You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
49 lines
64 KiB
49 lines
64 KiB
0.25.0 |
|
#First line must be a semantic version number |
|
#all other lines are ignored. |
|
#0.25.0 - G-053 allocation participation (directed work, post-G-053): bounded -multiple occurrence ranges now PARTICIPATE in positional allocation for leaders and values instead of being post-loop enforcement only. Previously greedy collection overran a bounded max unless a type screen happened to stop it (untyped {3 3} followed by a -multiple tail collected 4 words and occurrencecount then reported the overrun the allocation itself created), and a later REQUIRED ranged argument was starved to a single reserved clause. Three cooperating sites: (a) get_dict_can_assign_value caps a satisfied -multiple argument at its bounded max - it yields the word (no-consume) so the caller's retreat advances to the next argument, serving both the leaders and values loops; the yield carries a G-082 rejection record of new kind 'occurrence', and both loops' overflow selection sites render it as the pointed occurrence-limit report with the G-053 occurrencecount errorcode (count = the attempted occurrence) instead of the generic toomanyarguments - so genuine over-supply keeps its pointed class; (b) reservations are min-occurrence aware - the allocator's tail_needs and the derived valmin floor (leader/value split + option-scan reservation) reserve min-occurrences x min-clause-length for a required ranged argument (e.g a following required -multiple {3 3} reserves 3 clauses; {2 2} pair clauses reserve 4 words); an explicit '@values -min' still overrides the derived floor, and legacy boolean -multiple (no MULTIPLE_RANGES entry) reserves exactly as before; (c) the greedy leader scan caps a bounded -multiple last leader at max occurrences (tracked via leader_posn_names_assigned) so remaining words flow to the opts/values sections. First occurrences are never capped (resolve guarantees range max >= 1); {0 1} at-most-once scalars are unaffected (no collection). define -help -multiple documents the allocation semantics. New multipleranges.test allocation pins: untyped/typed {3 3} cap incl the cap-beats-type-screen 4-ints case, greedy-up-to-max {1 2}, required-range and pair-clause reservations, leaders-side cap + split floor, pointed over-supply report, legacy-greed-unchanged guard. Full punk/args suite 408/0. |
|
#0.24.0 - G-084 increment 2 (parsekey completeness, part 2 - the settled cross-member design + integrity closes): -multiple on a member of a shared-parsekey OPTION group is now a DEFINE-TIME error unless ALL members belong to one @opts group marked -parsekeymode error (whose G-083 mutual exclusivity makes per-member collection well-defined; a single -multiple member with its own parsekey is unrestricted). This settles the G-084 cross-member accumulation design decision as rejection rather than accumulate-in-received-order: cross-member collection on one storage key is ambiguous (collect-vs-replace undefined; one member's collected list would silently displace another's). The check runs AFTER the G-053 -multiple canonicalisation so range forms get correct verdicts: {0 1} is replace-shape and allowed on a shared group, collect shapes (boolean 1, max>1, unbounded) are rejected. Value -parsekey integrity settled at define time: duplicate value parsekeys are rejected (trailing values consume positionally so distinct values sharing one result slot silently overwrote each other), and a dash-led value parsekey is rejected (it landed in the options namespace of the result/received dicts and ABORTED parse with an internal error in the option-shaped validation paths). parse_status/parse_report storage-key attribution: parse_status_build now resolves a storage key claimed by exactly ONE argument (a value's -parsekey, an option's declared -parsekey or its '='-suffixed trimmed tail) back to that argument, so argstatus/Source report received with the value-in-effect instead of default/0; shared-parsekey group members stay unresolved (per-member attribution does not survive the storage fold - the remaining recorded G-084 display gap, needs engine-level per-member tracking in the parse result; parse_report's display bridge still shows the group value on each member row). undefine/undefine_deflist now return "" (previously returned the final 'dict unset' expression - the WHOLE cache dict, hundreds of KB, to any caller capturing the value). define -help now documents -parsekey (storage-key rename for options and values, leader rejection, shared groups and -parsekeymode interaction, last-defined-member-wins defaults precedence, the -multiple rule, value parsekey uniqueness/non-dash rules). Internal: dead values-loop identifiers removed (val_ident/val_ident_is_parsekey/values-side seen_pks - assigned per word, never read since the api_valname keying landed), the per-parse parsekey reverse map now iterates only VAL_NAMES (leader map removed entirely - a leader -parsekey is rejected at define time), a dead parsekey-from--default derivation removed in resolve, stale pre-G-084 comments corrected. parsekey.test: parsekey_shared_key_multiple_rejected strengthened (message pins, {0 1}-allowed/{0 4}-rejected range coverage, the -parsekeymode error escape hatch incl collection and optionconflict), new parsekey_value_parsekey_integrity; parsereport.test parsereport_storagekey_attribution + parsestatus.test parsestatus_storagekey_attribution pin the attribution (incl the shared-group default/default remaining-gap row). Full punk/args suite 401/0. |
|
#0.23.0 - G-084 increment 1 (parsekey completeness, part 1): value -parsekey is now supported (was: accepted at define time but aborted parse). A value record declaring -parsekey now renames its result `values` slot and its `received` key to the parsekey (e.g `@values {v1 -parsekey renamed -type string}` parses `hello` to `values={renamed hello}`, `received` keyed by `renamed`), and -multiple value members collect under the parsekey (e.g the `variable` setvalues form `?name value...?` pairs collect under `name_value`). The fix keys VAL_DEFAULTS and the values_dict init/storage by parsekey (matching the already-parsekey-keyed VAL_REQUIRED), tracks the parsekey in a new api_valnames_received list (the internal valnames_received stays argname-keyed for the -multiple first/continuation gate and G-053 occurrence enforcement), uses api_valnames_received for the required-satisfaction check and the received dict, and adds a per-form val_pk2name/leader_pk2name reverse map so the valmin clause-length fallback and the post-parse validation loop can resolve a parsekey back to its argname for ARG_INFO/arg_checks lookups. -parsekey on a leader is now REJECTED at define time with a clear message (was: silently ignored) - no live caller uses a leader -parsekey, so the goal's no-silent-ignore contract is met by rejection rather than parallel hot-path surgery; the dead req_name derivation in the leaders resolve branch is removed. The @values directive line still rejects -parsekey (a group-default parsekey is not a feature; per-arg value -parsekey is). The defaulted-members precedence rule for a shared-parsekey group (none received) is now documented as last-defined-member-wins (deterministic in definition order) and the in-code `? review` is removed. The tclcore moduledoc `#todo - fix -parsekey for leaders and values` is resolved (value supported, leader rejected). parsekey.test: the value GAP flipped to parsekey_value_result_key_settled, the leader GAP flipped to parsekey_leader_parsekey_rejected, the @values-line GAP split to parsekey_values_line_rejects_parsekey; the defaults GAP flipped to a settled pin. dev/parsekey-knownbugs.test: the value and leader disabled pins retired (settled in parsekey.test); the cross-member -multiple accumulation pin remains (increment 3). Full punk/args suite 399/0 (1 skipped), punk/ns 125/125, broader punk sweep 989/0. |
|
