X-Git-Url: https://perl5.git.perl.org/perl5.git/blobdiff_plain/a7b39f85d7caac0822fdcd78400e131a95f11148..4097fff1bfc2a94d2104729c0f5416fba088130b:/hv.h diff --git a/hv.h b/hv.h index e6b9bb8..dcc6eb1 100644 --- a/hv.h +++ b/hv.h @@ -8,6 +8,26 @@ * */ +/* These control hash traversal randomization and the environment variable PERL_PERTURB_KEYS. + * Currently disabling this functionality will break a few tests, but should otherwise work fine. + * See perlrun for more details. */ + +#if defined(PERL_PERTURB_KEYS_DISABLED) +# define PL_HASH_RAND_BITS_ENABLED 0 +# define PERL_HASH_ITER_BUCKET(iter) ((iter)->xhv_riter) +#else +# define PERL_HASH_RANDOMIZE_KEYS 1 +# if defined(PERL_PERTURB_KEYS_RANDOM) +# define PL_HASH_RAND_BITS_ENABLED 1 +# elif defined(PERL_PERTURB_KEYS_DETERMINISTIC) +# define PL_HASH_RAND_BITS_ENABLED 2 +# else +# define USE_PERL_PERTURB_KEYS 1 +# define PL_HASH_RAND_BITS_ENABLED PL_hash_rand_bits_enabled +# endif +# define PERL_HASH_ITER_BUCKET(iter) (((iter)->xhv_riter) ^ ((iter)->xhv_rand)) +#endif + /* entry in hash value chain */ struct he { /* Keep hent_next first in this structure, because sv_free_arenas take @@ -23,12 +43,17 @@ struct he { /* hash key -- defined separately for use as shared pointer */ struct hek { - U32 hek_hash; /* hash of key */ - I32 hek_len; /* length of hash key */ - char hek_key[1]; /* variable-length hash key */ + U32 hek_hash; /* computed hash of key */ + I32 hek_len; /* length of the hash key */ + /* Be careful! Sometimes we store a pointer in the hek_key + * buffer, which means it must be 8 byte aligned or things + * dont work on aligned platforms like HPUX + * Also beware, the last byte of the hek_key buffer is a + * hidden flags byte about the key. */ + char hek_key[1]; /* variable-length hash key */ /* the hash-key is \0-terminated */ /* after the \0 there is a byte for flags, such as whether the key - is UTF-8 */ + is UTF-8 or WAS-UTF-8, or an SV */ }; struct shared_he { @@ -38,7 +63,7 @@ struct shared_he { /* Subject to change. Don't access this directly. - Use the funcs in mro.c + Use the funcs in mro_core.c */ struct mro_alg { @@ -61,6 +86,9 @@ struct mro_meta { U32 pkg_gen; /* Bumps when local methods/@ISA change */ const struct mro_alg *mro_which; /* which mro alg is in use? */ HV *isa; /* Everything this class @ISA */ + HV *super; /* SUPER method cache */ + CV *destroy; /* DESTROY method if destroy_gen non-zero */ + U32 destroy_gen; /* Generation number of DESTROY cache */ }; #define MRO_GET_PRIVATE_DATA(smeta, which) \ @@ -91,12 +119,17 @@ struct xpvhv_aux { */ I32 xhv_name_count; struct mro_meta *xhv_mro_meta; - HV * xhv_super; /* SUPER method cache */ +#ifdef PERL_HASH_RANDOMIZE_KEYS U32 xhv_rand; /* random value for hash traversal */ U32 xhv_last_rand; /* last random value for hash traversal, used to detect each() after insert for warnings */ +#endif + U32 xhv_aux_flags; /* assorted extra flags */ }; +#define HvAUXf_SCAN_STASH 0x1 /* stash is being scanned by gv_check */ +#define HvAUXf_NO_DEREF 0x2 /* @{}, %{} etc (and nomethod) not present */ + /* hash structure: */ /* This structure must match the beginning of struct xpvmg in sv.h. */ struct xpvhv { @@ -112,7 +145,7 @@ struct xpvhv { =for apidoc AmU||HEf_SVKEY This flag, used in the length slot of hash entries and magic structures, specifies the structure