// file : libbuild2/target.ixx -*- C++ -*- // license : MIT; see accompanying LICENSE file #include // memcpy() #include namespace build2 { LIBBUILD2_SYMEXPORT timestamp mtime (const char*); // filesystem.cxx // target_key // inline const string& target_key:: effective_name (string& r, bool force_ext) const { const target_type& tt (*type); // Note that if the name is not empty, then we always use that, even // if the type is dir/fsdir. // if (name->empty () && (tt.is_a () || tt.is_a ())) { r = dir->leaf ().string (); } // If we have the extension and the type expects the extension to be // always specified explicitly by the user, then add it to the name. // // Overall, we have the following cases: // // 1. Extension is fixed: man1{}. // // 2. Extension is always specified by the user: file{}. // // 3. Default extension that may be overridden by the user: hxx{}. // // 4. Extension assigned by the rule but may be overridden by the // user: obje{}. // // By default we only include the extension for (2). // else if (ext && !ext->empty () && (force_ext || tt.fixed_extension == &target_extension_none || tt.fixed_extension == &target_extension_must)) { r = *name + '.' + *ext; } else return *name; // Use name as is. return r; } // rule_hints // inline const string& rule_hints:: find (const target_type& tt, operation_id o, bool ut) const { // Look for fallback during the same iteration. // const value_type* f (nullptr); for (const value_type& v: map) { if (!(v.type == nullptr ? ut : tt.is_a (*v.type))) continue; if (v.operation == o) return v.hint; if (f == nullptr && v.operation == default_id && (o == update_id || o == clean_id)) f = &v; } return f != nullptr ? f->hint : empty_string; } inline void rule_hints:: insert (const target_type* tt, operation_id o, string h) { auto i (find_if (map.begin (), map.end (), [tt, o] (const value_type& v) { return v.operation == o && v.type == tt; })); if (i == map.end ()) map.push_back (value_type {tt, o, move (h)}); else i->hint = move (h); } inline const string& target:: find_hint (operation_id o) const { using flag = target_type::flag; const target_type& tt (type ()); // First check the target itself. // if (!rule_hints.empty ()) { // If this is a group that "gave" its untyped hints to the members, then // ignore untyped entries. // bool ut ((tt.flags & flag::member_hint) != flag::member_hint); const string& r (rule_hints.find (tt, o, ut)); if (!r.empty ()) return r; } // Then check the group. // if (const target* g = group) { if (!g->rule_hints.empty ()) { // If the group "gave" its untyped hints to the members, then don't // ignore untyped entries. // bool ut ((g->type ().flags & flag::member_hint) == flag::member_hint); return g->rule_hints.find (tt, o, ut); } } return empty_string; } // match_extra // inline void match_extra:: reinit (bool f) { clear_data (); fallback = f; cur_options = all_options; new_options = 0; posthoc_prerequisite_targets = nullptr; } inline void match_extra:: free () { clear_data (); } // target // inline const string* target:: ext_locked () const { return *ext_ ? &**ext_ : nullptr; } inline const string* target:: ext () const { slock l (ctx.targets.mutex_); return ext_locked (); } inline target_key target:: key () const { const string* e (ext ()); return target_key { &type (), &dir, &out, &name, e != nullptr ? optional (*e) : nullopt}; } inline target_key target:: key_locked () const { const string* e (ext_locked ()); return target_key { &type (), &dir, &out, &name, e != nullptr ? optional (*e) : nullopt}; } inline names target:: as_name () const { return key ().as_name (); } inline void target:: as_name (names& r) const { return key ().as_name (r); } inline auto target:: prerequisites () const -> const prerequisites_type& { return prerequisites_state_.load (memory_order_acquire) == 2 ? prerequisites_ : empty_prerequisites; } inline bool target:: prerequisites (prerequisites_type&& p) const { target& x (const_cast (*this)); // MT-aware. uint8_t e (0); if (x.prerequisites_state_.compare_exchange_strong ( e, 1, memory_order_acq_rel, memory_order_acquire)) { x.prerequisites_ = move (p); x.prerequisites_state_.fetch_add (1, memory_order_release); return true; } else { // Spin the transition out so that prerequisites() doesn't return empty. // for (; e == 1; e = prerequisites_state_.load (memory_order_acquire)) /*this_thread::yield ()*/ ; return false; } } inline bool target:: matched (action a, memory_order mo) const { assert (ctx.phase == run_phase::match || ctx.phase == run_phase::execute); const opstate& s (state[a]); size_t c (s.task_count.load (mo)); size_t b (ctx.count_base ()); // Note: cannot do (c - b)! if (ctx.phase == run_phase::match) { // While it will normally be applied, it could