Exercise both real drivers with an earlier single-file decoy and a later canonical package directory, exact sorted unit ownership, direct export composition, deterministic package artifacts, transitive linking, stable file-only rejection, and a directory-package test import.
Build w6a and w6l from package-main directories and expose the wcc backend through a narrow package API so w6c and wwdump no longer import implementation files. Retarget the remaining load-bearing fixtures and example sources to directory packages; retain the one intentional flat compiler collision as an explicitly composed raw unit.
Go has no per-file test targets: the PACKAGE is the unit of testing
(`go test ./...`) and same-package test files compose together. The
59-target LIBRARY_TESTS fan-out (per-file -w workdirs, one Make rule
per suite) collapses to one line: `ww test -j $(JOBS) -w
out/wwbuild/wwtest-lib -I lib/ww lib/...` — the regex dir-route
precedent generalized. The lib/regex line and the per-file pattern
rule instance dissolve with it (test/lang keeps its own rule).
Measured before committing (-j4, strings edit row): old per-file
warm 4.7-4.9s; bare walk 5.2-5.6s — a real regression, so the
coordinator first gained the brief's persistent per-package workdir:
`-w DIR` on a package target forwards to wwpackage, which keys
DIR/<dir>_<pkg> per group and hands it to each inner `ww test -c`
build. Reuse stays entirely with the driver's existing
content-identity contract — the coordinator adds pure path policy,
no cache machinery. Both driver stages drop their package-target -w
rejects (forward instead); -w with -c stays rejected at the
coordinator (two ownership contracts). After: 3.9-4.0s on the edit
row, 1.0s warm no-op, 2.5s cold — faster than the old flow on every
row.
package_test's tree -w reject row becomes the positive contract
(cold+warm byte-stable stream, cs/ww same) plus the -c conflict
reject. libbyteid roster shape DECIDED: per-file fx entries stay —
every enrolled file is still standalone-buildable, so coverage is
byte-for-byte unchanged; the dir-mode entry form arrives only with
the B3 shared-helper split that first needs it. Docs: owner table,
target table, -w contract paragraph.
Go does not distinguish bootstrap observers from unit tests: an
in-process test is a unit test. 738 is pure parsefile calls (arena +
lexinit + parserinit, AST assertions, zero subprocess or artifact
claims) — the 200_parse class, misfiled among the residuals. The
fleet partition now reads six units / six bootstrap / three platform
and the residual class is empty; COMPILER_WRAPPER_SOURCES drops it
by the existing filter-out. Doc roster updated (also refreshes the
corpus prose to the current 1745/3490 pin and retires the stale
794-era paragraph).
Pure move: log_test.ww dissolves along its banner seams into
funcs_test.ww (the six logger-arg rows — lprintlnbasic/single/empty,
lprintflnbasic/indexed/mods — plus the fatal-family subprocess TODO;
ref/hare/log/funcs.ha owns lprintln/lprintfln and the fatal/lfatal
forms) and global_test.ww (defaultwiredtoglobal, setloggerswap,
printflnglobal; ref/hare/log/global.ha owns default/global/setlogger).
Moved blocks are byte-identical; the 9 banner lines are deleted.
Imports minimized per file (log + memio; the stale fmt/io imports
drop).
Helper inventory: streq spans both halves (all six funcs rows;
setloggerswap + printflnglobal on the global side), so it is
duplicated under per-file names per the sanctioned strconv
streq/fstreq precedent — fstreq in funcs_test.ww (decl + 6 call
sites), gstreq in global_test.ww (decl + 6 call sites); bodies
byte-identical apart from the name. Line-multiset diff old-vs-concat
(heads/banners excluded, names normalized): the only residue is the
9-line streq helper counted twice.
Order-safety, proven on the composed run: dir-mode executes
declaration order per file, files byte-lexicographic (funcs_test <
global_test < silent_test). Observed order: lprintlnbasic,
lprintlnsingle, lprintlnempty, lprintflnbasic, lprintflnindexed,
lprintflnmods, defaultwiredtoglobal, setloggerswap, printflnglobal,
silentwritesnothing, silentignoresprintfln — defaultwiredtoglobal
still precedes every log.setlogger caller (setloggerswap and
printflnglobal, both later in its own file; funcs and silent rows
never call setlogger).
