139 bar the one UTF-16 surrogate test (Unicode-scalar char model).
5 corpus rows certified vs JVM; make test + shakesmoke green, 0 new divergences.
java.util.Calendar (a UTC calendar over epoch-ms): getInstance, the int
field constants, set/get/setTime/getTime/getTimeInMillis. java.time pieces
for a zoned round-trip: DateTimeFormatter/ISO_ZONED_DATE_TIME, formatter
.withZone/.parse, LocalDateTime/parse (with formatter), Instant/from,
.toLocalDate/.atStartOfDay. java.sql.Date is now a distinct class (its own
host value + dispatch tags) so a protocol extended to both java.util.Date
and java.sql.Date routes a sql.Date to its own impl. All UTC-consistent.
data.json print-sql-date + print-time-supports-format pass (suite 137/139).
Replace the 33-line pprint shim with a column-aware pretty printer and a
Common Lisp cl-format engine, ported from the ClojureScript implementation
(no STM — atom-backed fields) and adapted to JVM-Clojure interop. Provides
pprint/write/write-out/with-pprint-dispatch/formatter-out/cl-format and
simple/code dispatch.
core print routes column-aware into an active pretty-writer via a __write
hook (suppressed inside with-out-str captures); PrintWriter host class
forwards into the wrapped writer. Re-mint: pprint is baked into the seed.
Unblocks clojure.data.json/pprint (its pretty-printing test passes).
Directives ~R/~P/~C/~F/~E/~G/~$/~(~) not ported (unused by the targets).
make test + shakesmoke green, 0 new divergences, selfhost holds.
Shaking out clojure.data.json's own test suite (now 134/137):
- Reader.read(char[],off,len) bulk read + PushbackReader.unread(char[],off,len)
on the string/pushback reader jhosts; instance? java.io.PushbackReader/
Reader/StringReader (data.json re-wraps a reader unless it's already a
PushbackReader, so this is load-bearing for repeated reads).
- number protocol dispatch by actual type: a flonum is Double (not Long),
exact ratio is Ratio, exact integer is Long — value-host-tags split.
- Integer/toHexString|toOctalString|toBinaryString|toString; .isNaN/.isInfinite
as instance methods on numbers.
- EOFException ctor/class; .isArray on a class-name string.
- dispatch tags for the uuid/bigdec/inst host values so a protocol extended to
java.util.UUID / java.math.BigDecimal / java.util.Date / java.time.Instant
reaches its impl; canonical-host-tag strips java.math./java.time.
- instant/zoned/local time values compare = by epoch-ms (two parsed Instants).
- java.time.Instant/parse, java.sql.Date ctor + valueOf, TimeZone/getDefault,
DateTimeFormatter/ISO_INSTANT.
All runtime .ss (no re-mint). 9 corpus rows certified vs JVM; make test +
shakesmoke green, 0 new divergences.
- String.getChars copies into a char-array; String. builds from a char[]
(whole or offset/count slice); subSequence returns the substring.
- str of a StringBuilder returns its content (was the opaque host object;
.toString already worked).
- Appendable.append gains the 3-arg subsequence form append(csq,start,end)
on StringBuilder/StringWriter/file/port writers.
- reader combines a \uXXXX surrogate pair into the one Unicode scalar
(😃) instead of crashing on the lone high surrogate; a stray surrogate
maps to U+FFFD. (A high-plane char re-escaped as \uXXXXX stays the
irreducible UTF-16/scalar divergence.)
- TimeZone/getDefault.
These let clojure.data.json read and write JSON; its own suite goes from
not loading to 97/133 assertions passing (remainder: per-type writer
dispatch for uuid/bigdec/date, EOFException, niche date interop).
(instance? clojure.lang.Named :a), java.lang.CharSequence/Number,
java.util.Map/Set/List/Collection, clojure.lang.Associative now report
true for jolt's value model, matching the JVM (a Map is not a Collection;
a List excludes sets/maps). Libraries branch on these — data.json's
default-write-key-fn keys on clojure.lang.Named, so map keys serialized
with the leading colon before this.