#0.22.0 - G-053 range-valued -multiple (occurrence arity): -multiple now accepts a {min max} range (max -1 = unbounded) alongside the legacy booleans 0/1, so a definition can declare at most once ({0 1}, a second occurrence is a parse error), bounded repetition ({2 4}), or one-or-more ({1 -1}) instead of choosing between silent last-wins (0) and unbounded collection (1). The spec compiler canonicalises once at resolve: the stored -multiple becomes the computed boolean (list-shape collect: true for max>1 or -1, false for legacy 0 and max==1) so every existing collect-vs-replace / scalar-vs-list / leader-value-single-multiple hot-path truth-test stays correct, and the range companions (min/max/maxbounded) live in a separate per-form MULTIPLE_RANGES dict (NOT in ARG_INFO, so they do not ride along when ARG_INFO is round-tripped as a spec via resolved_def copyfrom). Legacy 0/1 (and boolean strings true/false/yes/no) are coerced to the boolean and stay byte-unchanged (no MULTIPLE_RANGES entry for unlimited cases). Resolve validation: max must be positive or -1, min <= max, and the -optional/range-min reconciliation - a non-zero min forces presence so -optional set is a contradiction (reject with a clear message; declare -optional 0). Parse enforcement: a new PUNKARGS VALIDATION occurrencecount failure class (payload <argname> count <n> min <m> | max <m>) fires in a single post-loop pass per section (opts/leaders/values); over-max is a hard contradiction (fires in both normal and viability-probe modes, parse_status_classify maps it to invalid), under-min is pure end-of-input exhaustion (SUPPRESSED in the G-152 viability probe and classified incomplete so a viable form reports incomplete, not invalid). The usage-table Multi column reflects the range (0-1 / 2-4 / 1+ for unbounded-with-floor; the greencheck stays for legacy 1), the string renderer emits MULTI:0-1 etc., and the synopsis distinguishes at-most-once (?arg?, no ellipsis) from repeating (arg...). -multipleunique/-multipleuniqueset compose with max>1 ranges unchanged. define -help documents the range form and the -optional/range-min rule. New testsuite multipleranges.test (28 tests: 13 define-time canonicalisation+validation, 12 parse-time enforcement incl parse_status verdicts and legacy-required-still-trailingvaluecount guard, 3 rendering); full punk/args suite 399/0, punk/ns 125/125. |
|
#0.21.0 - G-083 increment 3 (argument-relations usage rendering + lsearch moduledoc adoption): the -conflicts and -parsekeymode error vocabulary now surfaces in usage/arg_error and synopsis output, and the lsearch tclcore moduledoc models its documented option incompatibilities with the new vocabulary. Rendering: (a) a per-arg -conflicts list appends a 'conflicts with: <targets>' hint to the argument's help text in the usage table (targets resolved to display names via lookup_optset, so a parsekey target shows its member flag, not the raw parsekey); (b) a named @opts group marked -parsekeymode error is annotated 'mutually exclusive (distinct members may not be combined)' in its group header, while override (default) groups carry no such annotation; (c) the synopsis one-line form carries no conflict detail (conflicts are a usage-table concern, not a synopsis-line one). Characterized in relations.test (conflicts_usage_hint, parsekeymode_error_usage_group_header, parsekeymode_override_no_header_annotation, conflicts_synopsis_no_hint). lsearch moduledoc: -sorted gains -conflicts {-glob -regexp} and -bisect gains -conflicts {-all -not} (per-arg conflicts, the whole group stays -parsekeymode override so -glob/-regexp remain last-wins); the '(documentation incomplete - punk::args fixes required for grouped mutually exclusive options and prefix calculation)' caveat is dropped from the @cmd -help. Pinned in relations.test (lsearch_sorted_conflicts_glob + lsearch_bisect_conflicts_all raise optionconflict, lsearch_glob_regexp_last_wins stays last-wins, lsearch_caveat_dropped). clock clicks unchanged - its active positional-choice definition already models exclusivity (exactly-one-of) and the acceptance clause names lsearch only. Legacy untouched by default (no -conflicts/parsekeymode-error definition carries no new rendering). Full punk/args suite green; tclcoreparity 10/10. |
|
#0.20.0 - G-083 increment 2 (argument-relations parse-time enforcement): the -conflicts and -parsekeymode error vocabulary declared at define time (0.19.0) is now enforced at parse. A new optionconflict failure class joins the PUNKARGS VALIDATION errorcode vocabulary, mirroring optionmissing's shape: {optionconflict <arg_a> <arg_b> received <receivednames>} - it names both offending received arguments, for both per-arg -conflicts violations (any pair, cross-group) and -parsekeymode error group co-occurrence (distinct members of a shared-parsekey group). The check runs in a single post-resolution site in get_dict_form (after the optionmissing/valuemissing block), so the ordinary option path and the mash (short-flag bundling) path share it - both raise identically. Checked against RECEIVED arguments only (defaults never conflict), after prefix/abbreviation resolution, on optset identity (a new optsets_received tracker for options, since flagsreceived collapses shared-parsekey members onto one api_opt). Runs unconditionally - a received conflict is a hard contradiction in any mode, not end-of-input exhaustion, so the candidacy/viability probe reports it as status invalid (parse_status_classify maps optionconflict to invalid, not incomplete). Define-time check added: a group marked -parsekeymode error must also declare a non-empty -parsekey (the strict mode only applies to a shared-parsekey group). Legacy untouched by default: -parsekeymode override is the default and absent -conflicts means no check runs, so the full existing suite (including the pinned parsekey_repeat_ordering last-wins / prepend-defaults idiom) passes unchanged. relations.test extended with 7 parse-time enforcement pins (raise cases, defaults-never-conflict, cross-group, parsekeymode error raise + one-received-ok + override-legacy, parse_status invalid); full punk/args suite green. |
|
#0.19.0 - G-083 increment 1 (argument-relations define-time vocabulary): new per-argument key -conflicts <list> (parsekeys or flag names that must not be RECEIVED together with this argument; checked against received args only at parse time, defaults never conflict - parse-time optionconflict enforcement lands in a later G-083 increment) and new @opts-level -parsekeymode override|error (per named OPT_GROUPS group; error = distinct-member co-occurrence within a shared-parsekey group raises optionconflict at parse; override = legacy last-wins, the default). Both keys are cross-validated at resolve: -conflicts targets must name a defined argname or declared -parsekey; -parsekeymode requires -group and a value of override|error. Define-time integrity hole closed: a -parsekey value colliding with a distinct defined argument's name (one that does not share that -parsekey) is now a resolve error instead of silently forming an implicit shared-key group (parsekey_collides_with_defined_optname_GAP flipped to a define-time error pin in parsekey.test). Legacy behaviour untouched by default (no -conflicts, default -parsekeymode override, no colliding parsekey). New testsuite relations.test pins the define-time vocabulary; full punk/args suite green. |
|