Consumers: Makefile LIBRARY_TESTS replaces the log_test.ww entry with
the two new entries in place; test/byteid/libbyteid_test.ww roster row
becomes two fx rows, NENTEXPECT 59->60.
Validation: standalone ww test lib/log/funcs_test.ww (6 passed) and
lib/log/global_test.ww (3 passed); dir mode ww test lib/log (11
passed, no duplicate symbols); byteid compile-only build exit 0, 60
roster appends == NENTEXPECT.
Pure move: bufio_test.ww dissolves along its banner seams into
stream_test.ww (13 stream* rows) and scanner_test.ww (13 scan*/
newscanner*/emptystream/errorsource/boundarycases/multifill/
readaheadcantgrow rows), mirroring ref/hare/bufio/{stream_test+test.ha,
scanner_test+test.ha} ownership. streamscannerunread stays stream-side:
Hare's stream-level unread test lives in stream_test+test.ha:86.
Moved blocks are byte-identical; the 28 file-scope banner lines (and
their // continuation lines) are deleted, the 2 interior banners inside
boundarycases stay. Imports minimized per file (the unused `import
bytes;` drops; utf8 is scanner-only).
Helper inventory: putstr spans both halves (7 scanner rows / 8 call
sites; 8 stream rows / 8 call sites), so it is duplicated under
per-file names per the sanctioned strconv streq/fstreq precedent —
scanner_test.ww keeps `putstr`, the stream copy is `sputstr` (decl,
head comment, 8 call sites; body otherwise byte-identical). The
error-returning source (errvt/errread/errwrite/errsource) moves with
its sole consumer errorsource into scanner_test.ww; the close-recording
source (srcclosed/closevt/closesink/closewrite/closeclose/closesource)
moves with streamclosenopropagate into stream_test.ww. Line-multiset
diff old-vs-concat (heads/banners excluded, sputstr normalized): the
only residue is the 10-line putstr helper counted twice.
Consumers: Makefile LIBRARY_TESTS replaces the bufio_test.ww entry with
the two new entries in place; test/byteid/libbyteid_test.ww roster row
becomes two fx rows, NENTEXPECT 58->59.
Validation: standalone ww test lib/bufio/scanner_test.ww (13 passed)
and lib/bufio/stream_test.ww (13 passed); dir mode ww test lib/bufio
(26 passed, no duplicate symbols); byteid compile-only build exit 0,
59 roster appends == NENTEXPECT.
taggedcallslot gated on an N_IDENT callee + fnretlookup, so every
N_DOT / wwi-decl'd cross-module callee (and every fn-pointer call)
returned 0 and its tagged call-result ARG fell off the cursor-push
path into a 16B scratch spill where cstage PUSHQes AX/DX (cstage keys
the same push on args[i]->type, cgen.c tagged_arg_size). Read the
stamped result type instead — the #209/#211 predicate-to-stamp
discipline, same class as the fn-value cgident fix.
Graduates BOTH remaining DATABYTEID_DIVERGED pins
(r700_strings_byteindex frame 80->64, r839_xmod_nominal_match frame
32->16) — the ledger is EMPTY; the byteid gate ran 1389 data fixtures
at 0 pinned-divergent. Lang rows cover the same-module control,
fn-pointer, float-member, and str-member shapes (byte-identical each).
Separate latent lead banked, NOT closed here: the tagged
subset-widening call arg ((i64|bool) passed as (i64|bool|void))
byte-diverges pre-existing — cstage stages through a 16B local,
wwstage passes direct; runtime-equivalent.
Pure move: utf8_test.ww dissolves along its banner seams into sibling
rune_test.ww (runesz/utf8sz), encode_test.ww (encoderune),
decode_test.ww (decoder next/prev/validate/remaining/slice/position
plus the encode->decode round-trip), and types_test.ww (strerror),
mirroring ref/hare/encoding/utf8/{rune,encode,decode,types}.ha
ownership (types.ha:12 owns strerror). Blocks are byte-identical;
only the banner lines are deleted. The streq helper moves with its
sole consumer, strerror_cases, into types_test.ww. The impl utf8.ww
is untouched (no banners).