A slash-free dotted symbol with a Capitalized final segment (java.util.Map,
clojure.lang.Named, java.time.Instant) now self-evaluates to its name string
instead of resolving to nil — jolt models a class as its name, so a library
can extend a protocol to, or instance?-check, a host class jolt has no shim
for. hc-resolve-global classifies these as :class; the analyzer emits a const.
extends? now matches when either the query or the registered tag is a dotted
suffix of the other, so (extends? P java.util.Collection) finds the impl
extend registered under the canonical short tag.
Add DateTimeFormatter/ISO_INSTANT (UTC, trailing Z).
These unblock loading clojure.data.json, which dispatches JSONWriter on
java.util.Map/Collection/CharSequence/Instant and defaults a formatter to
ISO_INSTANT.
read-string/read now return real sets for #{...} literals (top-level and
nested) instead of the reader's {:jolt/type :jolt/set} form — the data
seams convert set forms to sets (recursing, preserving metadata and source
map-key order); clojure.edn already did this. The compiler keeps reading
via the raw reader, so set literals in code stay forms the analyzer lowers.
format %x now emits lowercase hex (Chez number->string is uppercase); %X
unchanged.
extend and extends? handle a nil target type (host tag "nil"), matching
extend-type — protocols can be extended to nil via the function form, not
just the macro.
Found porting transit/data.json and shaking out aero.
conj/assoc/dissoc/disj/pop/into and empty now thread the receiver's
metadata onto the result, matching Clojure (each op constructs a new
collection with meta() carried forward; coll.empty() is
EMPTY.withMeta(meta())). The metadata side-table is now weak so meta on
intermediate collections is reclaimed with them, and empty-list-t carries
an (unused) field so a metadata-bearing () is a distinct identity from the
shared singleton instead of leaking meta onto every ().
Unblocks metadata-driven walks (aero/integrant): (into (empty form) ...)
now preserves a vector/map/set's metadata, so a postwalk whose outer fn
reads (meta x) sees it.
The reader lowered ^meta on a vector/map/set literal to a runtime
(with-meta form meta) list, so read-string/edn of data with metadata
returned the form and lost the metadata. Attach it to the value instead,
as Clojure does; the analyzer re-emits (with-meta coll meta) for a
meta-carrying collection literal in code, so a literal still carries its
metadata at runtime and ^Type/^long arglist hints (consumed by
analyze-arity directly) are unaffected.
Also: pr honors *print-meta*, and clojure.walk/clojure.edn re-attach
metadata to the collections they rebuild (matches Clojure; a
metadata-driven config lib like aero relies on it).
clojure.test reported a host-condition crash with a blank message (ex-message is
nil for non-ex-info conditions); fall back to jolt.host/condition-message.
condition-message itself left a Chez format-template message (open-input-file's
'failed for ~a: ~(~a~)') unformatted; apply its irritants.
jfile lacked .isAbsolute (path starts with /). clojure.core/default-data-readers
and *data-readers* were nil, so a library merging them (aero's reader opts)
couldn't resolve #inst. Define them keyed by symbol, like Clojure.
(instance? clojure.lang.IObj/IMapEntry/IRecord/IPersistentMap/... x) all returned
false, so libraries branching on these interfaces (e.g. a metadata-preserving
walk guarded by IObj) silently misbehaved. Register an instance-check arm mapping
the interface tokens to jolt predicates, matched by last dotted segment.
(tagged-literal t f) compared by identity, so two equal tagged literals were not
= and didn't dedup as map keys / set members; JVM's TaggedLiteral is value-equal.
Add a register-eq-arm! for jtagged (tag + recursive form); hash is already
structural. Fixes an infinite loop in any fixpoint that dedups structures
containing tagged literals (e.g. aero's #ref expansion).
io/file already resolved a relative path against JOLT_PWD (the user's cwd before
the launcher cd's to the repo root), but slurp and io/reader on a relative string
path opened it against the process cwd — so a program run from its own directory
couldn't read its own relative files (e.g. aero's (read-config "config.edn")).
Route the string branches through project-relative, matching io/file. Internal
callers pass already-resolved paths through read-file-string and are unaffected.
clojure.core/*print-meta* was missing, so (binding [*print-meta* true] …) failed
to compile ("var of non-var"). Add it (default false, like the JVM); corpus
covers binding it.
