This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Given that we now have a union, there's no need for all the HvARRAY()
[perl5.git] / hv.h
CommitLineData
a0d0e21e 1/* hv.h
79072805 2 *
4bb101f2 3 * Copyright (C) 1991, 1992, 1993, 1996, 1997, 1998, 1999,
1d325971 4 * 2000, 2001, 2002, 2005, by Larry Wall and others
79072805
LW
5 *
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
8 *
79072805
LW
9 */
10
5cbe4eec 11/* entry in hash value chain */
79072805 12struct he {
6f51351d
NC
13 /* Keep hent_next first in this structure, because sv_free_arenas take
14 advantage of this to share code between the he arenas and the SV
15 body arenas */
5cbe4eec
MF
16 HE *hent_next; /* next entry in chain */
17 HEK *hent_hek; /* hash key */
18 SV *hent_val; /* scalar value that was hashed */
bbce6d69
PP
19};
20
5cbe4eec 21/* hash key -- defined separately for use as shared pointer */
ff68c719 22struct hek {
5cbe4eec
MF
23 U32 hek_hash; /* hash of key */
24 I32 hek_len; /* length of hash key */
25 char hek_key[1]; /* variable-length hash key */
e05949c7 26 /* the hash-key is \0-terminated */
1e54db1a
JH
27 /* after the \0 there is a byte for flags, such as whether the key
28 is UTF-8 */
79072805
LW
29};
30
cbae3960
NC
31struct shared_he {
32 struct he shared_he_he;
33 struct hek shared_he_hek;
34};
bfcb3514
NC
35
36/* Subject to change.
37 Don't access this directly.
38*/
39struct xpvhv_aux {
7423f6db 40 HEK *xhv_name; /* name, if a symbol table */
bfcb3514
NC
41 HE *xhv_eiter; /* current entry of iterator */
42 I32 xhv_riter; /* current root of iterator */
43};
44
5cbe4eec 45/* hash structure: */
6ee623d5 46/* This structure must match the beginning of struct xpvmg in sv.h. */
79072805 47struct xpvhv {
311a25d9 48 NV xnv_nv; /* numeric value, if any */
463ee0b2
LW
49 STRLEN xhv_fill; /* how full xhv_array currently is */
50 STRLEN xhv_max; /* subscript of last element of xhv_array */
e4305a63 51 union {
311a25d9
NC
52 IV xivu_iv; /* integer value or pv offset */
53 UV xivu_uv;
54 void * xivu_p1;
55 } xiv_u;
79072805
LW
56 MAGIC* xmg_magic; /* magic for scalar array */
57 HV* xmg_stash; /* class package */
79072805
LW
58};
59
311a25d9 60#define xhv_keys xiv_u.xivu_iv
e4305a63 61
c5f87f51 62#if 0
59813432 63typedef struct xpvhv xpvhv_allocated;
9f1501b2
NC
64#else
65typedef struct {
66 STRLEN xhv_fill; /* how full xhv_array currently is */
67 STRLEN xhv_max; /* subscript of last element of xhv_array */
68 union {
311a25d9
NC
69 IV xivu_iv; /* integer value or pv offset */
70 UV xivu_uv;
71 void * xivu_p1;
72 } xiv_u;
9f1501b2
NC
73 MAGIC* xmg_magic; /* magic for scalar array */
74 HV* xmg_stash; /* class package */
75} xpvhv_allocated;
76#endif
e4305a63 77
5cbe4eec 78/* hash a key */
5afd6d42
JH
79/* FYI: This is the "One-at-a-Time" algorithm by Bob Jenkins
80 * from requirements by Colin Plumb.
81 * (http://burtleburtle.net/bob/hash/doobs.html) */
82/* The use of a temporary pointer and the casting games
83 * is needed to serve the dual purposes of
84 * (a) the hashed data being interpreted as "unsigned char" (new since 5.8,
85 * a "char" can be either signed or signed, depending on the compiler)
86 * (b) catering for old code that uses a "char"
830b38bd 87 *
504f80c1
JH
88 * The "hash seed" feature was added in Perl 5.8.1 to perturb the results
89 * to avoid "algorithmic complexity attacks".