#0.18.0 - G-151 annotated success render: new punk::args::parse_report - landing report for a SUCCESSFUL parse, one row per value-holding argument (received, or filled from -default) in declaration-section order (leaders, opts, values) with columns Argument | Source | Value - the success-side sibling of the arg_error usage table (a flag-like word consumed as a VALUE by position shows attributed to its consuming argument - the make.tcl dry-run 'kitname = punk91 -confirm 0' class). Input: a punk::args::parse result dict (its id key must resolve to a registered definition), or words + 'withid <id>' (the parse is performed first; a rejected line raises the same validation error parse would - only successful parses have a landing report). Source column: received / received xN (-multiple aggregation count) / default - the plain-mode-safe twin of the colour marking (received CLR(check), default CLR(parsekey_hint)); absent optionals with no default-in-effect get no row. Returns: table (arg_error-styled bordered block via textblock; the default), tableobject, string (plain lines 'name (source) = value'; also the degraded no-textblock path), dict (machine form: id/form/rows with class/source/received/positions/multiple/hasvalue/value - never elided). Display elision is type-aware, deterministic and always marked (display-hint honesty): strings single-line-ized (ansistring VIEW -lf 1 -vt 1 -sp 0 - embedded controls/ANSI render visibly so a value cannot disturb the table; real spaces stay spaces) then grapheme-capped via the new ansistring TRUNCATE (punk::ansi 0.2.0) with trailing '...' plus ' (len N)' citing the raw value's string length; -multiple aggregations and -type list values element-capped with '(+N more)' (whole list-quoted elements kept while they fit, first element char-capped if it alone overflows); -type dict values pair-capped with '(+N pairs)'; markers stay intact so cells can exceed very small caps by the marker width; plain-string fallbacks when punk::ansi is unavailable. Width cascade per the G-149 plumbing precedent: caller -valuewidth > new @cmd directive key -reportvaluewidth > built-in 48; 0 means no cap; deliberately never console-derived (deterministic and pipe-safe - the G-007 emit-then-query lesson). Value-in-effect lookup bridges storage-key folds at display level (-parsekey renames, aliased optionsets' last-|-member finalopts fold) so received aliased/parsekey'd arguments keep their row - G-084 owns the storage-key model itself. Failure-side rendering (arg_error/usage) unchanged. make.tcl retiring its interim one-line 'dry-run: line accepted ...' report in favour of this render is consumer follow-through after a bootsupport promotion (not part of this change). New testsuite parsereport.test; define -help documents @cmd -reportvaluewidth beside the other directive-options. |
|
#0.17.0 - G-149 configurable multi-form synopsis hint labels: the per-form hint labels rendered beside a multi-form definition's synopsis lines in the usage/arg_error table (hard-coded ' i -form N ...' previously) are now a template substituted per form - %formindex% replaced with the form's 0-based ordinal, %formname% with its name; an empty template suppresses the labels entirely (the cell keeps just the Synopsis: label, as single-form renders do). Precedence: caller option > definition > built-in default. New arg_error/usage option -formhint <template> (supplied-empty means suppress; usage pulls it from arg_error's definition via resolved_def); new @cmd directive key -formhint <template> for definition-supplied hints (rides the existing arbitrary-key cmd_info merge, so resolved_def round-trips it; not display-deferred). Unconfigured definitions render byte-identical to 0.16.0 (default template reproduces the exact prior bytes, verified by capture diff). Motivation (G-143 origin): script-level (script) ids - e.g make.tcl's - are unreachable from any repl, so the repl-oriented ' i -form N ...' default is not actionable there; consumers can now name their own per-form navigation. The string (non-table) renderer never carried hint labels and is unchanged. make.tcl adoption is consumer follow-through after a bootsupport promotion (not part of this change). New testsuite formhint.test; define -help documents the @cmd key beside the other directive-options. |
|
#0.16.0 - G-073 unavailable choices: new per-argument key -choiceunavailable <list> - names that are recognised but not available in the current runtime/context. Display: entries render among the choices under a dedicated 'Unavailable (recognised, not selectable here)' group heading (table title / string-renderer Group heading - the distinguishing mechanism) with their ordinary -choicelabels notes, and join the prefix-calculation pool (display shortest-prefix marking reflects their reserving effect); synopsis choice-literal display and value-in-effect/goodarg marking never include them (structural: they are not in -choices/-choicegroups, and -default may not name one). Parse: choiceword_match gains an optional 8th argument unavailablelist and an 'unavailable' result key - a word landing on an unavailable name (exact, or unique prefix; case per -nocase) reports unavailable 1 with the name as canonical while matched stays 0, so matched-only consumers treat it as an ordinary non-match; a prefix shared between an available and an unavailable name is a plain ambiguous non-match. Validation rejects such a landing with the tailored PUNKARGS VALIDATION {choiceunavailable <name> choices <list>} error (-badarg/-badval; message names the entry, its -choicelabels note - e.g a version boundary - and the available values; private::unavailable_choice_msg shared by all raise surfaces). The G-071 allocation screen honours the list (optional-arg landings recorded as G-082 rejection kind 'unavailable' and preferred at the overflow raise sites); required args are unscreened as before, so their landing reports via final validation identically. Ambiguous-prefix messages annotate unavailable longer matches '(recognised but unavailable)'. -choicerestricted 0 keeps full passthrough (no tailored rejection). Resolve-time validation: requires -choices/-choicegroups; entries must not collide with choices, -choicealiases alias names, or -choiceprefixreservelist entries; -default may not be an unavailable name. form_literal_affinity/formcheck screens pass the list through (uniform pool semantics). In-tree adopter: punk::args::moduledoc::tclcore 'string is' forward classes (dict on 8.6). New testsuite choiceunavailable.test; tclcoreparity.test gains forward-class pins + the user-sanctioned 'di' strictness exemption; define -help documents the key. |
|
#0.15.2 - G-082 follow-on: a REQUIRED typed value or leader whose present word fails the basic-type allocation screen (int/double/bool/number/dict) now reports the pointed typemismatch (message + errorcode with -badarg/-badval, same shape as the G-082 overflow-site selection) instead of the generic missingrequiredvalue/missingrequiredleader shortfall - the in-loop required-arg raise sites consult the rejection record returned by get_dict_can_assign_value for the current word. Unchanged by design: end-of-input shortages (valuemissing/count classes), a word failing a required literal-typed member, and multi-member clause partials (no rejection record - conservatism preserved); required-arg choice words were never screened (G-071 asymmetry) so choiceviolation reporting is untouched. parse_status consequences: this shape's top-level status flips incomplete->invalid with failureclass typemismatch (the formstatus verdict was already 'invalid' - the G-152 probe confirmed each such 'incomplete' away before reporting; one probe re-parse saved); multiform noformmatch classes unchanged (already invalid), per-form failureclass/message now the pointed typemismatch. parse_status argdoc status-key note updated. Pins flipped deliberately: parsestatus.test typed-value allocation failure, formviability.test single-form record (top-status/formstatus divergence demonstration re-anchored on a required-option shape), punk/ns cmdhelp.test badarg message match; new errorselection.test required-position coverage (value + leader) with unchanged-shape guards (end-of-input, literal-typed, clause partial). |
|
#0.15.1 - G-164 viability-probe alternative-allocation re-probe: when the valmin reservation seats a word that literally names a declared required option into a required value slot at end-of-input, and the resulting optionmissing stands in the G-152 viability probe (a value word consumed, defeating the position guard), the probe now re-parses with the stolen word(s) consumed as the options they name (private::get_dict_form's viabilitycheck argument carries {1 reseat_words} on the re-probe pass) and grants 'incomplete' (viable) on pure end-of-input exhaustion after that re-seating. Cures the 'i hash -t' false negative: the test form classes viable since {-t foo} parses; sibling forms stay invalid. Options still missing after the re-seating are ordinary exhaustion in the re-probe. Runtime seating is unchanged - the re-probe runs only inside the probe, only via reservation-seated words (the trailing-flag-as-value rule is load-bearing; flagvalues_plain_then_dash and the runtime optionmissing path pinned unchanged). A stolen word naming a value-TAKING required option stays conservative (invalid) - re-seated, it seats as a value again at end-of-input. Verdict contract documentation updated in the parse/parse_status argdoc. formviability.test gains 7 pins (re-probe, delete-form, other-missing-options exhaustion, unnamed stolen word, value-taking conservatism, runtime-unchanged, single-form formstatus); the cmdhelp.test GAP pin cmdhelp_GAP_multiform_flagled_testform_false_notviable flips to cmdhelp_multiform_flagled_testform_viable. |