Consumers: Makefile LIBRARY_TESTS replaces the utf8_test.ww entry
with the four new entries at the same position; the
test/byteid/libbyteid_test.ww roster row becomes four fx rows,
NENTEXPECT 52->55 (+3).
Pure move: bytes_test.ww dissolves along its banner seams into sibling
equal_test.ww, index_test.ww (u8 + []u8 arms of index/rindex),
contains_test.ww (contains/hasprefix/hassuffix), tokenize_test.ww
(tokenize/rtokenize/peektoken/remainingtokens + splitn/rsplitn/split +
cut/rcut), and trim_test.ww, mirroring ref/hare/bytes/{equal,index,
contains,tokenize,trim}.ha ownership (Hare keeps splitn and cut in
tokenize.ha; contains.ha owns hasprefix/hassuffix). Blocks are
byte-identical; only the banner lines are deleted. Helpers stay with
their sole consumers: expect_token/expect_done/expect_tok in
tokenize_test.ww, beq in trim_test.ww.
Consumers: Makefile LIBRARY_TESTS replaces the bytes_test.ww entry
with the five new entries at the same position; the
test/byteid/libbyteid_test.ww roster row becomes five fx rows,
NENTEXPECT 48->52 (+4).
Pure move: the hasprefix/hassuffix, contains, byteindex/rbyteindex/
index/rindex, and compare sections of strings_test.ww become sibling
suffix_test.ww, contains_test.ww, index_test.ww, compare_test.ww,
mirroring ref/hare/strings/{suffix,contains,index,compare}.ha
ownership. Blocks are byte-identical; only the section banner lines
are deleted (the file names now carry them).
The remaining families (dup, concat, trim, sub, utf8, iter, tokenize,
pad, replace, cut) stay in strings_test.ww: they all consume the
file-local streq helper, and a helper cannot be shared across split
test files — a single-file test root does not pull same-package
siblings (undefined: helper), while dir-mode composes all package
test files into one unit (duplicate fn helper). Splitting them would
force either helper duplication or renames, both out of scope for a
pure move.
Consumers: Makefile LIBRARY_TESTS gains the four new entries beside
the strings_test.ww row; test/byteid/libbyteid_test.ww roster gains
four fx rows, NENTEXPECT 44->48 (+4).
hton/ntoh -> network.ww (ref/hare/endian/network.ha:5-36); be* byte
get/put -> big.ww (ref/hare/endian/big.ha:5-48); le* -> little.ww
(ref/hare/endian/little.ha:5-48). ww does not port endian.ha's host
dispatch, so no host file appears.
Pure move: every fn is byte-identical; the three section banner
lines are deleted (file names now carry them). The module-doc header
rides with network.ww — its amd64-is-LE sentence is the WHY for
hton being a swap.
Consumers: Makefile WWFIXTURE_SRC entry replaced in place. The
byteid roster is untouched: lib/endian is enrolled as an
import-probe keyed on `package endian;`, which every new file
still declares.
Mirror ref/hare/path's impl layout in the test siblings:
abs/isroot/string/local rows -> buffer_test.ww (ref/hare/path/
buffer.ha:28-77); push/parent/pop-split rows -> stack_test.ww
(ref/hare/path/stack.ha:9-180); dirname/basename rows ->
posix_test.ww (ref/hare/path/posix.ha:17-71).
Pure move: every @test block is byte-identical; only the section
banner lines are deleted (file names now carry them) and each
section's ref-cite WHY comment rides with its @test. The suite
header's Hare-vector/parallel-array WHY paragraphs ride with
stack_test.ww (push is their subject); the stale `ww run
lib/path/path_test.ww` sentence is dropped, not relocated.
Consumers: Makefile LIBRARY_TESTS entry replaced in place. The
byteid roster is untouched: lib/path is enrolled there as an
import-probe, not a fixture.