830b38bd
JH
90 *
91 * If USE_HASH_SEED is defined, hash randomisation is done by default
92 * If USE_HASH_SEED_EXPLICIT is defined, hash randomisation is done
93 * only if the environment variable PERL_HASH_SEED is set.
94 * For maximal control, one can define PERL_HASH_SEED.
90ce55a1 95 * (see also perl.c:perl_parse()).
5afd6d42 96 */
830b38bd
JH
97#ifndef PERL_HASH_SEED
98# if defined(USE_HASH_SEED) || defined(USE_HASH_SEED_EXPLICIT)
99# define PERL_HASH_SEED PL_hash_seed
100# else
101# define PERL_HASH_SEED 0
102# endif
504f80c1 103#endif
bf6bd887
PP
104#define PERL_HASH(hash,str,len) \
105 STMT_START { \
5afd6d42
JH
106 register const char *s_PeRlHaSh_tmp = str; \
107 register const unsigned char *s_PeRlHaSh = (const unsigned char *)s_PeRlHaSh_tmp; \
bf6bd887 108 register I32 i_PeRlHaSh = len; \
504f80c1 109 register U32 hash_PeRlHaSh = PERL_HASH_SEED; \
a6fe520e
MF
110 while (i_PeRlHaSh--) { \
111 hash_PeRlHaSh += *s_PeRlHaSh++; \
112 hash_PeRlHaSh += (hash_PeRlHaSh << 10); \
113 hash_PeRlHaSh ^= (hash_PeRlHaSh >> 6); \
114 } \
115 hash_PeRlHaSh += (hash_PeRlHaSh << 3); \
116 hash_PeRlHaSh ^= (hash_PeRlHaSh >> 11); \
4fee5c71 117 (hash) = (hash_PeRlHaSh + (hash_PeRlHaSh << 15)); \
bf6bd887
PP
118 } STMT_END
119
3d78eb94
NC
120/* Only hv.c and mod_perl should be doing this. */
121#ifdef PERL_HASH_INTERNAL_ACCESS
4b5190b5
NC
122#define PERL_HASH_INTERNAL(hash,str,len) \
123 STMT_START { \
124 register const char *s_PeRlHaSh_tmp = str; \
125 register const unsigned char *s_PeRlHaSh = (const unsigned char *)s_PeRlHaSh_tmp; \
126 register I32 i_PeRlHaSh = len; \
008fb0c0 127 register U32 hash_PeRlHaSh = PL_rehash_seed; \
4b5190b5
NC
128 while (i_PeRlHaSh--) { \
129 hash_PeRlHaSh += *s_PeRlHaSh++; \
130 hash_PeRlHaSh += (hash_PeRlHaSh << 10); \
131 hash_PeRlHaSh ^= (hash_PeRlHaSh >> 6); \
132 } \
133 hash_PeRlHaSh += (hash_PeRlHaSh << 3); \
134 hash_PeRlHaSh ^= (hash_PeRlHaSh >> 11); \
135 (hash) = (hash_PeRlHaSh + (hash_PeRlHaSh << 15)); \
136 } STMT_END
137#endif
138
954c1994 139/*
ccfc67b7
JH
140=head1 Hash Manipulation Functions
141
954c1994
GS
142=for apidoc AmU||HEf_SVKEY
143This flag, used in the length slot of hash entries and magic structures,
d1be9408 144specifies the structure contains an C<SV*> pointer where a C<char*> pointer
954c1994
GS
145is to be expected. (For information only--not to be used).
146
ccfc67b7
JH
147=head1 Handy Values
148
954c1994
GS
149=for apidoc AmU||Nullhv
150Null HV pointer.
