This commit was manufactured by cvs2svn to create tag 'r234c1'.
[python/dscho.git] / PC / getpathp.c
blob7c87da85481880d8ff7a1c7c5b48f6d216e32ea9
2 /* Return the initial module search path. */
3 /* Used by DOS, OS/2, Windows 3.1, Windows 95/98, Windows NT. */
5 /* ----------------------------------------------------------------
6 PATH RULES FOR WINDOWS:
7 This describes how sys.path is formed on Windows. It describes the
8 functionality, not the implementation (ie, the order in which these
9 are actually fetched is different)
11 * Python always adds an empty entry at the start, which corresponds
12 to the current directory.
14 * If the PYTHONPATH env. var. exists, it's entries are added next.
16 * We look in the registry for "application paths" - that is, sub-keys
17 under the main PythonPath registry key. These are added next (the
18 order of sub-key processing is undefined).
19 HKEY_CURRENT_USER is searched and added first.
20 HKEY_LOCAL_MACHINE is searched and added next.
21 (Note that all known installers only use HKLM, so HKCU is typically
22 empty)
24 * We attempt to locate the "Python Home" - if the PYTHONHOME env var
25 is set, we believe it. Otherwise, we use the path of our host .EXE's
26 to try and locate our "landmark" (lib\\os.py) and deduce our home.
27 - If we DO have a Python Home: The relevant sub-directories (Lib,
28 plat-win, lib-tk, etc) are based on the Python Home
29 - If we DO NOT have a Python Home, the core Python Path is
30 loaded from the registry. This is the main PythonPath key,
31 and both HKLM and HKCU are combined to form the path)
33 * Iff - we can not locate the Python Home, have not had a PYTHONPATH
34 specified, and can't locate any Registry entries (ie, we have _nothing_
35 we can assume is a good path), a default path with relative entries is
36 used (eg. .\Lib;.\plat-win, etc)
39 The end result of all this is:
40 * When running python.exe, or any other .exe in the main Python directory
41 (either an installed version, or directly from the PCbuild directory),
42 the core path is deduced, and the core paths in the registry are
43 ignored. Other "application paths" in the registry are always read.
45 * When Python is hosted in another exe (different directory, embedded via
46 COM, etc), the Python Home will not be deduced, so the core path from
47 the registry is used. Other "application paths" in the registry are
48 always read.
50 * If Python can't find its home and there is no registry (eg, frozen
51 exe, some very strange installation setup) you get a path with
52 some default, but relative, paths.
54 ---------------------------------------------------------------- */
57 #include "Python.h"
58 #include "osdefs.h"
60 #ifdef MS_WINDOWS
61 #include <windows.h>
62 #include <tchar.h>
63 #endif
65 #include <sys/types.h>
66 #include <sys/stat.h>
67 #include <string.h>
69 /* Search in some common locations for the associated Python libraries.
71 * Py_GetPath() tries to return a sensible Python module search path.
73 * The approach is an adaptation for Windows of the strategy used in
74 * ../Modules/getpath.c; it uses the Windows Registry as one of its
75 * information sources.
78 #ifndef LANDMARK
79 #define LANDMARK "lib\\os.py"
80 #endif
82 static char prefix[MAXPATHLEN+1];
83 static char progpath[MAXPATHLEN+1];
84 static char dllpath[MAXPATHLEN+1];
85 static char *module_search_path = NULL;
88 static int
89 is_sep(char ch) /* determine if "ch" is a separator character */
91 #ifdef ALTSEP
92 return ch == SEP || ch == ALTSEP;
93 #else
94 return ch == SEP;
95 #endif
98 /* assumes 'dir' null terminated in bounds. Never writes
99 beyond existing terminator.
101 static void
102 reduce(char *dir)
104 size_t i = strlen(dir);
105 while (i > 0 && !is_sep(dir[i]))
106 --i;
107 dir[i] = '\0';
111 static int
112 exists(char *filename)
114 struct stat buf;
115 return stat(filename, &buf) == 0;
118 /* Assumes 'filename' MAXPATHLEN+1 bytes long -
119 may extend 'filename' by one character.