Mirror ref/hare/time's impl layout in the test siblings: constants ->
duration_test.ww (ref/hare/time/duration.ha:9-18 owns the constants);
now/sleep rows -> instant_test.ww (instant is the subject type,
ref/hare/time/instant.ha; now/sleep bodies live in
ref/hare/time/+linux/functions.ha, a platform split ww does not have);
add/diff/compare rows -> arithm_test.ww (ref/hare/time/arithm.ha).
Pure move: every @test block is byte-identical; only the section
banner lines are deleted (the file names now carry them). The stale
file header naming the retired `ww run lib/time/timetest.ww` entry
point is dropped, not relocated (rule 8: it described a WHAT that no
longer exists).
Consumers: Makefile LIBRARY_TESTS entry replaced in place;
test/byteid roster 44->46 (+2).
Pure file-boundary move: the two silent-logger rows
(silentwritesnothing, silentignoresprintfln) leave lib/log/log_test.ww
for lib/log/silent_test.ww, mirroring ref/hare/log/silent.ha's
ownership (ref/hare/log has no test files of its own; the file name
follows the impl layout as directed). Only the two banner lines are
deleted; both moved blocks are byte-identical. silent_test.ww imports
log + memio only (neither row uses streq/fmt/io).
The specced funcs_test/global_test halves stay merged in log_test.ww:
the streq helper is consumed by six lprint* rows AND the
setloggerswap/printflnglobal global rows, and same-package test files
compile as one unit in dir mode (duplicate fn rejected), so that
boundary cannot be a pure move; reported upstream rather than
relocating the majority of rows.
Ordering: defaultwiredtoglobal must precede any log.setlogger caller.
All setlogger callers (setloggerswap, printflnglobal) remain after it
in log_test.ww; the silent rows never touch setlogger, and dir-mode
concat order (pkgsort: log_test.ww < silent_test.ww) keeps
defaultwiredtoglobal first overall.
Consumers: Makefile LIBRARY_TESTS gains lib/log/silent_test.ww beside
the log_test.ww entry; test/byteid/libbyteid_test.ww roster gains
fx("lib/log/silent_test.ww") after the log_test row, roster 44->45
(+1, NENTEXPECT bumped).
Validation: standalone ww test lib/log/log_test.ww (9 passed) and
lib/log/silent_test.ww (2 passed); dir mode ww test lib/log (11
passed, defaultwiredtoglobal first); test/byteid/libbyteid_test.ww
compile-only build exit 0, 45 roster appends == NENTEXPECT.
Pure file-boundary move of lib/io/stream.ww's single 'empty stream'
banner section (_empty_read, _empty_write, _empty_vt, empty) into
lib/io/empty.ww, mirroring ref/hare/io/empty.ha. No code, name,
signature, or behavior change. Comment-only adjustments: the banner
line is deleted (the file name carries it); the moved block's
'Lives in stream.ww' self-reference now says empty.ww; stream.ww's
head drops its empty bullet (content duplicated by empty.ww's own
comments); io.ww's ownership map gains the empty.ww line; types.ww's
'three files' count becomes four. empty.ww needs no imports (package
io types only).
Makefile: lib/io/empty.ww added beside every lib/io/stream.ww
occurrence (WWFIXTURE_SRC line 75; w6c_ww/wwdump_ww bootstrap prereq
lists at lines 184/204). lib/io has no _test.ww, so no LIBRARY_TESTS
or byteid-roster change; byteid coverage rides the bootstrap gates.
Validation: out/bin/w6c lib/io/io.ww (standalone contract) exit 0;
importer tests green: bufio 26, memio 11, log 11, getopt 12 (calls
io.empty), fmt 49.
wwstage dotchainresolve gated its *T roots on an N_TPTR-over-N_TNAME
tnode; an inferred `let p = alloc(S{...})!` rides the checker-
SYNTHESIZED N_TPTR whose pointee is the struct BODY node (the #24
TNAME-normalize covers only the direct struct-lit binding), so the
root gate failed and chained p.field.pseudo reads fell to the
unfused deref arms — shape-only divergence vs cstage's type-keyed
fold (both stages runtime-correct; the historical field(SB) leak the
pin described was already fixed). Widen the local and global root
gates to accept the N_TSTRUCT pointee; correctness stays enforced by
the stamped-tinfo peel below (TY_PTR -> pointee TY_STRUCT).