151
ccfc67b7
JH
152=head1 Hash Manipulation Functions
153
954c1994 154=for apidoc Am|char*|HvNAME|HV* stash
9282b5fd
SH
155Returns the package name of a stash, or NULL if C<stash> isn't a stash.
156See C<SvSTASH>, C<CvSTASH>.
954c1994
GS
157
158=for apidoc Am|void*|HeKEY|HE* he
159Returns the actual pointer stored in the key slot of the hash entry. The
160pointer may be either C<char*> or C<SV*>, depending on the value of
161C<HeKLEN()>. Can be assigned to. The C<HePV()> or C<HeSVKEY()> macros are
162usually preferable for finding the value of a key.
163
164=for apidoc Am|STRLEN|HeKLEN|HE* he
165If this is negative, and amounts to C<HEf_SVKEY>, it indicates the entry
166holds an C<SV*> key. Otherwise, holds the actual length of the key. Can
167be assigned to. The C<HePV()> macro is usually preferable for finding key
168lengths.
169
170=for apidoc Am|SV*|HeVAL|HE* he
171Returns the value slot (type C<SV*>) stored in the hash entry.
172
173=for apidoc Am|U32|HeHASH|HE* he
174Returns the computed hash stored in the hash entry.
175
176=for apidoc Am|char*|HePV|HE* he|STRLEN len
177Returns the key slot of the hash entry as a C<char*> value, doing any
178necessary dereferencing of possibly C<SV*> keys. The length of the string
179is placed in C<len> (this is a macro, so do I<not> use C<&len>). If you do
180not care about what the length of the key is, you may use the global
181variable C<PL_na>, though this is rather less efficient than using a local
182variable. Remember though, that hash keys in perl are free to contain
183embedded nulls, so using C<strlen()> or similar is not a good way to find
184the length of hash keys. This is very similar to the C<SvPV()> macro
185described elsewhere in this document.
186
187=for apidoc Am|SV*|HeSVKEY|HE* he
188Returns the key as an C<SV*>, or C<Nullsv> if the hash entry does not
189contain an C<SV*> key.
190
191=for apidoc Am|SV*|HeSVKEY_force|HE* he
192Returns the key as an C<SV*>. Will create and return a temporary mortal
193C<SV*> if the hash entry contains only a C<char*> key.
194
195=for apidoc Am|SV*|HeSVKEY_set|HE* he|SV* sv
196Sets the key to a given C<SV*>, taking care to set the appropriate flags to
197indicate the presence of an C<SV*> key, and returns the same
198C<SV*>.
199
200=cut
201*/
bf6bd887 202
bf5b86ae 203/* these hash entry flags ride on hent_klen (for use only in magic/tied HVs) */
d1be9408 204#define HEf_SVKEY -2 /* hent_key is an SV* */
bf6bd887
PP
205
206
79072805 207#define Nullhv Null(HV*)
43e6e717 208#define HvARRAY(hv) ((hv)->sv_u.svu_hash)
79072805 209#define HvFILL(hv) ((XPVHV*) SvANY(hv))->xhv_fill
463ee0b2 210#define HvMAX(hv) ((XPVHV*) SvANY(hv))->xhv_max