A pushback-reader (drainable, so clojure.edn/read over it works); getLineNumber
is a stub since jolt doesn't track line numbers. Config readers (aero) wrap their
input in one.
A (defmethod m …) where m is :refer'd from another ns passed the bare symbol to
defmethod-setup, which resolved it in the current ns and created a shadow multifn
— the method never reached the real one. Resolve an unqualified name through the
refer table (then current ns) so it lands on the referred multifn.
The AOT build strips the ns form, so the refer table is empty in a binary; emit
chez-register-refer!/-refer-all! per app ns alongside the existing alias
registrations. build-app's fixture gains a defmethod on a referred multifn.
with-meta/meta only recognized the eager collections, so (with-meta (map-indexed
…) m) — a lazy seq — threw "value does not support metadata". A lazy seq is IObj
on the JVM; add it to the metadatable set (keyed in the side-table by identity,
like cseq/procedure).
The reader dropped the namespace on ::kw (read ::foo as :foo), so auto-resolved
keywords never matched their qualified form — code that round-trips them (spec
keys, aero's :aero.core/* expansion keys) silently broke. Resolve ::name against
the current ns and ::alias/name through the alias table, as Clojure does. The
runtime loader reads form-by-form with the ns set after the ns form; the
cross-compile reads all forms up front, so ei-emit-ns*/ei-emit-ns-records set the
ns before reading.
clojure.edn/read over a reader discarded its opts map — :readers/:default/:eof
were ignored, so a custom :default never saw the tag. Route the reader arity
through read-string so opts apply, and pass the tag to :default as a symbol (not
the internal :#name keyword), matching Clojure.
Seed re-minted (the ::halt transducer key in clojure.core now reads as
:clojure.core/halt). Corpus gains ::-keyword rows; the unit case that asserted the
old ns-dropping behavior now asserts the qualified result.
types.clj held the inferencer, the success-type checker, and the driver in one
716-line namespace. Move the self-contained checker into jolt.passes.types.check:
the error-domain predicates (not-number?/not-seqable?/not-callable?), the op
tables, type-name, check-invoke, and the user-fn registry. These are pure over
inferred types and the run's env cells, with no inference, so a check-rule edit
can no longer perturb the inferencer.
The infer-coupled probes stay in types.clj — isolated-diag-count and
check-user-call re-run inference, so moving them would make check depend on the
inferencer and reintroduce the cycle. Verbatim move; new ns wired into
ei-compiler-ns-files; seed re-minted to the byte-fixpoint.
The stderr seam wrote to current-error-port without flushing, so a process that
never returns from -main (a server loop) left its log lines in a buffer that only
drained at exit — they never appeared. Flush each write, like the JVM's
auto-flushing System.err.
The source loader sets the current ns and registers :as aliases per file. The
build flattened every app namespace into one image with no such markers, so all
app forms ran under the last-set ns ("user"). Two breakages followed, both only
in a built binary:
- defmulti/defmethod resolve their target var through chez-current-ns, so they
registered the multifn under "user" while compiled var-derefs used the baked
ns — the multifn the app saw was uninitialized ("not a fn nil" on dispatch).
- a quoted alias-qualified symbol (a (defmethod ig/foo …) on an aliased multifn)
resolves its ns through chez-resolve-alias, but the stripped (ns …) form left
the alias table empty, so it landed in ns "ig".
bld-ns-prelude now emits (set-chez-ns! ns) plus chez-register-alias! for each
ns's :as aliases before that ns's forms, in both the normal and tree-shake emit
paths. The build-app fixture gains a :default multimethod and an aliased cross-ns
defmethod so buildsmoke covers both across all build modes.
walk treated a record as a plain map (record? implies map?), rebuilding it via
(into (empty form) ...) which yields a bare map and drops the type. Add a record
branch before the map branch that conj-es the walked entries back onto the
original, matching JVM clojure.walk's IRecord case. Type-dispatched walks need it
— integrant resolves #ig/ref by detecting its Ref record while postwalking the
config, so without this every ref silently fails to resolve.
clojure.walk is baked into the prelude, so the seed is re-minted. Corpus gains
five JVM-certified rows for record type/instance? survival through pre/postwalk.
infer's :invoke case was ~120 lines of cond arms hand-coding eight call
patterns, all destructured positionally with (nth r 0)/(nth r 1) on the
[type node'] tuples infer returns. Split each pattern into a named helper
(infer-pred-fold/-kw-lookup/-get-lookup/-reduce-hof/-seq-hof/-conj-into/-call)
behind an infer-invoke dispatcher that keeps the cond guards verbatim, and add
ty/nd accessors for the tuple so a silent transposition can't hide.