|
#0.15.0 - @values -flagvalues (G-150 follow-on, user-directed): a @values section may declare -flagvalues 1 - an unmatched dash-led word in option position is then consumed as a trailing value instead of raising invalidoption (declared options still match first; the '--' end-of-options marker is unaffected; off by default). Motivation: '--help'-style forms robustly swallowing ignored trailing words of any shape - previously such a word seated as a value only positionally (at end-of-input via the valmin reservation, or within a bounded valmax) and errored 'Unexpected option' otherwise ('{--help -t foo}' on the unconstrained shape). Mechanism: VAL_FLAGVALUES on the form dict (template default false; @values compiler key, boolean-validated at resolve time) short-circuits the unmatched-dash-word fallback in the parse engine. In-tree consumer: punk::auto_exec::hash's help form. New testsuite flagvalues.test (9 tests: seating matrix, declared-option precedence, off-by-default contrast, optionmissing after value seating, key validation). Documented in the module header notes. |
|
#0.14.0 - G-152 suffix-viability verdict: multiform per-form failure statuses (the 'formstatus' records, the noformmatch errorcode's 'classes' list and its -formerrors records) are now SOUND suffix-viability verdicts - a form reports 'incomplete' (viable) if and only if every supplied word was validly consumed as a prefix of that form and the failure is pure exhaustion at end-of-input; anything already supplied contradicting the form reports 'invalid'. Previously the count/missing-required failure classes granted 'incomplete' before final validation ever saw the consumed words, so a wrong action word landed 'incomplete' on a value-carrying sibling form identically to a genuinely-viable prefix ({frobnicate} vs {info} were indistinguishable), and a word rejected by an allocation type screen (missingrequiredvalue) also classed 'incomplete'. Mechanism: private::get_dict_form gains a viability-probe mode (optional 4th arg) suppressing exactly the pure-exhaustion raises - below-minimum leading/trailing count shortfalls and missing-required leader/option/value checks, each position-guarded so appended words could actually reach the missing element (leaders only with no option/value words consumed; options only before value words and any eopts --; values always), plus an adhoc option's value lying beyond the supplied words - while every contradiction raise stands, including final validation of the consumed words (the soundness backstop). Multiform candidacy and parse_status's single-form formstatus record confirm each classify-'incomplete' failure with one probe re-parse before reporting it. Documented conservatism: input ending INSIDE a multi-member type clause reports invalid (clause allocation cannot affirm the partial words). Rendered noformmatch messages mark each form line '(viable - needs more arguments)' or '(not viable)' after the stable "form '<fid>':" anchor; the existing incomplete-first ranking thereby puts still-satisfiable forms first. Contract documented in the parse ('Multiform failure contract' help section) and parse_status (formstatus/status key) argdocs; parse_status's top-level 'status' for single-form parses keeps the engine classification (display-scheme driver) - only formstatus carries the verdict. New testsuite formviability.test (acceptance scenarios, option-position guards, single-form record soundness, clause conservatism, message marking); forms.test verdict pins deliberately updated (ms/idle 'incomplete'->'invalid' for words failing their type screens). |
|
#0.13.0 - G-144: auto-generated synopses for (script)-prefixed constructed ids now lead with the declared @cmd -name (the invocation name - e.g 'make.tcl modules' for (script)::punkboot::modules) instead of the raw id, across punk::args::synopsis full/summary renders and the usage/arg_error synopsis sections (table and string) that funnel through them; explicit @form -synopsis overrides still win, so make.tcl/punkexe displays are unchanged today (their override retirement is a follow-on). The fallback is gated on the (script) prefix only: (autodef) ids are deliberately excluded - their id-minus-tag IS the invocation (arg_error's existing tag strip surfaces it) while their @cmd -name is a display label ('destroy', 'Object: ::x') that must not lead a synopsis line; (shared)/(package)/(default)/(widgetcommand) doc ids declare no invocation-style -name today and render the id as before (a prefixed id with no -name falls back to the id). Real command ids render byte-identical. Retires the 'open generation deficiency' recorded at 0.12.5 (head-render preference for @cmd -name); define's @form -synopsis directive doc and synopsis -help updated. New synopsis.test pins: script-id full/summary/multiform/surfaces, no-name constructed id, autodef unchanged, real-id byte-identical. |
|
#0.12.7 - update_definitions: tag-prefixed id 'namespaces' ((autodef)::x plus script-level id families like (script)::punkboot / (script)::punkexe - the make.tcl G-030 convention, now also the launch-dispatch doc home per G-118 item 10) are handled first-class in the nslist walk. Previously only the (autodef) tag was stripped; other tags fell through to the unqualified-ns branch, emitting a spurious "warning: update_definitions received unqualified ns: (script)" on stderr for EVERY id query and probing a garbage ::(script)... namespace - and, decisively, the namespace_docpackages association was never consulted with the TAGGED key the scan records, so real_id/usage on a registered-but-not-yet-loaded script-level id returned ""/no-such-id even though get_spec resolved it. Now: docpackages lookup with the original tagged key (loads the documenting package - punk::args::usage works for (script)::punkexe::* ids on first query), then generic tag strip for the real-namespace probes, with a bare/tag-only qualifier skipping them silently. Real unqualified namespaces still warn. |
|
#0.12.6 - arg_error error-message placement (user direction 2026-07-18): the supplied error message and any per-form match detail it carries ("Bad arguments ... Candidate forms ..." + per-form failure lines) now render exactly once, BELOW the usage display, for both the table and string return forms. Previously the block rendered above the usage AND was appended again below it (the historical "put original error at bottom of table too"), duplicating potentially long multi-form detail. The bottom append is also now suppressed when the message is empty (usage-only calls no longer gain a trailing newline + empty line). tableobject returns are unchanged (they never carried the message). Also (same user direction, format explicitly not frozen - no external consumers yet): the string renderer's argument rows are now single-line 'name TYPE:... ?DEFAULT:...? ?MULTI:yes?' - the multi-line table-cell type constraints (-minsize/-maxsize/-range(s), previously line-broken mid-row so DEFAULT:/MULTI: landed glued to the last constraint line, e.g 'i -return string dict filter' arg vars) parenthesize onto the base type as TYPE:list(-minsize 2 -maxsize 2), empty DEFAULT/MULTI fields are omitted instead of rendering empty labels (MULTI shows 'yes' rather than the check glyph), and whitespace-only help blocks no longer emit indented filler lines. The <unnamed>/arbitrary-flag indicator rows get the same DEFAULT-omission. Choices also gain structure in the string renderer (previously a flat join left choice names and multi-line choicelabels at inconsistent depths, machine-indistinguishable once ANSI is stripped - only the table/ANSI prefix highlighting disambiguated): each choice renders as a markdown-subset bullet '- `choice`' at fixed indent 6 (the backticked token may carry a trailing choiceinfo marker glyph), the label's first line follows the closing backtick on the bullet line after ' - ', label continuations indent to 8 (2 past the bullet dash), and groups head their bullets as "Group 'name':". The resulting structural contract for parsers: column-0 lines bearing TYPE: are argument rows, indented lines are documentation, '- `...`' bullets are choices. Table cells unchanged (construction shared, restructure applied at the string-path consumption site). Feeds punk::ns 0.9.0 cmdhelp -return text (LLM-oriented plain text). Pinned in rendering.test (rendering_stringmode_singleline_argrows, rendering_stringmode_choice_bullets); errorselection.test message pins gained leading * (message no longer starts the raised string); cmdhelp.test message pins sample the last line. No API change. |
