src/pl/plpython/plpy_procedure.c

   1 /*
   2  * Python procedure manipulation for plpython
   3  *
   4  * src/pl/plpython/plpy_procedure.c
   5  */
   6
   7 #include "postgres.h"
   8
   9 #include "access/htup_details.h"
  10 #include "access/transam.h"
  11 #include "catalog/pg_proc.h"
  12 #include "catalog/pg_type.h"
  13 #include "funcapi.h"
  14 #include "plpy_elog.h"
  15 #include "plpy_main.h"
  16 #include "plpy_procedure.h"
  17 #include "plpython.h"
  18 #include "utils/builtins.h"
  19 #include "utils/hsearch.h"
  20 #include "utils/inval.h"
  21 #include "utils/lsyscache.h"
  22 #include "utils/memutils.h"
  23 #include "utils/syscache.h"
  24
  25 static HTAB *PLy_procedure_cache = NULL;
  26
  27 static PLyProcedure *PLy_procedure_create(HeapTuple procTup, Oid fn_oid, bool is_trigger);
  28 static bool PLy_procedure_valid(PLyProcedure *proc, HeapTuple procTup);
  29 static char *PLy_procedure_munge_source(const char *name, const char *src);
  30
  31
  32 void
  33 init_procedure_caches(void)
  34 {
  35         HASHCTL         hash_ctl;
  36
  37         hash_ctl.keysize = sizeof(PLyProcedureKey);
  38         hash_ctl.entrysize = sizeof(PLyProcedureEntry);
  39         PLy_procedure_cache = hash_create("PL/Python procedures", 32, &hash_ctl,
  40                                                                           HASH_ELEM | HASH_BLOBS);
  41 }
  42
  43 /*
  44  * PLy_procedure_name: get the name of the specified procedure.
  45  *
  46  * NB: this returns the SQL name, not the internal Python procedure name
  47  */
  48 char *
  49 PLy_procedure_name(PLyProcedure *proc)
  50 {
  51         if (proc == NULL)
  52                 return "<unknown procedure>";
  53         return proc->proname;
  54 }
  55
  56 /*
  57  * PLy_procedure_get: returns a cached PLyProcedure, or creates, stores and
  58  * returns a new PLyProcedure.
  59  *
  60  * fn_oid is the OID of the function requested
  61  * fn_rel is InvalidOid or the relation this function triggers on
  62  * is_trigger denotes whether the function is a trigger function
  63  *
  64  * The reason that both fn_rel and is_trigger need to be passed is that when
  65  * trigger functions get validated we don't know which relation(s) they'll
  66  * be used with, so no sensible fn_rel can be passed.
  67  */
  68 PLyProcedure *
  69 PLy_procedure_get(Oid fn_oid, Oid fn_rel, bool is_trigger)
  70 {
  71         bool            use_cache = !(is_trigger && fn_rel == InvalidOid);
  72         HeapTuple       procTup;
  73         PLyProcedureKey key;
  74         PLyProcedureEntry *volatile entry = NULL;
  75         PLyProcedure *volatile proc = NULL;
  76         bool            found = false;
  77
  78         procTup = SearchSysCache1(PROCOID, ObjectIdGetDatum(fn_oid));
  79         if (!HeapTupleIsValid(procTup))
  80                 elog(ERROR, "cache lookup failed for function %u", fn_oid);
  81
  82         /*
  83          * Look for the function in the cache, unless we don't have the necessary
  84          * information (e.g. during validation). In that case we just don't cache
  85          * anything.