also be already executed // or being relocked to reapply match options (see lock_impl() for // background). // // Note that we can't just do >= offset_applied since offset_busy can // also mean it is being matched. // // See also matched_state_impl(), mtime() for similar logic. // return (c == (b + offset_applied) || c == (b + offset_executed) || (c >= (b + offset_busy) && s.match_extra.cur_options_.load (memory_order_relaxed) != 0)); } else { // Note that while the target could be being executed, we should see at // least offset_matched since it must have been "achieved" before the // phase switch. // return c >= (b + offset_matched); } } inline bool target:: group_state (action a) const { // We go an extra step and short-circuit to the target state even if the // raw state is not group provided the recipe is group_recipe and the // state is unknown (see mtime() for a discussion on why we do it). // // Note that additionally s.state may not be target_state::group even // after execution due to deferment (see execute_impl() for details). // // @@ Hm, I wonder why not just return s.recipe_group_action now that we // cache it. // const opstate& s (state[a]); // This special hack allows us to do things like query an ad hoc member's // state or mtime without matching/executing the member, only the group. // Requiring matching/executing the member would be too burdensome and // this feels harmless (ad hoc membership cannot be changed during the // execute phase). // // Note: this test must come first since the member may not be matched and // thus its state set (but it won't be postponed; see opstate::state). // if (ctx.phase == run_phase::execute && adhoc_group_member ()) { // Note: if the member state is postponed, then the group state may not // be yet known (see group_action() for details). // return s.state != target_state::postponed; } if (s.state == target_state::group) return true; if (s.state == target_state::unknown && group != nullptr) return s.recipe_group_action; return false; } inline pair target:: matched_state_impl (action a) const { // Note that the "tried" state is "final". // const opstate& s (state[a]); // Note: already synchronized. // size_t c (s.task_count.load (memory_order_relaxed)); size_t b (ctx.count_base ()); // Note: cannot do (c - b)! if (c == (b + offset_tried)) return make_pair (false, target_state::unknown); else { // The same semantics as in target::matched(). Note that in the executed // case we are guaranteed to be synchronized since we are in the match // phase. // assert (c == (b + offset_applied) || c == (b + offset_executed) || (c >= (b + offset_busy) && s.match_extra.cur_options_.load (memory_order_relaxed) != 0)); return make_pair (true, (group_state (a) ? group->state[a] : s).state); } } inline target_state target:: executed_state_impl (action a) const { target_state ts ((group_state (a) ? group->state : state)[a].state); // Translate postponed to unchanged, similar to execute_recipe(). // if (ts == target_state::postponed) ts = target_state::unchanged; return ts; } inline target_state target:: matched_state (action a, bool fail) const { assert (ctx.phase == run_phase::match); // Note that the target could be being asynchronously re-matched. // pair r (matched_state_impl (a)); if (fail && (!r.first || r.second == target_state::failed)) throw failed (); return r.second; } inline pair target:: try_matched_state (action a, bool fail) const { assert (ctx.phase == run_phase::match); pair r (matched_state_impl (a)); if (fail && r.first && r.second == target_state::failed) throw failed (); return r; } inline target_state target:: executed_state (action a, bool fail) const { assert (ctx.phase == run_phase::execute || ctx.phase == run_phase::load); target_state r (executed_state_impl (a)); if (fail && r == target_state::failed) throw failed (); return r; } inline bool target:: has_prerequisites () const { return !prerequisites ().empty (); } inline bool target:: has_group_prerequisites () const { return has_prerequisites () || (group != nullptr && group->has_prerequisites ()); } inline bool target:: unchanged (action a) const { assert (ctx.phase == run_phase::match); return matched_state_impl (a).second == target_state::unchanged; } inline ostream& operator<< (ostream& os, const target& t) { return os << t.key (); } // mark()/unmark() // // VC15 doesn't like if we use (abstract) target here. // static_assert (alignof (file) % 4 == 0, "unexpected target alignment"); inline void mark (const target*& p, uint8_t m) { uintptr_t i (reinterpret_cast (p)); i |= m & 0x03; p = reinterpret_cast (i); } inline uint8_t marked (const target* p) { uintptr_t i (reinterpret_cast (p)); return uint8_t (i & 0x03); } inline uint8_t unmark (const target*& p) { uintptr_t i (reinterpret_cast (p)); uint8_t m (i & 0x03); if (m != 0) { i &= ~uintptr_t (0x03); p = reinterpret_cast (i); } return m; } // include() // LIBBUILD2_SYMEXPORT include_type include_impl (action, const target&, const prerequisite&, const target*, lookup*); inline include_type include (action a, const target& t, const prerequisite& p, lookup* l) { // Most of the time no prerequisite-specific variables will be specified, // so let's optimize for that. // return p.vars.empty () ? include_type (true) : include_impl (a, t, p, nullptr, l); } inline include_type include (action a, const target& t, const prerequisite_member& pm, lookup* l) { return pm.prerequisite.vars.empty () ? include_type (true) : include_impl (a, t, pm.prerequisite, pm.member, l); } // group_prerequisites // inline group_prerequisites:: group_prerequisites (const target& t) : t_ (nullptr), g_ (nullptr) { // Take "snapshot" of prerequisites, both for target and group. // const auto& ps (t.prerequisites ()); if (!ps.empty ()) t_ = &ps; if (const target* g = t.group) { if (g->adhoc_member == nullptr) // Not ad hoc group member. { const auto& ps (g->prerequisites ()); if (!ps.empty ()) g_ = &ps; } } } inline group_prerequisites:: group_prerequisites (const target& t, const target* g) : t_ (nullptr), g_ (nullptr) { const auto& ps (t.prerequisites ()); if (!ps.empty ()) t_ = &ps; if (g != nullptr) { const auto& ps (g->prerequisites ()); if (!ps.empty ()) g_ = &ps; } } inline auto group_prerequisites:: begin () const -> iterator { auto* c (g_ != nullptr ? g_ : t_ != nullptr ? t_ : &empty_prerequisites); return iterator (t_, g_, c, c->begin ()); } inline auto group_prerequisites:: end () const -> iterator { auto* c (t_ != nullptr ? t_ : g_ != nullptr ? g_ : &empty_prerequisites); return iterator (t_, g_, c, c->end ()); } inline size_t group_prerequisites:: size () const { return ((t_ != nullptr ? t_->size () : 0) + (g_ != nullptr ? g_->size () : 0)); } // group_prerequisites::iterator // inline auto group_prerequisites::iterator:: operator++ () -> iterator& { if (++i_ == c_->end () && c_ == g_ && t_ != nullptr) { c_ = t_; i_ = c_->begin (); } return *this; } inline auto group_prerequisites::iterator:: operator-- () -> iterator& { if (i_ == c_->begin () && c_ == t_) { c_ = g_; i_ = c_->end (); } --i_; return *this; } // prerequisite_member // inline prerequisite prerequisite_member:: as_prerequisite () const { if (member == nullptr) return prerequisite; // An ad hoc group member cannot be used as a prerequisite (use the whole // group instead). // assert (!member->adhoc_group_member ()); // Feels like copying the prerequisite's variables to member is more // correct than not (consider for_install, for example). // prerequisite_type p (*member); p.vars = prerequisite.vars; return p; } inline prerequisite_key prerequisite_member:: key () const { return member != nullptr ? prerequisite_key {prerequisite.proj, member->key (), nullptr} : prerequisite.key (); } // prerequisite_members // LIBBUILD2_SYMEXPORT group_view resolve_members (action, const target&); // template inline group_view prerequisite_members_range::iterator:: resolve_members (const prerequisite& p) { // We want to allow iteration over members during execute provided the // same iteration has been performed during match. // const target* pt (r_->t_.ctx.phase == run_phase::match ? &search (r_->t_, p) : search_existing (p)); assert (pt != nullptr); return build2::resolve_members (r_->a_, *pt); } template inline void prerequisite_members_range::iterator:: switch_mode () { g_ = resolve_members (*i_); if (g_.members != nullptr) { // See empty see through groups as groups. // for (j_ = 1; j_ <= g_.count && g_.members[j_ - 1] == nullptr; ++j_) ; if (j_ > g_.count) g_.count = 0; } else assert (r_->mode_ != members_mode::always); // Group can't be resolved. } template inline auto prerequisite_members_range::iterator:: operator++ () -> iterator& { if (k_ != nullptr) // Iterating over an ad hoc group. k_ = k_->adhoc_member; if (k_ == nullptr && g_.count != 0) // Iterating over a normal group. { if (g_.members == nullptr) // Special case, see leave_group(). g_.count = 0; else { for (++j_; j_ <= g_.count && g_.members[j_ - 1] == nullptr; ++j_) ; if (j_ > g_.count) g_.count = 0; } } if (k_ == nullptr && g_.count == 0) // Iterating over the range. { ++i_; if (r_->mode_ != members_mode::never && i_ != r_->e_ && i_->type.see_through ()) switch_mode (); } return *this; } template inline bool prerequisite_members_range::iterator:: enter_group () { assert (k_ == nullptr); // No nested ad hoc group entering. // First see if we are about to enter an ad hoc group. // const target* t (g_.count != 0 ? j_ != 0 ? g_.members[j_ - 