|
#0.12.5 - eg usability: punk::args::eg resolves relative/unqualified command names (via punk::ns::cmdinfo from the calling context when punk::ns is loaded - handles subcommand words too - else global qualification). Rationale: in a namespace where punk::args::eg shadows the global 'eg' alias (e.g whilst in ::punk::args itself) 'eg list' hit the raw-id API and returned empty while 'i list' resolved - Tcl name resolution, not an alias deficiency. Unknown ids now return a message naming the id instead of empty; punk::ns::eg hands the typed name through when resolution yields no docid so the message can name it. The usage-table 'Example: eg <id>' header row (the short indication an @examples block exists) now uses scheme colour CLR(check) instead of hardcoded white - the raw id is kept (not @cmd -name, which is display text and not always invocable e.g 'Autoloading Built-in: parray'); a frame-embedded bottom-right marker awaits textblock table footer-block support (candidate goal). New examples.test pins: relative-name resolution, unknown-id message. |
|
#0.12.5 - synopsis notation documented + internal @form -synopsis overrides retired. ::punk::args::parse and ::parse_status dropped their @form -synopsis overrides - they had gone stale (omitted -cache/-caller) and their only added value (the -form member hint) is now expressed properly via -typesynopsis {<int|formname>...} on the -form option; the auto-generated lines stay current as options are added (drift pin: synopsis.test synopsis_parse_generated). @form -synopsis remains fully supported for documentors: define's directive help now documents that it is display-only text needing no Tcl-list validity, that display helpers substitute the command head, and the known warranted case (script-level/constructed ids rendering the raw id - the G-030 lesson recorded in src/modules/AGENTS.md; head-render preference for @cmd -name remains an open generation deficiency). ::punk::args::synopsis -help gains a notation legend ([...] optional, italic argname/<type> placeholders, unitalicised literals, (a|b) alternate grouping to disambiguate from -f|--file alias pipes, {a b} multi-word literal protection, --opt=<type>, text*/*text) and a new @examples block comparing Tcl man-page ?...? style (ambiguous when nested - same delimiter both ends) with the bracket style, viewable via punk::args::eg. New synopsis.test pins: non-list custom -synopsis passthrough + ns head-replacement survival, parse generated-line drift pin, examples block presence. |
|
#0.12.5 - parse -cache overhaul. (1) The VALIDATION trap's parse_cache write is now explicitly gated on -cache - previously it was unconditional but only 'worked' by scoping accident ('variable parse_cache' was declared solely on the -cache 1 branch, so -cache 0 failures wrote to a discarded proc-local). (2) Cached validation errors are now stored as per {errorstyle callertext} render variants under a main key of {arglist definition form} (-caller removed from the main key - results don't depend on it): previously the rendered message baked in the first caller's errorstyle and %caller% attribution and replayed them verbatim to later callers wanting a different style/caller (e.g. a minimal-style render replayed to a -errorstyle standard call). A missing variant falls through to a fresh parse whose trap renders and caches it. (3) @dynamic definitions (and legacy leading '-dynamic 1') are never cached - parse -cache 1 is ignored for them so re-evaluated substitutions always take effect (dev/dynamic-cache.test known-bug pin flipped). (4) New punk::args::parse_cache convenience proc (view summary/keys/dict, -action clear - sgr_cache style); parse_cache variable init now guarded with info exists like the sibling caches; parse's -errorstyle definition default corrected to 'standard' matching the operative defaultopts (doc previously claimed enhanced). New testsuite parsecache.test. |
|
#0.12.4 - resolve: retired the (already disabled) blanket 'called with undefined id' stderr warning - it fired on the supported inline usages (parse ... withdef, direct resolve calls) where a first-sighting cache miss is by design. In its place the miss path is now collision-aware: a first-seen rawdef whose @id is already registered with different content purges the id's existing cache entries (quiet undefine) before registering, mirroring define's re-creation behaviour - previously it repointed id_cache_rawdef while leaving the old rawdef_cache_about/rawdef_cache_argdata entries in place, so two rawdefs claimed the same id and a subsequent define with the original text was a silent no-op (stale cache-about hit) leaving withid lookups stuck on the inline definition. Guard tests id_cache_rawdef directly (id_exists also matches aliases, where the rawdef lookup would error). Known residual (comment marker retained): the miss path still records no -defspace, so @dynamic definitions first seen via resolve lack a defining namespace for re-substitution. New test define.test define_resolve_id_shadow (characterized pre-fix: re-define no-op reproduced with the purge disabled). |
|
#0.12.3 - fixed the stringstartswith(*) arm of get_dict_can_assign_value's multi-member tail-clause reservation walk: it matched against $tp (the type string) instead of $rv (the candidate value), so for a later optional clause mixing literal(...) and stringstartswith(...) members, genuinely matching trailing words were never reserved from a greedy -multiple argument (the optional tail clause was silently lost), while a prefix text that happened to prefix the literal type string (e.g 'str' vs 'stringstartswith(str)') wrongly reserved non-matching words and raised a spurious toomanyarguments overflow. Found flagged-only in the 0.12.2 comment review; characterized pre-fix then pin flipped (allocation.test allocation_tailclause_ssw_reservation + literal-only and no-match guards). Note the walk still only runs when the tail clause's typelist contains a literal* member (the pre-existing acknowledged approximation) - a stringstartswith-only tail clause remains unreserved. |
|
#0.12.2 - comment/documentation hygiene pass (no behaviour change). Removed ~1100 lines of superseded commented-out code: the old-signature private::check_clausecolumn, the abandoned _check_clausecolumn2 'interim version' experiment (its one keeper idea - collect per-clause validation results and report at end instead of raising on first failure - recorded in an in-place note for G-072-era rework), the pre-parsekey opts ordering loop, a disabled XXXliteral* switch arm in get_dict_can_assign_value's multi-member clause walk (open item noted: literalprefix() members unhandled there), and misc debug remnants. Expanded the leaders/values loop MAINTENANCE notes into a real refactor assessment: shared clause-assignment helper extraction is indicated (drift is real - G-082 duplicated its selection block, parsekey idents exist only values-side) but deferred to ride the G-084 leaders/values parsekey parity work under the G-046 hot-path constraints; until then edits to one loop must be mirrored or justified. Documentation corrections: get_dict PUNKARGS/doctools return shape now lists the real result keys (leaders opts values received solos multis id form + formstatus when candidacy ran); directive lists updated beyond @cmd/@leaders/@opts/@values; removed false pre-modern claims ('only supports -flag val pairs, not solo options', 'only the last value is used'); fixed the broken parse -help example (dict exists bracketing) and the module-header dofilestuff example (stray trailing 1 after -type existingfile); assorted typos. Goal cross-references annotated in code comments (G-053 -multiple occurrence ranges, G-084 values/leaders parsekey, corrected test-name refs to parsekey_repeat_ordering). Flagged (comment only, behaviour unchanged, no pins exist): suspected bug in the multi-member clause stringstartswith( arm - it string-matches against $tp (the type) instead of $rv (the value), so it always takes the success path. |