  86          */
  87         if (use_cache)
  88         {
  89                 key.fn_oid = fn_oid;
  90                 key.fn_rel = fn_rel;
  91                 entry = hash_search(PLy_procedure_cache, &key, HASH_ENTER, &found);
  92                 proc = entry->proc;
  93         }
  94
  95         PG_TRY();
  96         {
  97                 if (!found)
  98                 {
  99                         /* Haven't found it, create a new procedure */
 100                         proc = PLy_procedure_create(procTup, fn_oid, is_trigger);
 101                         if (use_cache)
 102                                 entry->proc = proc;
 103                 }
 104                 else if (!PLy_procedure_valid(proc, procTup))
 105                 {
 106                         /* Found it, but it's invalid, free and reuse the cache entry */
 107                         entry->proc = NULL;
 108                         if (proc)
 109                                 PLy_procedure_delete(proc);
 110                         proc = PLy_procedure_create(procTup, fn_oid, is_trigger);
 111                         entry->proc = proc;
 112                 }
 113                 /* Found it and it's valid, it's fine to use it */
 114         }
 115         PG_CATCH();
 116         {
 117                 /* Do not leave an uninitialized entry in the cache */
 118                 if (use_cache)
 119                         hash_search(PLy_procedure_cache, &key, HASH_REMOVE, NULL);
 120                 PG_RE_THROW();
 121         }
 122         PG_END_TRY();
 123
 124         ReleaseSysCache(procTup);
 125
 126         return proc;
 127 }
 128
 129 /*
 130  * Create a new PLyProcedure structure
 131  */
 132 static PLyProcedure *
 133 PLy_procedure_create(HeapTuple procTup, Oid fn_oid, bool is_trigger)
 134 {
 135         char            procName[NAMEDATALEN + 256];
 136         Form_pg_proc procStruct;
 137         PLyProcedure *volatile proc;
 138         MemoryContext cxt;
 139         MemoryContext oldcxt;
 140         int                     rv;
 141         char       *ptr;
 142
 143         procStruct = (Form_pg_proc) GETSTRUCT(procTup);
 144         rv = snprintf(procName, sizeof(procName),
 145                                   "__plpython_procedure_%s_%u",
 146                                   NameStr(procStruct->proname),
 147                                   fn_oid);
 148         if (rv >= sizeof(procName) || rv < 0)
 149                 elog(ERROR, "procedure name would overrun buffer");
 150
 151         /* Replace any not-legal-in-Python-names characters with '_' */
 152         for (ptr = procName; *ptr; ptr++)
 153         {
 154                 if (!((*ptr >= 'A' && *ptr <= 'Z') ||
 155                           (*ptr >= 'a' && *ptr <= 'z') ||
 156                           (*ptr >= '0' && *ptr <= '9')))
 157                         *ptr = '_';
 158         }
 159
 160         /* Create long-lived context that all procedure info will live in */
 161         cxt = AllocSetContextCreate(TopMemoryContext,
 162                                                                 "PL/Python function",
 163                                                                 ALLOCSET_DEFAULT_SIZES);
 164
 165         oldcxt = MemoryContextSwitchTo(cxt);
 166
 167         proc = (PLyProcedure *) palloc0(sizeof(PLyProcedure));
 168         proc->mcxt = cxt;
 169
 170         PG_TRY();
 171         {
 172                 Datum           protrftypes_datum;
 173                 Datum           prosrcdatum;
 174                 bool            isnull;
 175                 char       *procSource;
 176                 int                     i;
 177
 178                 proc->proname = pstrdup(NameStr(procStruct->proname));
 179                 MemoryContextSetIdentifier(cxt, proc->proname);
 180                 proc->pyname = pstrdup(procName);
 181                 proc->fn_xmin = HeapTupleHeaderGetRawXmin(procTup->t_data);
 182                 proc->fn_tid = procTup->t_self;
 183                 proc->fn_readonly = (procStruct->provolatile != PROVOLATILE_VOLATILE);
 184                 proc->is_setof = procStruct->proretset;
 185                 proc->is_procedure = (procStruct->prokind == PROKIND_PROCEDURE);
 186                 proc->is_trigger = is_trigger;
 187                 proc->src = NULL;
 188                 proc->argnames = NULL;
 189                 proc->args = NULL;
 190                 proc->nargs = 0;
 191                 proc->langid = procStruct->prolang;
 192                 protrftypes_datum = SysCacheGetAttr(PROCOID, procTup,
 193                                                                                         Anum_pg_proc_protrftypes,
 194                                                                                         &isnull);
 195                 proc->trftypes = isnull ? NIL : oid_array_to_list(protrftypes_datum);
 196                 proc->code = NULL;
 197                 proc->statics = NULL;
 198                 proc->globals = NULL;
 199                 proc->calldepth = 0;
 200                 proc->argstack = NULL;
 201
 202                 /*
 203                  * get information required for output conversion of the return value,
 204                  * but only if this isn't a trigger.