contains an C pointer where a C pointer -is to be expected. (For information only--not to be used). +is to be expected. (For information only--not to be used). =head1 Handy Values @@ -124,20 +157,20 @@ Null HV pointer. =head1 Hash Manipulation Functions =for apidoc Am|char*|HvNAME|HV* stash -Returns the package name of a stash, or NULL if C isn't a stash. -See C, C. +Returns the package name of a stash, or C if C isn't a stash. +See C>, C>. =for apidoc Am|STRLEN|HvNAMELEN|HV *stash Returns the length of the stash's name. =for apidoc Am|unsigned char|HvNAMEUTF8|HV *stash -Returns true if the name is in UTF8 encoding. +Returns true if the name is in UTF-8 encoding. =for apidoc Am|char*|HvENAME|HV* stash -Returns the effective name of a stash, or NULL if there is none. The +Returns the effective name of a stash, or NULL if there is none. The effective name represents a location in the symbol table where this stash -resides. It is updated automatically when packages are aliased or deleted. -A stash that is no longer in the symbol table has no effective name. This +resides. It is updated automatically when packages are aliased or deleted. +A stash that is no longer in the symbol table has no effective name. This name is preferable to C for use in MRO linearisations and isa caches. @@ -145,10 +178,10 @@ caches. Returns the length of the stash's effective name. =for apidoc Am|unsigned char|HvENAMEUTF8|HV *stash -Returns true if the effective name is in UTF8 encoding. +Returns true if the effective name is in UTF-8 encoding. =for apidoc Am|void*|HeKEY|HE* he -Returns the actual pointer stored in the key slot of the hash entry. The +Returns the actual pointer stored in the key slot of the hash entry. The pointer may be either C or C, depending on the value of C. Can be assigned to. The C or C macros are usually preferable for finding the value of a key. @@ -156,11 +189,12 @@ usually preferable for finding the value of a key. =for apidoc Am|STRLEN|HeKLEN|HE* he If this is negative, and amounts to C, it indicates the entry holds an C key. Otherwise, holds the actual length of the key. Can -be assigned to. The C macro is usually preferable for finding key +be assigned to. The C macro is usually preferable for finding key lengths. =for apidoc Am|SV*|HeVAL|HE* he -Returns the value slot (type C) stored in the hash entry. Can be assigned +Returns the value slot (type C) +stored in the hash entry. Can be assigned to. SV *foo= HeVAL(hv); @@ -178,14 +212,14 @@ not care about what the length of the key is, you may use the global variable C, though this is rather less efficient than using a local variable. Remember though, that hash keys in perl are free to contain embedded nulls, so using C or similar is not a good way to find -the length of hash keys. This is very similar to the C macro -described elsewhere in this document. See also C. +the length of hash keys. This is very similar to the C macro +described elsewhere in this document. See also C>. If you are using C to get values to pass to C to create a new SV, you should consider using C as it is more efficient. -=for apidoc Am|char*|HeUTF8|HE* he +=for apidoc Am|U32|HeUTF8|HE* he Returns whether the C value returned by C is encoded in UTF-8, doing any necessary dereferencing of possibly C keys. The value returned will be 0 or