|
#0.12.1 - G-082 single-form parse error selection: a word rejected by an OPTIONAL argument's allocation screen (G-071 restricted-choice screen, or the basic-type screen: int/double/bool/number/dict) that subsequently overflows the argument list now reports the specific rejection - choiceviolation naming the word and the allowed choices (with -badarg/-badval and the same message template as final validation's report for required args), or typemismatch naming the declared type - instead of the generic 'Received more values than can be assigned' toomanyarguments (regression vs 0.5.0 found during G-030 make.tcl dogfooding; the -optional 0 authoring workaround is no longer required). Mechanism: get_dict_can_assign_value returns a 'rejection' record (single-member clauses only; literal-category mismatches deliberately not recorded - overflow stays the clearer report for syntax-word shapes like lseq 'to'), the three allocation call sites (leader/value split trial, leaders loop, values loop) collect first-rejection-per-word into a per-form store, and the two overflow raise sites prefer a recorded rejection for the unassignable word. Genuinely-surplus words (all optionals satisfied) still report toomanyarguments; words consumed by a later slot parse unchanged; multiform selection (rank_form_failures) unchanged - per-form failures just carry the more specific class. New testsuite errorselection.test (characterized pre-fix, pins flipped deliberately). |
|
#0.12.0 - G-074: new punk::args::formcheck - on-demand multiform ambiguity analysis. Static pairwise pass over a definition's resolved FORMS: enumerates the positional word-slot chains each form can present (leaders then values, branching on -optional arguments and ?-wrapped optional clause members, -multiple capped; options excluded - order-free), derives candidate witness arglists for aligned equal-length chains (discriminator words + type-derived witnesses, screened via choiceword_match/type tests so screening can't diverge from parse acceptance), and only reports a form pair after CONFIRMING a witness with a real single-form parse against BOTH forms (parse_status; no user-supplied words, nothing added to define/resolve cost) - so fully discriminated pairs cannot false-alarm and every reported witness is a genuine multipleformmatches arglist; the documented miss direction is witness derivation (exotic types, forms requiring options, enumeration caps). Findings classify as type_weakness (a discriminator slot - literal()/literalprefix()/restricted choices - aligned with a permissive non-validating type: any/none/string/ansistring/globstring/expr/script; the 'lseq 1 count 5' class) vs structural (forms genuinely share an argument shape; the 'after cancel id|script' class). New @form key -overlapallowed <formname-list> sanctions a KNOWN overlap on either pair member: the finding reports with sanctioned 1 and leaves the unsanctioned list (the actionable/gate subset) - parse behaviour is never affected (multipleformmatches still raises); unknown form names rejected at definition resolve. Returns machine-parsable dict (id/form_names/pairs/findings with witness+per-slot relations/unsanctioned/summary) or -return summary for the report text. New testsuite formcheck.test; @form directive doc updated. Proving consumers: tclcore ::lseq (reports range/start_count AND range/count - both real, root cause the expr-typed end slot) and ::after (exactly the sanctioned cancelid/cancelscript documented overlap). |
|
#0.11.2 - bad-@dynamic warn-once + round-1 caching (user-reported: 'i join' emitted the "bad @dynamic tag" warning 4x, 'i join test' 6x): a @dynamic definition whose round-1 tstr output contains no round-2 parameters previously bypassed argdefcache_unresolved entirely, so EVERY resolve redid display masking plus the full round-1 tstr and re-warned - each 'i' invocation resolves several times (doc walk, advisory parse, get_spec for the render, synopsis), all cache hits for static definitions but full re-work for bad-dynamic ones. Such definitions are now cached as a zero-param unresolved entry: subsequent resolves take the cheap cached branch (consistent with the round-1 freezing legitimately dynamic definitions already get) and the warning emits once per definition per interp. Warning message corrected while there: @dynamic is NOT a complete no-op for a parse-inert definition - deferred display fields still re-expand per render instead of caching (expand_display_fields spec_dynamic gate), so the message now says removal is appropriate only if that display behaviour is unintended. Legitimately dynamic definitions unaffected (round-2 re-substitution pinned live). Also commented the dev diagnostics dump (puts of the full records list + ::testrecord global) on resolve's malformed-record error path - the raised message already carries the offending record; found when the malformed ::tcl_startOfNextWord tclcore definition triggered it (fixed in tclcore 0.3.2). New testsuite args/dynamic.test (warn-once + cached-parse pin, legit-dynamic round-2 freshness pin, stderr captured via channel transform). |
|
#0.11.1 - documentation-only: define -help gains a 'Registration styles' section (direct define vs deferred lappend-PUNKARGS + register::NAMESPACES registration - lazy @id scan and on-demand definition via update_definitions, the module-template/moduledoc style, PUNKARGS_aliases, punk::args::status timings; prefer deferred for anything large) and an 'Interpolation (tstr placeholders) and the defspace' section (display-field deferral vs parse-field expansion at first resolve vs @dynamic re-expansion; defspace rule: direct define = calling namespace, registered PUNKARGS = the argdoc child namespace when it exists even if the PUNKARGS variable is in the parent; unresolvable placeholders left silently literal - check the defspace when a literal ${...} appears; robust patterns for load-time-computed values: explicit tstr pre-expansion, build-time string-map token, or setting the variable in argdoc). Motivated by the ::after id-shape harvest misattribution (2026-07-13, corrected record in goals/G-055). |
|
#0.11.0 - G-041 increment 2 (doc-surface support): synopsis renderer honours the documented @form -synopsis override - the stored override now replaces the auto-calculated synopsis line in punk::args::synopsis full and summary renders (arg_error's synopsis section honoured it already; the per-form dict gains a 'synopsis' key when overridden, FORMARGS unchanged for dict consumers) - forms.test GAP forms_form_synopsis_override_stored_not_rendered_GAP flipped to forms_form_synopsis_override_rendered. Candidate-form ranking extended to choice discriminators: form_literal_affinity now treats a required argument with a RESTRICTED choice set as a discriminator (matched via choiceword_match, the shared G-040 resolver) alongside literal()/literalprefix() types - the tclcore models express subcommand-ish literals as -choices (e.g after's cancel/idle/info), so 'after cancel' now ranks the cancel forms first in noformmatch errors and parse_status best-candidate selection instead of falling back to declaration order. |
|