b79f7545
NC
211/* This quite intentionally does no flag checking first. That's your
212 responsibility. */
213#define HvAUX(hv) ((struct xpvhv_aux*)&(HvARRAY(hv)[HvMAX(hv)+1]))
bfcb3514
NC
214#define HvRITER(hv) (*Perl_hv_riter_p(aTHX_ (HV*)(hv)))
215#define HvEITER(hv) (*Perl_hv_eiter_p(aTHX_ (HV*)(hv)))
216#define HvRITER_set(hv,r) Perl_hv_riter_set(aTHX_ (HV*)(hv), r)
217#define HvEITER_set(hv,e) Perl_hv_eiter_set(aTHX_ (HV*)(hv), e)
b79f7545
NC
218#define HvRITER_get(hv) (SvOOK(hv) ? HvAUX(hv)->xhv_riter : -1)
219#define HvEITER_get(hv) (SvOOK(hv) ? HvAUX(hv)->xhv_eiter : 0)
7423f6db 220#define HvNAME(hv) HvNAME_get(hv)
bfcb3514
NC
221/* FIXME - all of these should use a UTF8 aware API, which should also involve
222 getting the length. */
9b9d0b15 223/* This macro may go away without notice. */
b79f7545
NC
224#define HvNAME_HEK(hv) (SvOOK(hv) ? HvAUX(hv)->xhv_name : 0)
225#define HvNAME_get(hv) ((SvOOK(hv) && (HvAUX(hv)->xhv_name)) \
226 ? HEK_KEY(HvAUX(hv)->xhv_name) : 0)
227#define HvNAMELEN_get(hv) ((SvOOK(hv) && (HvAUX(hv)->xhv_name)) \
228 ? HEK_LEN(HvAUX(hv)->xhv_name) : 0)
a0d0e21e 229
8aacddc1
NIS
230/* the number of keys (including any placeholers) */
231#define XHvTOTALKEYS(xhv) ((xhv)->xhv_keys)
232
8aacddc1
NIS
233/*
234 * HvKEYS gets the number of keys that actually exist(), and is provided
235 * for backwards compatibility with old XS code. The core uses HvUSEDKEYS
236 * (keys, excluding placeholdes) and HvTOTALKEYS (including placeholders)
237 */
ca732855
NC
238#define HvKEYS(hv) HvUSEDKEYS(hv)
239#define HvUSEDKEYS(hv) (HvTOTALKEYS(hv) - HvPLACEHOLDERS_get(hv))
8aacddc1 240#define HvTOTALKEYS(hv) XHvTOTALKEYS((XPVHV*) SvANY(hv))
ca732855
NC
241#define HvPLACEHOLDERS(hv) (*Perl_hv_placeholders_p(aTHX_ (HV*)hv))
242#define HvPLACEHOLDERS_get(hv) (SvMAGIC(hv) ? Perl_hv_placeholders_get(aTHX_ (HV*)hv) : 0)
243#define HvPLACEHOLDERS_set(hv,p) Perl_hv_placeholders_set(aTHX_ (HV*)hv, p)
8aacddc1 244
bf6bd887
PP
245#define HvSHAREKEYS(hv) (SvFLAGS(hv) & SVphv_SHAREKEYS)
246#define HvSHAREKEYS_on(hv) (SvFLAGS(hv) |= SVphv_SHAREKEYS)
247#define HvSHAREKEYS_off(hv) (SvFLAGS(hv) &= ~SVphv_SHAREKEYS)
248
19692e8d
NC
249/* This is an optimisation flag. It won't be set if all hash keys have a 0
250 * flag. Currently the only flags relate to utf8.
251 * Hence it won't be set if all keys are 8 bit only. It will be set if any key
252 * is utf8 (including 8 bit keys that were entered as utf8, and need upgrading
253 * when retrieved during iteration. It may still be set when there are no longer
254 * any utf8 keys.
fdcd69b6 255 * See HVhek_ENABLEHVKFLAGS for the trigger.