non-0, not necessarily 1 (or even a value with any low bits set), @@ -208,6 +242,20 @@ C. =cut */ +#define PERL_HASH_DEFAULT_HvMAX 7 + +/* During hsplit(), if HvMAX(hv)+1 (the new bucket count) is >= this value, + * we preallocate the HvAUX() struct. + * The assumption being that we are using so much space anyway we might + * as well allocate the extra bytes and speed up later keys() + * or each() operations. We don't do this to small hashes as we assume + * that a) it will be easy/fast to resize them to add the iterator, and b) that + * many of them will be objects which won't be traversed. Larger hashes however + * will take longer to extend, and the size of the aux struct is swamped by the + * overall length of the bucket array. + * */ +#define PERL_HV_ALLOC_AUX_SIZE (1 << 9) + /* these hash entry flags ride on hent_klen (for use only in magic/tied HVs) */ #define HEf_SVKEY -2 /* hent_key is an SV* */ @@ -215,7 +263,7 @@ C. # define Nullhv Null(HV*) #endif #define HvARRAY(hv) ((hv)->sv_u.svu_hash) -#define HvFILL(hv) Perl_hv_fill(aTHX_ (const HV *)(hv)) +#define HvFILL(hv) Perl_hv_fill(aTHX_ MUTABLE_HV(hv)) #define HvMAX(hv) ((XPVHV*) SvANY(hv))->xhv_max /* This quite intentionally does no flag checking first. That's your responsibility. */ @@ -226,6 +274,9 @@ C. #define HvEITER_set(hv,e) Perl_hv_eiter_set(aTHX_ MUTABLE_HV(hv), e) #define HvRITER_get(hv) (SvOOK(hv) ? HvAUX(hv)->xhv_riter : -1) #define HvEITER_get(hv) (SvOOK(hv) ? HvAUX(hv)->xhv_eiter : NULL) +#define HvRAND_get(hv) (SvOOK(hv) ? HvAUX(hv)->xhv_rand : 0) +#define HvLASTRAND_get(hv) (SvOOK(hv) ? HvAUX(hv)->xhv_last_rand : 0) + #define HvNAME(hv) HvNAME_get(hv) #define HvNAMELEN(hv) HvNAMELEN_get(hv) #define HvENAME(hv) HvENAME_get(hv) @@ -274,7 +325,7 @@ C. ((SvOOK(hv) && HvAUX(hv)->xhv_name_u.xhvnameu_name && HvAUX(hv)->xhv_name_count != -1) \ ? HEK_UTF8(HvENAME_HEK_NN(hv)) : 0) -/* the number of keys (including any placeholders) */ +/* the number of keys (including any placeholders) - NOT PART OF THE API */ #define XHvTOTALKEYS(xhv) ((xhv)->xhv_keys) /* @@ -338,7 +389,9 @@ C. ((HeKLEN(he) == HEf_SVKEY) ? \ HeKEY_sv(he) : \ newSVpvn_flags(HeKEY(he), \ - HeKLEN(he), SVs_TEMP)) : \ + HeKLEN(he), \ + SVs_TEMP | \ + ( HeKUTF8(he) ? SVf_UTF8 : 0 ))) : \ &PL_sv_undef) #define HeSVKEY_set(he,sv) ((HeKLEN(he) = HEf_SVKEY), (HeKEY_sv(he) = sv)) @@ -354,7 +407,8 @@ C. #define HVhek_UTF8 0x01 /* Key is utf8 encoded. */ #define HVhek_WASUTF8 0x02 /* Key is bytes here, but was supplied as utf8. */ #define HVhek_UNSHARED 0x08 /* This key isn't a shared hash key. */ -#define HVhek_FREEKEY 0x100 /* Internal flag to say key is malloc()ed. */ +/* the following flags are options for functions, they are not stored in heks */ +#define HVhek_FREEKEY 0x100 /* Internal flag to say key is Newx()ed. */ #define HVhek_PLACEHOLD 0x200 /* Internal flag to create placeholder. * (may change, but Storable is a core module) */ #define HVhek_KEYCANONICAL 0x400 /* Internal flag - key is in canonical form. @@ -407,8 +461,7 @@ C. (val), (hash))) #define hv_exists_ent(hv, keysv, hash) \ - (hv_common((hv), (keysv), NULL, 0, 0, HV_FETCH_ISEXISTS, 0, (hash)) \ - ? TRUE : FALSE) + cBOOL(hv_common((hv), (keysv), NULL, 0, 0, HV_FETCH_ISEXISTS, 0, (hash))) #define hv_fetch_ent(hv, keysv, lval, hash) \ ((HE *) hv_common((hv), (keysv), NULL, 0, 0, \ ((lval) ? HV_FETCH_LVALUE : 0), NULL, (hash))) @@ -426,9 +479,10 @@ C. (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), \ (val), (hash))) + + #define hv_exists(hv, key, klen) \ - (hv_common_key_len((hv), (key), (klen), HV_FETCH_ISEXISTS, NULL, 0) \ - ? TRUE : FALSE) + cBOOL(hv_common_key_len((hv), (key), (klen), HV_FETCH_ISEXISTS, NULL, 0)) #define hv_fetch(hv, key, klen, lval) \ ((SV**) hv_common_key_len((hv), (key), (klen), (lval) \ @@ -439,6 +493,40 @@ C. (MUTABLE_SV(hv_common_key_len((hv), (key), (klen), \ (flags) | HV_DELETE, NULL, 0))) +/* Provide 's' suffix subs for constant strings (and avoid needing to count + * chars). See STR_WITH_LEN in handy.h - because these are macros we cant use + * STR_WITH_LEN to do the work, we have to unroll it. */ +#define hv_existss(hv, key) \ + hv_exists((hv), ("" key ""), (sizeof(key)-1)) + +#define hv_fetchs(hv, key, lval) \ + hv_fetch((hv), ("" key ""), (sizeof(key)-1), (lval)) + +#define hv_deletes(hv, key, flags) \ + hv_delete((hv), ("" key ""), (sizeof(key)-1), (flags)) + +#define hv_name_sets(hv, name, flags) \ + hv_name_set((hv),("" name ""),(sizeof(name)-1), flags) + +#define hv_stores(hv, key, val) \ + hv_store((hv), ("" key ""), (sizeof(key)-1), (val), 0) + +#ifdef PERL_CORE +# define hv_storehek(hv, hek, val) \ + hv_common((hv), NULL, HEK_KEY(hek), HEK_LEN(hek), HEK_UTF8(hek), \ + HV_FETCH_ISSTORE|HV_FETCH_JUST_SV, (val), HEK_HASH(hek)) +# define hv_fetchhek(hv, hek, lval) \ + ((SV **) \ + hv_common((hv), NULL, HEK_KEY(hek), HEK_LEN(hek), HEK_UTF8(hek), \ + (lval) \ + ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \ + : HV_FETCH_JUST_SV, \ + NULL, HEK_HASH(hek))) +# define hv_deletehek(hv, hek, flags) \ + hv_common((hv), NULL, HEK_KEY(hek), HEK_LEN(hek), HEK_UTF8(hek), \ + (flags)|HV_DELETE, NULL, HEK_HASH(hek)) +#endif + /* This refcounted he structure is used for storing the hints used for lexical pragmas. Without threads, it's basically struct he + refcount. With threads, life gets more complex as the structure needs to be shared @@ -477,10 +565,10 @@ struct refcounted_he { }; /* -=for apidoc m|SV *|refcounted_he_fetch_pvs|const struct refcounted_he *chain|const char *key|U32 flags +=for apidoc m|SV *|refcounted_he_fetch_pvs|const struct refcounted_he *chain|"literal string" key|U32 flags -Like L, but takes a literal string instead of -a string/length pair, and no precomputed hash. +Like L, but takes a literal string +instead of a string/length pair, and no precomputed hash. =cut */ @@ -489,10 +577,10 @@ a string/length pair, and no precomputed hash. Perl_refcounted_he_fetch_pvn(aTHX_ chain, STR_WITH_LEN(key), 0, flags) /* -=for apidoc m|struct refcounted_he *|refcounted_he_new_pvs|struct refcounted_he *parent|const char *key|SV *value|U32 flags +=for apidoc m|struct refcounted_he *|refcounted_he_new_pvs|struct refcounted_he *parent|"literal string" key|SV *value|U32 flags -Like L, but takes a literal string instead of -a string/length pair, and no precomputed hash. +Like L, but takes a literal string +instead of a string/length pair, and no precomputed hash. =cut */ @@ -565,11 +653,5 @@ Creates a new HV. The reference count is set to 1. #include "hv_func.h" /* - * Local variables: - * c-indentation-style: bsd - * c-basic-offset: 4 - * indent-tabs-mode: nil - * End: - * * ex: set ts=8 sts=4 sw=4 et: */