121 static int
122 ismodule(char *filename) /* Is module -- check for .pyc/.pyo too */
124 if (exists(filename))
125 return 1;
127 /* Check for the compiled version of prefix. */
128 if (strlen(filename) < MAXPATHLEN) {
129 strcat(filename, Py_OptimizeFlag ? "o" : "c");
130 if (exists(filename))
131 return 1;
133 return 0;
136 /* guarantees buffer will never overflow MAXPATHLEN+1 bytes */
137 static void
138 join(char *buffer, char *stuff)
140 size_t n, k;
141 if (is_sep(stuff[0]))
142 n = 0;
143 else {
144 n = strlen(buffer);
145 if (n > 0 && !is_sep(buffer[n-1]) && n < MAXPATHLEN)
146 buffer[n++] = SEP;
148 k = strlen(stuff);
149 if (n + k > MAXPATHLEN)
150 k = MAXPATHLEN - n;
151 strncpy(buffer+n, stuff, k);
152 buffer[n+k] = '\0';
155 /* gotlandmark only called by search_for_prefix, which ensures
156 'prefix' is null terminated in bounds. join() ensures
157 'landmark' can not overflow prefix if too long.
159 static int
160 gotlandmark(char *landmark)
162 int n, ok;
164 n = strlen(prefix);
165 join(prefix, landmark);
166 ok = ismodule(prefix);
167 prefix[n] = '\0';
168 return ok;
171 /* assumes argv0_path is MAXPATHLEN+1 bytes long, already \0 term'd.
172 assumption provided by only caller, calculate_path() */
173 static int
174 search_for_prefix(char *argv0_path, char *landmark)
176 /* Search from argv0_path, until landmark is found */
177 strcpy(prefix, argv0_path);
178 do {
179 if (gotlandmark(landmark))
180 return 1;
181 reduce(prefix);
182 } while (prefix[0]);
183 return 0;
186 #ifdef MS_WINDOWS
188 /* a string loaded from the DLL at startup.*/
189 extern const char *PyWin_DLLVersionString;
192 /* Load a PYTHONPATH value from the registry.
193 Load from either HKEY_LOCAL_MACHINE or HKEY_CURRENT_USER.
195 Works in both Unicode and 8bit environments. Only uses the
196 Ex family of functions so it also works with Windows CE.
198 Returns NULL, or a pointer that should be freed.
200 XXX - this code is pretty strange, as it used to also
201 work on Win16, where the buffer sizes werent available
202 in advance. It could be simplied now Win16/Win32s is dead!
205 static char *
206 getpythonregpath(HKEY keyBase, int skipcore)
208 HKEY newKey = 0;
209 DWORD dataSize = 0;
210 DWORD numKeys = 0;
211 LONG rc;
212 char *retval = NULL;
213 TCHAR *dataBuf = NULL;
214 static const TCHAR keyPrefix[] = _T("Software\\Python\\PythonCore\\");
215 static const TCHAR keySuffix[] = _T("\\PythonPath");
216 size_t versionLen;
217 DWORD index;
218 TCHAR *keyBuf = NULL;
219 TCHAR *keyBufPtr;
220 TCHAR **ppPaths = NULL;
222 /* Tried to use sysget("winver") but here is too early :-( */
223 versionLen = _tcslen(PyWin_DLLVersionString);
224 /* Space for all the chars, plus one \0 */
225 keyBuf = keyBufPtr = malloc(sizeof(keyPrefix) +
226 sizeof(TCHAR)*(versionLen-1) +
227 sizeof(keySuffix));
228 if (keyBuf==NULL) goto done;
230 memcpy(keyBufPtr, keyPrefix, sizeof(keyPrefix)-sizeof(TCHAR));
231 keyBufPtr += sizeof(keyPrefix)/sizeof(TCHAR) - 1;
232 memcpy(keyBufPtr, PyWin_DLLVersionString, versionLen * sizeof(TCHAR));
233 keyBufPtr += versionLen;
234 /* NULL comes with this one! */
235 memcpy(keyBufPtr, keySuffix, sizeof(keySuffix));
236 /* Open the root Python key */
237 rc=RegOpenKeyEx(keyBase,
238 keyBuf, /* subkey */
239 0, /* reserved */
240 KEY_READ,
241 &newKey);
242 if (rc!=ERROR_SUCCESS) goto done;
243 /* Find out how big our core buffer is, and how many subkeys we have */
244 rc = RegQueryInfoKey(newKey, NULL, NULL, NULL, &numKeys, NULL, NULL,
245 NULL, NULL, &dataSize, NULL, NULL);
246 if (rc!=ERROR_SUCCESS) goto done;
247 if (skipcore) dataSize = 0; /* Only count core ones if we want them! */
248 /* Allocate a temp array of char buffers, so we only need to loop
249 reading the registry once
251 ppPaths = malloc( sizeof(TCHAR *) * numKeys );
252 if (ppPaths==NULL) goto done;