19692e8d
NC
256 */
257#define HvHASKFLAGS(hv) (SvFLAGS(hv) & SVphv_HASKFLAGS)
258#define HvHASKFLAGS_on(hv) (SvFLAGS(hv) |= SVphv_HASKFLAGS)
259#define HvHASKFLAGS_off(hv) (SvFLAGS(hv) &= ~SVphv_HASKFLAGS)
574c8022 260
bf5b86ae
GS
261#define HvLAZYDEL(hv) (SvFLAGS(hv) & SVphv_LAZYDEL)
262#define HvLAZYDEL_on(hv) (SvFLAGS(hv) |= SVphv_LAZYDEL)
263#define HvLAZYDEL_off(hv) (SvFLAGS(hv) &= ~SVphv_LAZYDEL)
264
4b5190b5
NC
265#define HvREHASH(hv) (SvFLAGS(hv) & SVphv_REHASH)
266#define HvREHASH_on(hv) (SvFLAGS(hv) |= SVphv_REHASH)
267#define HvREHASH_off(hv) (SvFLAGS(hv) &= ~SVphv_REHASH)
268
a0d0e21e
LW
269/* Maybe amagical: */
270/* #define HV_AMAGICmb(hv) (SvFLAGS(hv) & (SVpgv_badAM | SVpgv_AM)) */
271
272#define HV_AMAGIC(hv) (SvFLAGS(hv) & SVpgv_AM)
273#define HV_AMAGIC_on(hv) (SvFLAGS(hv) |= SVpgv_AM)
274#define HV_AMAGIC_off(hv) (SvFLAGS(hv) &= ~SVpgv_AM)
275
276/*
277#define HV_AMAGICbad(hv) (SvFLAGS(hv) & SVpgv_badAM)
278#define HV_badAMAGIC_on(hv) (SvFLAGS(hv) |= SVpgv_badAM)
279#define HV_badAMAGIC_off(hv) (SvFLAGS(hv) &= ~SVpgv_badAM)
280*/
bf6bd887
PP
281
282#define Nullhe Null(HE*)
283#define HeNEXT(he) (he)->hent_next
ff68c719
PP
284#define HeKEY_hek(he) (he)->hent_hek
285#define HeKEY(he) HEK_KEY(HeKEY_hek(he))
bbce6d69 286#define HeKEY_sv(he) (*(SV**)HeKEY(he))
ff68c719 287#define HeKLEN(he) HEK_LEN(HeKEY_hek(he))
da58a35d 288#define HeKUTF8(he) HEK_UTF8(HeKEY_hek(he))
19692e8d 289#define HeKWASUTF8(he) HEK_WASUTF8(HeKEY_hek(he))
4b5190b5 290#define HeKREHASH(he) HEK_REHASH(HeKEY_hek(he))
da58a35d 291#define HeKLEN_UTF8(he) (HeKUTF8(he) ? -HeKLEN(he) : HeKLEN(he))
19692e8d 292#define HeKFLAGS(he) HEK_FLAGS(HeKEY_hek(he))
bf6bd887 293#define HeVAL(he) (he)->hent_val
ff68c719 294#define HeHASH(he) HEK_HASH(HeKEY_hek(he))
1e422769
PP
295#define HePV(he,lp) ((HeKLEN(he) == HEf_SVKEY) ? \
296 SvPV(HeKEY_sv(he),lp) : \
297 (((lp = HeKLEN(he)) >= 0) ? \
298 HeKEY(he) : Nullch))
299
bbce6d69
PP
300#define HeSVKEY(he) ((HeKEY(he) && \
301 HeKLEN(he) == HEf_SVKEY) ? \
302 HeKEY_sv(he) : Nullsv)
303
304#define HeSVKEY_force(he) (HeKEY(he) ? \
305 ((HeKLEN(he) == HEf_SVKEY) ? \
306 HeKEY_sv(he) : \
79cb57f6 307 sv_2mortal(newSVpvn(HeKEY(he), \
bbce6d69 308 HeKLEN(he)))) : \
3280af22 309 &PL_sv_undef)
1e422769 310#define HeSVKEY_set(he,sv) ((HeKLEN(he) = HEf_SVKEY), (HeKEY_sv(he) = sv))
bbce6d69 311
ff68c719 312#define Nullhek Null(HEK*)
71be2cbc 313#define HEK_BASESIZE STRUCT_OFFSET(HEK, hek_key[0])
ff68c719
PP
314#define HEK_HASH(hek) (hek)->hek_hash
315#define HEK_LEN(hek) (hek)->hek_len
316#define HEK_KEY(hek) (hek)->hek_key
19692e8d
NC
317#define HEK_FLAGS(hek) (*((unsigned char *)(HEK_KEY(hek))+HEK_LEN(hek)+1))
318
319#define HVhek_UTF8 0x01 /* Key is utf8 encoded. */
320#define HVhek_WASUTF8 0x02 /* Key is bytes here, but was supplied as utf8. */
4b5190b5 321#define HVhek_REHASH 0x04 /* This key is in an hv using a custom HASH . */
19692e8d 322#define HVhek_FREEKEY 0x100 /* Internal flag to say key is malloc()ed. */
e16e2ff8
NC
323#define HVhek_PLACEHOLD 0x200 /* Internal flag to create placeholder.