Graduates alias_g73_heapfill out of DATABYTEID_DIVERGED (2 pins
remain, both the tagged-spill family); adds compile fixtures pinning
the newly-converged siblings (nested read+store, slice .cap,
inferred-global root; pin 1730/22/3460) and a runtime lang row that
mutates then re-points the heap base — a read bypassing the pointer
returns the wrong len.
The three modules pass bare (loud SKIPs on absent env); their arranged
legs stay owned by test/libenv. Closes the deliberate scope cut from
the os wave.
test/libenv/libenv_test.ww owns the env-arranged legs of the lib/os
getenv and lib/dirs XDG suites (rows asserted ok, not SKIP, plus the
created-chain layout), the dirs too-long abort on both stages, and the
make-owned runner for the self-arranged stat suite. Wired beside the
sep observers under test-compiler.
Strengthens the pre-symmetry-era cstage-only stdout scrape to both
frontends with gated exits and the rule-10 cs==ww byte-id; the libc
binding smoke likewise runs both stages.
One commit for both carriers: the asserttyped manifest and the -c/-r
parse-error gate are the same wwstage-only wwdump_ww subprocess
surface. The 901_*.ww raw gap companions stay in test/wcc, fed via
testenv.repo().
One commit for both carriers: flag-misuse parity and the
missing-package fatal are the same driver-CLI subprocess surface. The
-V oracle parses WW_VERSION out of cmd/wcc/ww.h at run time; the
missingpkg ww_ww twin stays owned by bootstrap carrier 993_ww_ww.c.
Ports 708_param_shadow_mod, 712_redecl and 961_opaque_guards into
test/tool/rejects_test.ww on testenv, one commit because the three
carriers share the single asymmetric-polarity claim (cstage rejects,
wwstage accepts) that a both-stage //ww:error fixture cannot carry.
The neg_selfimp leg drops: test/wcc/data/r948_selfimport/case.ww owns
the identical both-frontend claim.
The unowned remainder is the four run-stderr assertions — msg content
for fail_msg/abort_msg, MANDATORY EMPTINESS for fail_nomsg/abort_bare
(rt_abort(NULL,0) writes nothing) — which the fixture protocol cannot
see (run exit only). The rows build the r957_* fixture sources in
place via the cstage driver; byte-id extends the lowering to wwstage
(the carrier's own argument). Owned legs cited, not duplicated: all
eight exit rows by the r957_assert_*/r957_abort_* fixtures, the
checker rejects by r957_assert_reject_*, byte-id by the
test-data-byteid blanket.
Only the ast_proof leg was unowned: wwdump -a / wwdump_ww -a dump
byte-id plus the ordered dot-over-structlit-over-tname node shape for
a trailing op on a qualified struct literal (inexpressible as a
runtime scenario while the #77 literal-temporary cgen gap holds; the
multilevel row pins the source-order a.b.C tname flatten). The three
build/run scenarios are owned by the r84_qualstruct_* fixtures plus
the test-data-byteid blanket — their vestigial want_exit fields were
already dead in the carrier.
954_tuprecv_run, 956_tuprecv_f64_run and 956_modqualdestr_run shared
one retained dimension: w6c_ww compiles the destructure shape cleanly
AND emits no asserttyped stderr diagnostic (#121 A-narrow / #6a-A
stamp nets) — a channel neither test/lang nor the fixture protocol
has. All 11 rows fold into one observer with the drew-C4 non-vacuity
self-check on the shared predicate. Owned legs cited, not duplicated:
value rows + byte-id ride test/lang/tuprecv_test.ww,
test/lang/tuprecv_f64_test.ww and the test-lang-byteid blanket; the
modqual values + byte-id ride the r956_modqual_* fixtures and
test-data-byteid — those four rows compile the fixture files in place
so probe and fixture cannot drift. Bundled: one file, one concern
(the wwstage stamp-absence dimension).