253 memset(ppPaths, 0, sizeof(TCHAR *) * numKeys);
254 /* Loop over all subkeys, allocating a temp sub-buffer. */
255 for(index=0;index<numKeys;index++) {
256 TCHAR keyBuf[MAX_PATH+1];
257 HKEY subKey = 0;
258 DWORD reqdSize = MAX_PATH+1;
259 /* Get the sub-key name */
260 DWORD rc = RegEnumKeyEx(newKey, index, keyBuf, &reqdSize,
261 NULL, NULL, NULL, NULL );
262 if (rc!=ERROR_SUCCESS) goto done;
263 /* Open the sub-key */
264 rc=RegOpenKeyEx(newKey,
265 keyBuf, /* subkey */
266 0, /* reserved */
267 KEY_READ,
268 &subKey);
269 if (rc!=ERROR_SUCCESS) goto done;
270 /* Find the value of the buffer size, malloc, then read it */
271 RegQueryValueEx(subKey, NULL, 0, NULL, NULL, &reqdSize);
272 if (reqdSize) {
273 ppPaths[index] = malloc(reqdSize);
274 if (ppPaths[index]) {
275 RegQueryValueEx(subKey, NULL, 0, NULL,
276 (LPBYTE)ppPaths[index],
277 &reqdSize);
278 dataSize += reqdSize + 1; /* 1 for the ";" */
281 RegCloseKey(subKey);
283 /* original datasize from RegQueryInfo doesn't include the \0 */
284 dataBuf = malloc((dataSize+1) * sizeof(TCHAR));
285 if (dataBuf) {
286 TCHAR *szCur = dataBuf;
287 DWORD reqdSize = dataSize;
288 /* Copy our collected strings */
289 for (index=0;index<numKeys;index++) {
290 if (index > 0) {
291 *(szCur++) = _T(';');
292 dataSize--;
294 if (ppPaths[index]) {
295 int len = _tcslen(ppPaths[index]);
296 _tcsncpy(szCur, ppPaths[index], len);
297 szCur += len;
298 dataSize -= len;
301 if (skipcore)
302 *szCur = '\0';
303 else {
304 /* If we have no values, we dont need a ';' */
305 if (numKeys) {
306 *(szCur++) = _T(';');
307 dataSize--;
309 /* Now append the core path entries -
310 this will include the NULL
312 rc = RegQueryValueEx(newKey, NULL, 0, NULL,
313 (LPBYTE)szCur, &dataSize);
315 /* And set the result - caller must free
316 If MBCS, it is fine as is. If Unicode, allocate new
317 buffer and convert.
319 #ifdef UNICODE
320 retval = (char *)malloc(reqdSize+1);
321 if (retval)
322 WideCharToMultiByte(CP_ACP, 0,
323 dataBuf, -1, /* source */
324 retval, reqdSize+1, /* dest */
325 NULL, NULL);
326 free(dataBuf);
327 #else
328 retval = dataBuf;
329 #endif
331 done:
332 /* Loop freeing my temp buffers */
333 if (ppPaths) {
334 for(index=0;index<numKeys;index++)
335 if (ppPaths[index]) free(ppPaths[index]);
336 free(ppPaths);
338 if (newKey)
339 RegCloseKey(newKey);
340 if (keyBuf)
341 free(keyBuf);
342 return retval;
344 #endif /* MS_WINDOWS */
346 static void
347 get_progpath(void)
349 extern char *Py_GetProgramName(void);
350 char *path = getenv("PATH");
351 char *prog = Py_GetProgramName();
353 #ifdef MS_WINDOWS
354 extern HANDLE PyWin_DLLhModule;
355 #ifdef UNICODE
356 WCHAR wprogpath[MAXPATHLEN+1];
357 /* Windows documents that GetModuleFileName() will "truncate",
358 but makes no mention of the null terminator. Play it safe.
359 PLUS Windows itself defines MAX_PATH as the same, but anyway...
361 wprogpath[MAXPATHLEN]=_T('\0');
362 if (PyWin_DLLhModule &&
363 GetModuleFileName(PyWin_DLLhModule, wprogpath, MAXPATHLEN)) {
364 WideCharToMultiByte(CP_ACP, 0,
365 wprogpath, -1,
366 dllpath, MAXPATHLEN+1,
367 NULL, NULL);
369 wprogpath[MAXPATHLEN]=_T('\0');
370 if (GetModuleFileName(NULL, wprogpath, MAXPATHLEN)) {
371 WideCharToMultiByte(CP_ACP, 0,
372 wprogpath, -1,
373 progpath, MAXPATHLEN+1,
374 NULL, NULL);
375 return;
377 #else
378 /* static init of progpath ensures final char remains \0 */
379 if (PyWin_DLLhModule)
380 if (!GetModuleFileName(PyWin_DLLhModule, dllpath, MAXPATHLEN))
381 dllpath[0] = 0;
382 if (GetModuleFileName(NULL, progpath, MAXPATHLEN))
383 return;
384 #endif
385 #endif
386 if (prog == NULL || *prog == '\0')
387 prog = "python";
389 /* If there is no slash in the argv0 path, then we have to
390 * assume python is on the user's $PATH, since there's no
391 * other way to find a directory to start the search from. If
392 * $PATH isn't exported, you lose.