#0.10.0 - G-041 multi-form candidacy: punk::args::parse/get_dict with the default -form * (or any multi-form selection) now attempts every permitted form instead of effectively parsing form 0 only - a clean match against exactly one form is auto-selected, no match raises a 'noformmatch' PUNKARGS VALIDATION error naming each candidate form's first-line failure (candidates ranked best-first: leading-literal affinity with the supplied words, then incomplete before invalid, then declaration order - private::form_selection/form_literal_affinity/rank_form_failures), and several clean matches raise 'multipleformmatches' naming the forms (no silent preference - deliberate, callers pass -form to disambiguate). -form now accepts a list of form names/indices as documented (get_dict, parse, parse_status, arg_error; supplied order preserved - arg_error renders the argument table for the first listed form and marks all listed forms' synopsis entries; parse's error render passes the ranked candidates). Parse results gain a 'form' key (the parsed form; keys appended after 'id' - positional consumers unaffected) and, when candidacy ran, a 'formstatus' key reporting every attempted form's outcome; parse_status 'form' is now the matched/best-candidate form (per-argument statuses and badarg marking built for it) and gains the documented 'formstatus' key (per-form status/failureclass/badarg/message with caller attribution) - the per-form compatibility surface the G-044 completion/hinting consumer needs. Engine body extracted verbatim to private::get_dict_form (single-form parse; argspecs resolved before selection - no caller-frame use). forms.test GAP pins flipped (auto-selection, -form list restriction) + new coverage (shared-prologue arity discrimination, noformmatch ranking/errorcode, multipleformmatches, formstatus in results and parse_status). |
|
#0.9.0 - G-071: allocation choice screen - get_dict_can_assign_value no longer blanket-accepts any word for an argument with a RESTRICTED choice set; where allocation has an alternative (the argument is -optional, or a further occurrence of a -multiple argument) the candidate word is screened with choiceword_match (the shared G-040 implementation: exact/alias/prefix/nocase semantics; -choicemultiple words screened per list member within min/max), and a non-matching word yields onward to later elements/clauses instead of being consumed and failing validation. Fixes the lseq-class in-form failures: an optional choice noise word between required values plus a trailing optional-member clause ('lseq 0 10 2', '0 10 by 2', '1 5 by 0' now parse; noise-word variants and prefix normalization unchanged; genuinely invalid arglists now get the plain excess-values rejection instead of blaming the unrelated optional). REQUIRED arguments are deliberately not screened - the word must fill them and validation's choiceviolation reporting stays informative. -choicerestricted 0 behaviour unchanged. allocation.test GAPs flipped; parse_status -form documented-order behaviour pinned (a prior mis-premise: -form was already supported before the withid/withdef tail per its synopsis). |
|
#0.8.3 - G-071: silenced four more unconditional debug puts on get_dict's allocation retreat paths ("get_dict cannot assign val:... (111)/(222)/(333)/(444)") - these fired on stderr during NORMAL successful parses whenever an optional element's candidate word was rejected and allocation retreated (e.g. every noise-word skip when parsing if/lseq-modelled commands), polluting interactive output. Same class as the 0.2.3/0.8.2 finds. The retreat logic they marked is the target of the G-071 allocator-correctness work; behaviour unchanged, output only. New allocation.test characterization suite (reduced lseq-range/if-shape fixtures): current mis-allocation and misblame pinned as GAP tests (flip with G-071), noise-word variants and if-shape guards pinned as regression guards, parse_status -form absence pinned as GAP. |
|
#0.8.2 - fixed stray debug output: an unconditional 'puts' in private::get_dict_can_assign_value ("checking tp '<type>' against value '<value>'") fired on multi-element clause type checks (e.g. parsing 'try ... trap ... on ...' style arglists against the tclcore moduledoc), polluting interactive output - now commented like its 0.2.3 companion. Found during G-041 prework probing (real-vs-model divergence sweep of if/switch/try/lseq). |
|
#0.8.1 - G-045 documentation-only: the container quoting rules pinned by defquoting.test promoted into the define -help documentation (-help key section): braced values fully literal (backslash sequences survive as typed) with only tstr placeholders and their backslash escape special; double-quoted values get Tcl backslash semantics at record parse (\n -> newline, doubled backslash collapses) while $ and [] stay literal with no variable/command substitution outside tstr placeholders; the backslash-escaped placeholder idiom renders a literal placeholder. |
|
#0.8.0 - G-045: new bare @normalize directive - opts a definition into indent normalization of BLOCK-FORM multi-line field values (first line whitespace-only, as authored by opening a braced literal with a newline): the structural first newline and a whitespace-only trailing line are dropped, the content lines' common leading whitespace is the block's base indent, the first content line is unindented fully and subsequent lines re-based to the standard 4-space continuation convention (deeper relative indents preserved; whitespace-only inner lines become empty). Head-form values are never altered - their base indent is ambiguous (uniform continuation indent may be the deliberate +2 relative convention over base 4) - which makes @normalize a no-op on conforming file-style definitions (idempotence pinned). Fields in a record's -unindentedfields are exempt. Implemented as a resolve pre-pass over split records (private::normalize_records / rebase_multiline_value); @normalize with options is an error. Intended for constructed/string-built definitions (no whole-block indent treatment otherwise); ::punk::helptopic::define_docs converts to it as the consumer proof (punk 0.2.5). New testsuite normalize.test; define doc documents the directive; rendering.test P4 notes stay pinned as the unopted default. |
|
#0.7.0 - G-045: record-continuation token -& - an unquoted trailing -& element on a definition record line continues the record on the next line. Implemented in private::split_definition_records: the token is dropped and the next line joins after a single space with its leading whitespace collapsed, exactly how the Tcl parser joins backslash-newline continuations before a braced definition reaches the splitter - so a -& record assembles byte-identical to its backslash-continued equivalent (proven by parse+render equality test). Motivation: constructed/string-built definitions cannot author backslash-newline ergonomically (the building code's own quoting consumes it); -& is plain text and survives any construction. Collision rules: the token must be a bare word preceded by whitespace (or the whole line), trailing whitespace after it is tolerated (more forgiving than raw backslash-newline); a braced/quoted -& is data ({-&} is the escape for a literal trailing value); -& mid-line, -& as a word suffix (abc-&), and -& on lines inside still-open braced/quoted values are all data. Backslash continuation authoring is unchanged (continuation is additive). New testsuite recordcontinuation.test; define doc documents the token alongside backslash continuation. |
|
#0.6.1 - G-045: @cmd honours -unindentedfields for -help - arg_error's display-time indent transform (undent " "+help, max 4) is now gated by "-help" membership in the @cmd line's -unindentedfields list (same gate argument -help already had), so left-margin-authored cmd help renders its first line flush with continuations in both the table and string renderers. Previously the option was accepted on @cmd but ignored (rendering.test rendering_unindentedfields_cmd_help_GAP - flipped to rendering_unindentedfields_cmd_help). No in-tree definitions set @cmd -unindentedfields, so existing rendering is unchanged. Note: @cmd -summary has no indent transform in any renderer, so -unindentedfields membership for -summary is accepted and vacuously honoured. define doc for -unindentedfields now states where the option is valid. |
|