The accessors are applied only to genuine infer results (the new helpers and
infer-fn-seeded); the :map/:let/:loop arms interleave non-infer pairs
(binding tuples, accumulator pairs) with infer results, so those keep nth.
Pure restructuring — the guards, order, and bodies are unchanged; seed
re-minted to the byte-fixpoint, gate green, 0 new corpus divergences.
The inliner and the type inferencer each recognized (:k m) and (get m :k)
lookups with their own copy of the callee tests — the get-callee check was
duplicated verbatim across both. Lift kw-callee?/get-callee? into
jolt.passes.fold (alongside scalar-const?) and call them from both passes so
the head recognition can't drift.
Only the head predicates move. The deliberate differences stay: the inliner
still accepts any scalar key in the get-form (its scalar-replacement targets
can be string/number-keyed maps) while the inferencer requires keyword keys
for struct field typing, and the inferencer keeps its two arms separate so each
rebuilds args for its form. The backend's value-as-fn ifn-kind is left alone.
Two small clarity moves from the review:
- dynamic-vars.ss -> dynamic-var-defaults.ss. It holds the default VALUES of a few
core dynamic vars (*clojure-version*, *assert*, …); the near-identical dyn-binding.ss
holds the binding-stack machinery. The names were easy to confuse.
- Pull the ~60-line %-format engine out of the natives-misc.ss grab-bag into
natives-format.ss (its ->long/pad-left/fmt-float helpers were local to it).
rt.ss loads + MODULES.md updated. Runtime .ss, no re-mint; make test green, format +
the dynamic vars verified.
natives-parity.ss was a grab-bag (its own header called it 'parity'). Dissolve it
into homes that say what they hold:
- cat -> natives-transduce.ss (was natives-xform.ss, renamed for the
whole transducer surface: volatiles + cat + transduce/sequence)
- transient? -> transients.ss
- rseq -> natives-seq.ss
- hash family -> natives-misc.ss (the public hash API over jolt-hash)
The remainder — the #?() feature set, the reader-conditional + re-matcher tagged-map
ctors, and macroexpand — is a coherent reader/macro runtime-support unit, kept as
natives-reader.ss (np- prefix -> nr-). rt.ss loads + two comment refs updated.
Runtime .ss, no re-mint. make test green; hash/transient?/rseq/cat/macroexpand and
#?() reader features all verified.
The trio split on a fine axis (registry core / statics / object classes) but the
names didn't say so — 'static-statics'/'static-objects' and headers that read
'Continues X'. Rename:
host-static-statics.ss -> host-static-methods.ss (Class/member statics + fields)
host-static-objects.ss -> host-static-classes.ss (instantiable object classes)
host-static.ss stays the registry core. Headers rewritten to state each file's role
and what it covers instead of chaining. rt.ss loads + the one comment reference +
MODULES.md updated. No code moved; runtime .ss, make test green.
Under --direct-link a top-level def binds jv$<fqn> and app->app calls bound directly
to it. emit-invoke raw-applied that binding for any var callee, but only a fn-valued
def is a Scheme procedure: (def cfg {...}) then (cfg :a) emitted (jv$cfg :a), applying
a pmap -> "attempt to apply non-procedure". Maps/sets/keywords are invokable in Clojure
via jolt-invoke, which the indirect path used, so this only bit closed-world builds.
Track which direct-linked vars hold fn literals (direct-link-fns, registered at the def
site when the init op is :fn) and only raw-apply those. A non-fn callee falls through to
the jolt-invoke branch, which still uses the direct jv$ binding as the invoke target —
so the var-deref is still skipped, just not the dispatch.