 205                  */
 206                 if (!is_trigger)
 207                 {
 208                         Oid                     rettype = procStruct->prorettype;
 209                         HeapTuple       rvTypeTup;
 210                         Form_pg_type rvTypeStruct;
 211
 212                         rvTypeTup = SearchSysCache1(TYPEOID, ObjectIdGetDatum(rettype));
 213                         if (!HeapTupleIsValid(rvTypeTup))
 214                                 elog(ERROR, "cache lookup failed for type %u", rettype);
 215                         rvTypeStruct = (Form_pg_type) GETSTRUCT(rvTypeTup);
 216
 217                         /* Disallow pseudotype result, except for void or record */
 218                         if (rvTypeStruct->typtype == TYPTYPE_PSEUDO)
 219                         {
 220                                 if (rettype == VOIDOID ||
 221                                         rettype == RECORDOID)
 222                                          /* okay */ ;
 223                                 else if (rettype == TRIGGEROID || rettype == EVENT_TRIGGEROID)
 224                                         ereport(ERROR,
 225                                                         (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
 226                                                          errmsg("trigger functions can only be called as triggers")));
 227                                 else
 228                                         ereport(ERROR,
 229                                                         (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
 230                                                          errmsg("PL/Python functions cannot return type %s",
 231                                                                         format_type_be(rettype))));
 232                         }
 233
 234                         /* set up output function for procedure result */
 235                         PLy_output_setup_func(&proc->result, proc->mcxt,
 236                                                                   rettype, -1, proc);
 237
 238                         ReleaseSysCache(rvTypeTup);
 239                 }
 240                 else
 241                 {
 242                         /*
 243                          * In a trigger function, we use proc->result and proc->result_in
 244                          * for converting tuples, but we don't yet have enough info to set
 245                          * them up.  PLy_exec_trigger will deal with it.
 246                          */
 247                         proc->result.typoid = InvalidOid;
 248                         proc->result_in.typoid = InvalidOid;
 249                 }
 250
 251                 /*
 252                  * Now get information required for input conversion of the
 253                  * procedure's arguments.  Note that we ignore output arguments here.
 254                  * If the function returns record, those I/O functions will be set up
 255                  * when the function is first called.
 256                  */
 257                 if (procStruct->pronargs)
 258                 {
 259                         Oid                *types;
 260                         char      **names,
 261                                            *modes;
 262                         int                     pos,
 263                                                 total;
 264
 265                         /* extract argument type info from the pg_proc tuple */
 266                         total = get_func_arg_info(procTup, &types, &names, &modes);
 267
 268                         /* count number of in+inout args into proc->nargs */
 269                         if (modes == NULL)
 270                                 proc->nargs = total;
 271                         else
 272                         {
 273                                 /* proc->nargs was initialized to 0 above */
 274                                 for (i = 0; i < total; i++)
 275                                 {
 276                                         if (modes[i] != PROARGMODE_OUT &&
 277                                                 modes[i] != PROARGMODE_TABLE)
 278                                                 (proc->nargs)++;
 279                                 }
 280                         }
 281
 282                         /* Allocate arrays for per-input-argument data */
 283                         proc->argnames = (char **) palloc0(sizeof(char *) * proc->nargs);
 284                         proc->args = (PLyDatumToOb *) palloc0(sizeof(PLyDatumToOb) * proc->nargs);
 285
 286                         for (i = pos = 0; i < total; i++)
 287                         {
 288                                 HeapTuple       argTypeTup;
 289                                 Form_pg_type argTypeStruct;
 290
 291                                 if (modes &&
 292                                         (modes[i] == PROARGMODE_OUT ||
 293                                          modes[i] == PROARGMODE_TABLE))
 294                                         continue;       /* skip OUT arguments */
 295
 296                                 Assert(types[i] == procStruct->proargtypes.values[pos]);
 297
 298                                 argTypeTup = SearchSysCache1(TYPEOID,
 299                                                                                          ObjectIdGetDatum(types[i]));
 300                                 if (!HeapTupleIsValid(argTypeTup))
 301                                         elog(ERROR, "cache lookup failed for type %u", types[i]);
 302                                 argTypeStruct = (Form_pg_type) GETSTRUCT(argTypeTup);
 303
 304                                 /* disallow pseudotype arguments */
 305                                 if (argTypeStruct->typtype == TYPTYPE_PSEUDO)
 306                                         ereport(ERROR,
 307                                                         (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
 308                                                          errmsg("PL/Python functions cannot accept type %s",
 309                                                                         format_type_be(types[i]))));
 310
 311                                 /* set up I/O function info */
 312                                 PLy_input_setup_func(&proc->args[pos], proc->mcxt,
 313                                                                          types[i], -1,  /* typmod not known */
 314                                                                          proc);
 315
 316                                 /* get argument name */
 317                                 proc->argnames[pos] = names ? pstrdup(names[i]) : NULL;
 318
 319                                 ReleaseSysCache(argTypeTup);
 320
 321                                 pos++;
 322                         }
 323                 }
 324
 325                 /*
 326                  * get the text of the function.