#0.6.0 - synopsis display: small restricted choice sets now render as literal alternates - an argument whose choice pool (-choices plus -choicegroups members, deduplicated) has 1-3 members and -choicerestricted true (the default) displays those words unitalicised joined by | in synopses (e.g 'after cancel' shows literal cancel; a 3-choice option shows (left|centre|right)), matching the display style of literal()/literalprefix() type-alternatives. Larger or unrestricted (-choicerestricted 0) choice sets keep the italicised argname/<type> display, and an explicit -typesynopsis always takes precedence. New private helper punk::args::private::synopsis_choice_literals shared by both synopsis render paths (leaders/values via synopsis_form_arg_display, options inline in synopsis); applies only to single-element -type lists (multi-element clause display unchanged). define doc for -choices documents the rule. Tests: synopsis.test - new characterization coverage for literal/literalprefix/stringstartswith/stringendswith type-alternates, option alternate parenthesization, multi-element clause display, -typesynopsis (value element lists, option passthrough incl documenter ANSI), plus the new choice-literal rule (small sets in leader/option/value positions, choicegroups counting, >3 and unrestricted fallbacks, -typesynopsis precedence) |
|
#0.5.0 - G-049 parse-status data model: new punk::args::parse_status - runs a parse attempt (withid/withdef) and returns a documented status structure instead of raising on validation failure (overall ok/status valid|invalid|incomplete/scheme/message/errorcode-minus-argspecs/failureclass/badarg/id/form/receivednames + per-argument argstatus with class/status ok|bad|unparsed/received/positions/hasvalue/value-in-effect incl -default fill). arg_error: new -parsestatus option - both renderers (table and string) now derive goodarg/badarg row marking and choice value-in-effect highlighting from the structure (built internally from -badarg/-parsedargs when not supplied), replacing the transient goodargs/badarg locals; scheme colours resolve per-render into a local array (scheme renders no longer mutate the shared arg_error_CLR array - the -nocolour leak) and the DOCUMENTED -scheme choice value 'nocolour' (and 'nocolor') now takes effect instead of falling through to leftover colours. parse: new -caller option overriding the %caller% frame-walk substitution in validation failure messages (parse_status defaults it to the definition's @cmd -name). get_dict: missingrequiredvalue/missingrequiredleader allocation failures now carry -badarg (type-failed words get badarg marking, not just choice violations). Tests: parsestatus.test (new), usagemarking.test nocolour/leak GAP pins flipped + -parsestatus render parity |
|
#0.4.2 - G-046: display-field deferral - resolve no longer tstr-expands display-only content during argument resolution: ${...} in -help (@cmd/@examples/argument records) and @formdisplay -header/-body is masked with inert tokens (spec key DISPLAY_DEFERRED) and expanded on demand at display time (arg_error/eg/resolved_def/@default-copyfrom hooks; separate argdefcache_display cache for non-dynamic defs; @dynamic display content re-expands per render preserving provider refresh). First parse of heavily documented commands drops accordingly (punk::ansi::mark_columns ~4.3s -> ~12ms; tclcore ::lseq resolve ~184ms -> ~2ms) and -help content that calls punk::args-parsing commands (including against its own id) no longer stalls or loops - plus a display-time reentrancy guard (raw ${...} source substituted on nested expansion of the same id). Record splitter factored to private::split_definition_records. Also: @dynamic second-round multiline substitutions into deferred fields now get the 'line' paramindents alignment (rendering_atdynamic_multiline_help_insertion GAP flipped); prefix/alias choice normalization writeback no longer list-quotes single-element-clause values ({\Deleted} shape bug, choicegroups_imap_prefix_shape GAP flipped); -return string renderer aligns cmd-help continuations under the Description label and its Example line shows the example instead of the doc url. Tests: deferredhelp.test (new), rendering.test/choicegroups.test updated per G-046 acceptance |
|
#0.4.1 - fixed ensemble_subcommands_definition building the definition against unloaded argdocs: subhelp/doctype-punkargs choiceinfo entries and synopsis choicelabels were omitted for documented subcommands whose registered namespace (::punk::args::register::NAMESPACES) had not yet been lazily loaded - e.g first 'i ansistring' in a fresh shell rendered the autogenerated ensemble help without subcommand-help markers until something else loaded ::punk::ansi::ansistring. The generator now runs update_definitions for the namespaces its id_exists checks could resolve in before testing them (tests: punk/args ensembledef.test, punk/ns cmdhelp.test cmdhelp_ensemble_lazy_registered_argdocs) |
|
#0.4.0 - tidy-up (pre G-046): removed dead developer-test procs define2 and parseXXX and the unused ::punk::args::TEST definition; renamed the interactive Get_caller diagnostics test_get_dict/test_get_by_id/test_callers (+ test1 id) and lib::tstr_test_one with leading double underscores and excluded them from exports (punk::args::lib now exports {[a-z]*} instead of *). new punk::args::private namespace (namespace path to punk::args/lib/system) for internal helpers - _get_dict_can_assign_value, _split_type_expression, _check_clausecolumn, _synopsis_form_arg_display, _argerror_load_colours and lib::_parse_tstr_parts moved there with underscore prefixes dropped. documentation: PUNKARGS definitions added for every remaining exported command (errorstyle, undefine, undefine_deflist, raw_def, resolved_def_values, get_spec, is_dynamic, rawdef_id, rawdef_is_dynamic, id_exists, idaliases, set_idalias, unset_idalias, get_idalias, id_query, real_id, status, choiceword_match, get_dict; lib: string_is_dict, flatzip, zero_based_posns, choiceinfo_marks, indent, undent, undentleader, longestCommonPrefix, lunique; helpers and argdocbase B/N/I/NI); resolve's empty @cmd -help filled in; punk::args::argdocbase added to the registered definition namespaces |
|
#0.3.2 - Tcl 9.1 compatibility (TIP 746 removed expr behaviour from lseq operands): zero_based_posns now wraps its count-1 operand in expr |
|
#0.3.1 - fixed helpers::example -title path calling bare [a] (only resolvable where a global 'a' alias exists, e.g. punk shells) - now punk::ansi::a; found by the new examples.test running under plain tclsh |
|
#0.3.0 - new: -choicealiases {alias canonical ...} argument option (G-040) - aliases accepted exact under any -choiceprefix/-nocase setting, join the prefix-calculation pool when -choiceprefix is true, and normalize to their canonical choice in parse results; deny applies to the matched name (alias deny requires full alias; canonical reached via alias exempt from the canonical's own deny); usage display folds aliases into the canonical entry's label as an (alias:...) note and includes alias names in display prefix calculation; alias->existing-choice and no-collision validated at definition resolve. new public helper punk::args::choiceword_match - the single choice-word matching implementation extracted from get_dict, also consumed by punk::ns cmd_traverse for doc-lookup parity |
|
#0.2.3 - fixed "@opts -any 1" adhoc option passthrough: an option not in the definition crashed get_dict ("can't read argname") at the name-mapping stage instead of passing through - adhoc opts now fall back to their raw supplied name (argstate/arg_checks already carried them); commented out a debug puts stderr (">>>_get_dict_can_assign_value NOT alloc_ok...") that fired on every failed clause type assignment (e.g. any multiform parse miss) |
|
#0.2.2 - documentation-only: define doc clarifies -choicerestricted 0 passthrough (non-matching/ambiguous/denied/reserved words accepted as ordinary values), -choiceprefix result normalization to the canonical choice + minimal-prefix display highlighting, -choiceprefixdenylist behaviour for shorter prefixes in both restricted modes, and -choiceprefixreservelist phantom-entry idiom for per-choice minimum-prefix control
|
|
|