Seed source: re-minted. Regression in directlink-test.ss (jolt-cw1o).
The printer's two entry points (jolt-pr-str in rt.ss, jolt-pr-readable in printing.ss)
get register-pr-str-arm! / register-pr-readable-arm!, plus register-pr-arm! for the
types whose str and readable forms match (bigdec/inst/uuid/tagged/record/ns/var). The
normalize arms (sorted, lazy-seq, queue) and the uri readable arm register per-printer.
Also folds in the hash (dyn-binding var-cell), class (io uri/uuid/file), and get
(transients) arms missed earlier.
natives-array's get stays a case-lambda wrapper on purpose: its 2-arg path errors on
an out-of-bounds index while the 3-arg path returns the default, an arity distinction
the (coll k d) registry collapses — left as-is to preserve behaviour.
Completes jolt-lmot: all six dispatchers (hash/class/get/=/pr-str/pr-readable) off the
set!-rebind chains. make test green, 0 new corpus divergences; pr-str/str of inst,
uuid, bigdec, sorted-map, record-with-lazyseq, queue all verified.
jolt-get (4 sites: host-table/natives-misc/inst-time/records) and jolt=2 (6 sites:
records/vars/inst-time/natives-misc/bigdec/host-table) move off the set!-rebind
chains to register-get-arm! / register-eq-arm!. get's case-lambda becomes a stable
2/3-arg entry over a 3-arg dispatch; the equality arm pred is (a b) since either arg
may carry the type, and the host-table sorted arm normalizes-then-re-dispatches.
Behaviour-preserving (runtime .ss): var/inst/bigdec/record/uuid equality, record!=map,
sorted-map=plain-map, and all the get cases verified; make test green, 0 new corpus
divergences. Four of six dispatchers done; the printer (pr-str/pr-readable) remains.
Replace the set!-capture-and-rebind chains extending jolt-hash (3 sites: inst-time/
host-table/records) and jolt-class (3 sites: bigdec/natives-queue/host-table) with a
register-hash-arm! / register-class-arm! registry (the pattern already used by
register-str-render!). The base dispatcher walks its arms — disjoint types, so order
is immaterial — then falls to the base cases. The entry is stable, so the per-site
(def-var! "clojure.core" "class" …) re-points are gone.
Behaviour-preserving (runtime .ss, no re-mint): jinst hash, bigdec/queue/record
class, record hash, and a sorted-map hashing as its plain map all verified; make test
green, 0 new corpus divergences. First two of six dispatchers; get/=/pr-str follow.
set-optimize!/set-direct-link! are process-global flags in the back end, set then
reset around the emit. A strict-form build error (failing forms error the build by
design) skipped the reset, leaving the compiler in optimize/direct-link mode for any
later in-process caller. dynamic-wind guarantees the revert. Behaviour unchanged on
the success path; both --tree-shake and --opt --direct-link build + run identically.
Tree-shaking was split across emit-image.ss (the dce-* helpers + record producer)
and build.ss (bld-shake-all + the manifest splice), with the DCE record accessed by
raw (vector-ref r 0..3) in ~10 places and the manifest splice/drop driven by
substring-matching (load "…prelude.ss").
- New host/chez/dce.ss owns the whole DCE concern: a named record API
(dce-rec/-keep?/-fqn/-refs/-str — no more positional vector indexing), the ref
extraction + ref-set constants, dce-blob-records, and dce-shake decomposed into
dce-build-graph / dce-reachable / dce-bail-scan / dce-partition (was one 50-line
bld-shake-all doing five jobs with shared mutable state).
- emit-image.ss keeps only ei-emit-ns-records (it drives the ei-* compiler) and uses
the dce-rec constructor.
- The runtime manifest is now tagged ('prelude/'image/'compile-eval); bld-emit-runtime
dispatches on the tag instead of substring-matching file paths, so the core-splice
and compiler-drop can't silently break on a rename/reorder.
Behaviour-preserving (runtime .ss, no re-mint): build-app shakes identically
(56/460, 8.12MB), make test green, make shakesmoke green (4 git-lib apps
byte-identical, same sizes).