1] : nullptr : i_->target.load (memory_order_consume)); if (t != nullptr && t->adhoc_member != nullptr) k_ = t; // Increment that follows will make it t->member. else { // Otherwise assume it is a normal group. // g_ = resolve_members (*i_); if (g_.members == nullptr) // Members are not know. { g_.count = 0; return false; } // Note: 0-based to account for the increment that will follow. // for (j_ = 0; j_ != g_.count && g_.members[j_] == nullptr; ++j_) ; if (j_ == g_.count) g_.count = 0; } return true; } template inline void prerequisite_members_range::iterator:: leave_group () { if (k_ != nullptr) { // Skip until the last element (next increment will reach the end). // for (; k_->adhoc_member != nullptr; k_ = k_->adhoc_member) ; } else { // Pretend we are on the last member of a normal group. // j_ = 0; g_.count = 1; g_.members = nullptr; // Ugly "special case signal" for operator++. } } template inline bool prerequisite_members_range::iterator:: group () const { // Ad hoc. // if (k_ != nullptr) return k_->adhoc_member; // Explicit. // if (g_.count != 0 && g_.members != nullptr) { size_t j (j_ + 1); for (; j <= g_.count && g_.members[j - 1] == nullptr; ++j) ; return j <= g_.count; } return false; } inline auto prerequisite_members (action a, const target& t, members_mode m) { return prerequisite_members (a, t, t.prerequisites (), m); } inline auto reverse_prerequisite_members (action a, const target& t, members_mode m) { return prerequisite_members (a, t, reverse_iterate (t.prerequisites ()), m); } // mtime_target // inline void mtime_target:: mtime (timestamp mt) const { mtime_.store (mt.time_since_epoch ().count (), memory_order_release); } inline timestamp mtime_target:: load_mtime (const path& p) const { // We can only enforce "not group state" during the execute phase. During // match (e.g., the target is being matched), we will just have to pay // attention. // assert (ctx.phase == run_phase::match || (ctx.phase == run_phase::execute && !group_state (action () /* inner */))); duration::rep r (mtime_.load (memory_order_consume)); if (r == timestamp_unknown_rep) { assert (!p.empty ()); r = build2::mtime (p.string ().c_str ()).time_since_epoch ().count (); mtime_.store (r, memory_order_release); } return timestamp (duration (r)); } inline bool mtime_target:: newer (timestamp mt, target_state s) const { assert (s != target_state::unknown); // Should be executed. timestamp mp (mtime ()); // What do we do if timestamps are equal? This can happen, for example, // on filesystems that don't have subsecond resolution. There is not // much we can do here except detect the case where the target was // changed on this run. // return mt < mp || (mt == mp && s == target_state::changed); } inline bool mtime_target:: newer (timestamp mt) const { assert (ctx.phase == run_phase::execute); return newer (mt, executed_state_impl (action () /* inner */)); } // path_target // inline const path& path_target:: path (memory_order mo) const { // You may be wondering why don't we spin the transition out? The reason // is it shouldn't matter since were we called just a moment earlier, we // wouldn't have seen it. // return path_state_.load (mo) == 2 ? path_ : empty_path; } inline const path& path_target:: path (path_type p) const { uint8_t e (0); if (path_state_.compare_exchange_strong ( e, 1, memory_order_acq_rel, memory_order_acquire)) { path_ = move (p); path_state_.fetch_add (1, memory_order_release); } else { // Spin the transition out. // for (; e == 1; e = path_state_.load (memory_order_acquire)) /*this_thread::yield ()*/ ; assert (e == 2 && path_ == p); } return path_; } inline timestamp path_target:: load_mtime () const { return mtime_target::load_mtime (path ()); } inline const path& path_target:: path_mtime (path_type p, timestamp mt) const { // Because we use the presence of mtime to indicate the special "trust me, // this file exists" situation, the order in which we do things is // important. In particular, the fallback file_rule::match() will skip // assigning the path if there is a valid timestamp. As a result, with the // wrong order we may end up in a situation where the rule is matched but // the path is not assigned. // const path_type& r (path (move (p))); mtime (mt); return r; } // exe // inline auto exe:: process_path () const -> process_path_type { // It's unfortunate we have to return by value but hopefully the // compiler will see through it. Note also that returning empty // process path if path is empty. // return process_path_.empty () ? process_path_type (path ().string ().c_str (), path_type (), path_type ()) : process_path_type (process_path_, false /* init */); } inline void exe:: process_path (process_path_type p) { process_path_ = move (p); } }