 327                  */
 328                 prosrcdatum = SysCacheGetAttrNotNull(PROCOID, procTup,
 329                                                                                          Anum_pg_proc_prosrc);
 330                 procSource = TextDatumGetCString(prosrcdatum);
 331
 332                 PLy_procedure_compile(proc, procSource);
 333
 334                 pfree(procSource);
 335         }
 336         PG_CATCH();
 337         {
 338                 MemoryContextSwitchTo(oldcxt);
 339                 PLy_procedure_delete(proc);
 340                 PG_RE_THROW();
 341         }
 342         PG_END_TRY();
 343
 344         MemoryContextSwitchTo(oldcxt);
 345         return proc;
 346 }
 347
 348 /*
 349  * Insert the procedure into the Python interpreter
 350  */
 351 void
 352 PLy_procedure_compile(PLyProcedure *proc, const char *src)
 353 {
 354         PyObject   *crv = NULL;
 355         char       *msrc;
 356
 357         proc->globals = PyDict_Copy(PLy_interp_globals);
 358
 359         /*
 360          * SD is private preserved data between calls. GD is global data shared by
 361          * all functions
 362          */
 363         proc->statics = PyDict_New();
 364         if (!proc->statics)
 365                 PLy_elog(ERROR, NULL);
 366         PyDict_SetItemString(proc->globals, "SD", proc->statics);
 367
 368         /*
 369          * insert the function code into the interpreter
 370          */
 371         msrc = PLy_procedure_munge_source(proc->pyname, src);
 372         /* Save the mangled source for later inclusion in tracebacks */
 373         proc->src = MemoryContextStrdup(proc->mcxt, msrc);
 374         crv = PyRun_String(msrc, Py_file_input, proc->globals, NULL);
 375         pfree(msrc);
 376
 377         if (crv != NULL)
 378         {
 379                 int                     clen;
 380                 char            call[NAMEDATALEN + 256];
 381
 382                 Py_DECREF(crv);
 383
 384                 /*
 385                  * compile a call to the function
 386                  */
 387                 clen = snprintf(call, sizeof(call), "%s()", proc->pyname);
 388                 if (clen < 0 || clen >= sizeof(call))
 389                         elog(ERROR, "string would overflow buffer");
 390                 proc->code = Py_CompileString(call, "<string>", Py_eval_input);
 391                 if (proc->code != NULL)
 392                         return;
 393         }
 394
 395         if (proc->proname)
 396                 PLy_elog(ERROR, "could not compile PL/Python function \"%s\"",
 397                                  proc->proname);
 398         else
 399                 PLy_elog(ERROR, "could not compile anonymous PL/Python code block");
 400 }
 401
 402 void
 403 PLy_procedure_delete(PLyProcedure *proc)
 404 {
 405         Py_XDECREF(proc->code);
 406         Py_XDECREF(proc->statics);
 407         Py_XDECREF(proc->globals);
 408         MemoryContextDelete(proc->mcxt);
 409 }
 410
 411 /*
 412  * Decide whether a cached PLyProcedure struct is still valid
 413  */
 414 static bool
 415 PLy_procedure_valid(PLyProcedure *proc, HeapTuple procTup)
 416 {
 417         if (proc == NULL)
 418                 return false;
 419
 420         /* If the pg_proc tuple has changed, it's not valid */
 421         if (!(proc->fn_xmin == HeapTupleHeaderGetRawXmin(procTup->t_data) &&
 422                   ItemPointerEquals(&proc->fn_tid, &procTup->t_self)))
 423                 return false;
 424
 425         return true;
 426 }
 427
 428 static char *
 429 PLy_procedure_munge_source(const char *name, const char *src)
 430 {
 431         char       *mrc,
 432                            *mp;
 433         const char *sp;
 434         size_t          mlen;
 435         int                     plen;
 436
 437         /*
 438          * room for function source and the def statement
 439          */
 440         mlen = (strlen(src) * 2) + strlen(name) + 16;
 441
 442         mrc = palloc(mlen);
 443         plen = snprintf(mrc, mlen, "def %s():\n\t", name);
 444         Assert(plen >= 0 && plen < mlen);
 445
 446         sp = src;
 447         mp = mrc + plen;
 448
 449         while (*sp != '\0')
 450         {
 451                 if (*sp == '\r' && *(sp + 1) == '\n')
 452                         sp++;
 453
 454                 if (*sp == '\n' || *sp == '\r')
 455                 {
 456                         *mp++ = '\n';
 457                         *mp++ = '\t';
 458                         sp++;
 459                 }
 460                 else
 461                         *mp++ = *sp++;
 462         }
 463         *mp++ = '\n';
 464         *mp++ = '\n';
 465         *mp = '\0';
 466
 467         if (mp > (mrc + mlen))
 468                 elog(FATAL, "buffer overrun in PLy_procedure_munge_source");
 469
 470         return mrc;
 471 }