324 * (may change, but Storable is a core module) */
19692e8d
NC
325#define HVhek_MASK 0xFF
326
fdcd69b6
NC
327/* Which flags enable HvHASKFLAGS? Somewhat a hack on a hack, as
328 HVhek_REHASH is only needed because the rehash flag has to be duplicated
329 into all keys as hv_iternext has no access to the hash flags. At this
330 point Storable's tests get upset, because sometimes hashes are "keyed"
331 and sometimes not, depending on the order of data insertion, and whether
332 it triggered rehashing. So currently HVhek_REHAS is exempt.
333*/
334
335#define HVhek_ENABLEHVKFLAGS (HVhek_MASK - HVhek_REHASH)
336
19692e8d
NC
337#define HEK_UTF8(hek) (HEK_FLAGS(hek) & HVhek_UTF8)
338#define HEK_UTF8_on(hek) (HEK_FLAGS(hek) |= HVhek_UTF8)
339#define HEK_UTF8_off(hek) (HEK_FLAGS(hek) &= ~HVhek_UTF8)
340#define HEK_WASUTF8(hek) (HEK_FLAGS(hek) & HVhek_WASUTF8)
341#define HEK_WASUTF8_on(hek) (HEK_FLAGS(hek) |= HVhek_WASUTF8)
342#define HEK_WASUTF8_off(hek) (HEK_FLAGS(hek) &= ~HVhek_WASUTF8)
4b5190b5
NC
343#define HEK_REHASH(hek) (HEK_FLAGS(hek) & HVhek_REHASH)
344#define HEK_REHASH_on(hek) (HEK_FLAGS(hek) |= HVhek_REHASH)
d18c6117 345
5cbe4eec 346/* calculate HV array allocation */
36768cf4
NC
347#ifndef PERL_USE_LARGE_HV_ALLOC
348/* Default to allocating the correct size - default to assuming that malloc()
349 is not broken and is efficient at allocating blocks sized at powers-of-two.
350*/
d18c6117
GS
351# define PERL_HV_ARRAY_ALLOC_BYTES(size) ((size) * sizeof(HE*))
352#else
353# define MALLOC_OVERHEAD 16
354# define PERL_HV_ARRAY_ALLOC_BYTES(size) \
355 (((size) < 64) \
356 ? (size) * sizeof(HE*) \
357 : (size) * sizeof(HE*) * 2 - MALLOC_OVERHEAD)
358#endif
37d85e3a 359
e16e2ff8
NC
360/* Flags for hv_iternext_flags. */
361#define HV_ITERNEXT_WANTPLACEHOLDERS 0x01 /* Don't skip placeholders. */
362
37d85e3a
JH
363/* available as a function in hv.c */
364#define Perl_sharepvn(sv, len, hash) HEK_KEY(share_hek(sv, len, hash))
365#define sharepvn(sv, len, hash) Perl_sharepvn(sv, len, hash)
bfcb3514 366
d1db91c6
NC
367#define share_hek_hek(hek) \
368 (++(((struct shared_he *)(((char *)hek) \
369 - STRUCT_OFFSET(struct shared_he, \
370 shared_he_hek))) \
371 ->shared_he_he.hent_val), \
372 hek)
373
bfcb3514
NC
374/*
375 * Local variables:
376 * c-indentation-style: bsd
377 * c-basic-offset: 4
378 * indent-tabs-mode: t
379 * End:
380 *
381 * ex: set ts=8 sts=4 sw=4 noet:
382 */