394 #ifdef ALTSEP
395 if (strchr(prog, SEP) || strchr(prog, ALTSEP))
396 #else
397 if (strchr(prog, SEP))
398 #endif
399 strncpy(progpath, prog, MAXPATHLEN);
400 else if (path) {
401 while (1) {
402 char *delim = strchr(path, DELIM);
404 if (delim) {
405 size_t len = delim - path;
406 /* ensure we can't overwrite buffer */
407 len = min(MAXPATHLEN,len);
408 strncpy(progpath, path, len);
409 *(progpath + len) = '\0';
411 else
412 strncpy(progpath, path, MAXPATHLEN);
414 /* join() is safe for MAXPATHLEN+1 size buffer */
415 join(progpath, prog);
416 if (exists(progpath))
417 break;
419 if (!delim) {
420 progpath[0] = '\0';
421 break;
423 path = delim + 1;
426 else
427 progpath[0] = '\0';
430 static void
431 calculate_path(void)
433 char argv0_path[MAXPATHLEN+1];
434 char *buf;
435 size_t bufsz;
436 char *pythonhome = Py_GetPythonHome();
437 char *envpath = Py_GETENV("PYTHONPATH");
439 #ifdef MS_WINDOWS
440 int skiphome, skipdefault;
441 char *machinepath = NULL;
442 char *userpath = NULL;
443 char zip_path[MAXPATHLEN+1];
444 size_t len;
445 #endif
447 get_progpath();
448 /* progpath guaranteed \0 terminated in MAXPATH+1 bytes. */
449 strcpy(argv0_path, progpath);
450 reduce(argv0_path);
451 if (pythonhome == NULL || *pythonhome == '\0') {
452 if (search_for_prefix(argv0_path, LANDMARK))
453 pythonhome = prefix;
454 else
455 pythonhome = NULL;
457 else
458 strncpy(prefix, pythonhome, MAXPATHLEN);
460 if (envpath && *envpath == '\0')
461 envpath = NULL;
464 #ifdef MS_WINDOWS
465 /* Calculate zip archive path */
466 if (dllpath[0]) /* use name of python DLL */
467 strncpy(zip_path, dllpath, MAXPATHLEN);
468 else /* use name of executable program */
469 strncpy(zip_path, progpath, MAXPATHLEN);
470 zip_path[MAXPATHLEN] = '\0';
471 len = strlen(zip_path);
472 if (len > 4) {
473 zip_path[len-3] = 'z'; /* change ending to "zip" */
474 zip_path[len-2] = 'i';
475 zip_path[len-1] = 'p';
477 else {
478 zip_path[0] = 0;
481 skiphome = pythonhome==NULL ? 0 : 1;
482 machinepath = getpythonregpath(HKEY_LOCAL_MACHINE, skiphome);
483 userpath = getpythonregpath(HKEY_CURRENT_USER, skiphome);
484 /* We only use the default relative PYTHONPATH if we havent
485 anything better to use! */
486 skipdefault = envpath!=NULL || pythonhome!=NULL || \
487 machinepath!=NULL || userpath!=NULL;
488 #endif
490 /* We need to construct a path from the following parts.
491 (1) the PYTHONPATH environment variable, if set;
492 (2) for Win32, the zip archive file path;
493 (3) for Win32, the machinepath and userpath, if set;
494 (4) the PYTHONPATH config macro, with the leading "."
495 of each component replaced with pythonhome, if set;
496 (5) the directory containing the executable (argv0_path).
497 The length calculation calculates #4 first.
498 Extra rules:
499 - If PYTHONHOME is set (in any way) item (3) is ignored.
500 - If registry values are used, (4) and (5) are ignored.