826_alias_tuple_coerce, 944_alias_accept_run and
944_variant_chain_b95_run all reduced to the same irreducible shape: a
LIVE per-stage asymmetry (probed at HEAD) no symmetric fixture or
@test can host. Ported under the DIVERGE discipline — both sides
asserted loud (826 neg_0: cstage "not assignable to declared pair"
vs wwstage over-accept; #96: cstage runs vs w6c_ww "unsupported
address-of shape"; #277: cstage runs vs ww_ww "deferred #277"), so
a fix on either side fails the row and demands graduation. Owned legs
cited, not duplicated: r826_alias_tuple_* byte rows ride
test-data-byteid, 826's neg_1 went symmetric at HEAD and moves to a
//ww:error fixture per the audit split, the 944 value/reject corpora
live in test/lang/{alias_accept,variant_chain_b95}_test.ww and the
alias_*/vchain_* fixtures. Bundled: one file, one concern (held
divergence pins).
Both carriers were already slim: the r76_def_widen_* byte-id legs are
owned by the test-data-byteid blanket, 783's positive byte rows by the
r78_amp_fn_assign_* fixtures plus that same blanket, and its symmetric
non-tagged rejects by the corpus //ww:error fixtures. The unowned
remainder is three cstage-only loud-rejects (the 760 narrowing def-ref
and 783's two tagged-slot rows); the wwstage over-accepts (no def-init
assignability check; #214 void-variant leniency) are documented at the
site per the residual audit, never asserted as desired — the rows
graduate to //ww:error fixtures when the wwstage side tightens.
Bundled: one file, one concern (cstage-reject pins awaiting wwstage
graduation).
wwiroundtrip pins the @packed re-emit through a real -I tree on both
driver stages (importer-side layout exit 91), strengthened with the
`struct @packed {` needle in the retained sepwork pk2.wwi (a
single-file two-package form emits no .wwi, so the tree is
irreducible). identityreject pins the stage asymmetry as a DIVERGE
row: cstage rejects with "not assignable" (packed is type identity,
harec types.c:621), wwstage builds+runs through the documented
#224/#10 nominal-lossy leniency; a wwstage tighten fails the row and
demands graduation to a //ww:error fixture. The 15 layout/runtime
value rows are fixture-shaped and move to the declarative corpus per
the residual audit port split.
test/misc hosts the residual mixed-shape observers (divergence pins,
stamp/stderr probes, .wwi trees) on testenv, wired into test-compiler
beside the sep/xmod/asm families; its rule declares both full
per-stage toolchains plus the wwdump twins for the coming files.
The 689 port keeps the both-driver run-exit-6 backstop and the
per-stage absolute DATAW row count (linker dedup and byte-id sweeps
are both blind to a symmetric double-emit).
The 18 active rows keep compile + cs==ws byte-id and the tab-anchored
ADDQ $N,SP cleanup markers with both polarities (mem-class rows must
contain their slot's cleanup; the exactly-48B boundary row must not).
The per-row driver build+run legs (incl. the uninit-global zero-box
exit 91) move to the r929 run fixture corpus per the residual audit
port split; the four buildfail rows were already dead here, owned by
the r929_memarg_reject_* fixtures.
The 13 active rows keep compile + cs==ws byte-id and the hidden-RDI
"(BP), DI" presence/absence polarity (present iff sret-class; the two
exactly-32B boundary rows must stay register). The per-row driver
build+run legs move to the r926 run fixture corpus per the residual
audit port split; the seven buildfail rows were already dead here,
owned by the r926_sret_reject_* fixtures.
The five symmetric rows compile their r764_amp_fn_ident_* fixture
sources in place (runtime owned by those fixtures; pre-fix junk-store
still exits 0, so the LEAQ window is the discriminator) and keep the
cs==ws byte-id leg. The cross-module row stays cstage-only pending
#184; its wwstage/byte-id legs graduate with that task.
asmgate_test.ww absorbs 530_w6a_parsenum and 989_datargate_run: the
strtoll-semantics identity rows plus the bare-minus reject against
both assemblers with cross-stage .o byte-identity, and the undefined
DATAR-slot loud reject (no .o written) with the defined-slot
byte-identical ok path. The w6a_ww skip gates drop: the Make target
declares both assemblers.