503 /* Calculate size of return buffer */
504 if (pythonhome != NULL) {
505 char *p;
506 bufsz = 1;
507 for (p = PYTHONPATH; *p; p++) {
508 if (*p == DELIM)
509 bufsz++; /* number of DELIM plus one */
511 bufsz *= strlen(pythonhome);
513 else
514 bufsz = 0;
515 bufsz += strlen(PYTHONPATH) + 1;
516 bufsz += strlen(argv0_path) + 1;
517 #ifdef MS_WINDOWS
518 if (userpath)
519 bufsz += strlen(userpath) + 1;
520 if (machinepath)
521 bufsz += strlen(machinepath) + 1;
522 bufsz += strlen(zip_path) + 1;
523 #endif
524 if (envpath != NULL)
525 bufsz += strlen(envpath) + 1;
527 module_search_path = buf = malloc(bufsz);
528 if (buf == NULL) {
529 /* We can't exit, so print a warning and limp along */
530 fprintf(stderr, "Can't malloc dynamic PYTHONPATH.\n");
531 if (envpath) {
532 fprintf(stderr, "Using environment $PYTHONPATH.\n");
533 module_search_path = envpath;
535 else {
536 fprintf(stderr, "Using default static path.\n");
537 module_search_path = PYTHONPATH;
539 #ifdef MS_WINDOWS
540 if (machinepath)
541 free(machinepath);
542 if (userpath)
543 free(userpath);
544 #endif /* MS_WINDOWS */
545 return;
548 if (envpath) {
549 strcpy(buf, envpath);
550 buf = strchr(buf, '\0');
551 *buf++ = DELIM;
553 #ifdef MS_WINDOWS
554 if (zip_path[0]) {
555 strcpy(buf, zip_path);
556 buf = strchr(buf, '\0');
557 *buf++ = DELIM;
559 if (userpath) {
560 strcpy(buf, userpath);
561 buf = strchr(buf, '\0');
562 *buf++ = DELIM;
563 free(userpath);
565 if (machinepath) {
566 strcpy(buf, machinepath);
567 buf = strchr(buf, '\0');
568 *buf++ = DELIM;
569 free(machinepath);
571 if (pythonhome == NULL) {
572 if (!skipdefault) {
573 strcpy(buf, PYTHONPATH);
574 buf = strchr(buf, '\0');
577 #else
578 if (pythonhome == NULL) {
579 strcpy(buf, PYTHONPATH);
580 buf = strchr(buf, '\0');
582 #endif /* MS_WINDOWS */
583 else {
584 char *p = PYTHONPATH;
585 char *q;
586 size_t n;
587 for (;;) {
588 q = strchr(p, DELIM);
589 if (q == NULL)
590 n = strlen(p);
591 else
592 n = q-p;
593 if (p[0] == '.' && is_sep(p[1])) {
594 strcpy(buf, pythonhome);
595 buf = strchr(buf, '\0');
596 p++;
597 n--;
599 strncpy(buf, p, n);
600 buf += n;
601 if (q == NULL)
602 break;
603 *buf++ = DELIM;
604 p = q+1;
607 if (argv0_path) {
608 *buf++ = DELIM;
609 strcpy(buf, argv0_path);
610 buf = strchr(buf, '\0');
612 *buf = '\0';
613 /* Now to pull one last hack/trick. If sys.prefix is
614 empty, then try and find it somewhere on the paths
615 we calculated. We scan backwards, as our general policy
616 is that Python core directories are at the *end* of
617 sys.path. We assume that our "lib" directory is
618 on the path, and that our 'prefix' directory is
619 the parent of that.
621 if (*prefix=='\0') {
622 char lookBuf[MAXPATHLEN+1];
623 char *look = buf - 1; /* 'buf' is at the end of the buffer */
624 while (1) {
625 int nchars;
626 char *lookEnd = look;
627 /* 'look' will end up one character before the
628 start of the path in question - even if this
629 is one character before the start of the buffer
631 while (*look != DELIM && look >= module_search_path)
632 look--;
633 nchars = lookEnd-look;
634 strncpy(lookBuf, look+1, nchars);
635 lookBuf[nchars] = '\0';
636 /* Up one level to the parent */
637 reduce(lookBuf);
638 if (search_for_prefix(lookBuf, LANDMARK)) {
639 break;
641 /* If we are out of paths to search - give up */
642 if (look < module_search_path)
643 break;
644 look--;
650 /* External interface */
652 char *
653 Py_GetPath(void)
655 if (!module_search_path)
656 calculate_path();
657 return module_search_path;
660 char *
661 Py_GetPrefix(void)
663 if (!module_search_path)
664 calculate_path();
665 return prefix;
668 char *
669 Py_GetExecPrefix(void)
671 return Py_GetPrefix();
674 char *
675 Py_GetProgramFullPath(void)
677 if (!module_search_path)
678 calculate_path();
679 return progpath;