cpython/Modules/zipimport.c

1239 lines
37 KiB
C
Raw Normal View History

#include "Python.h"
#include "structmember.h"
#include "osdefs.h"
#include "marshal.h"
#include <time.h>
#define IS_SOURCE 0x0
#define IS_BYTECODE 0x1
#define IS_PACKAGE 0x2
struct st_zip_searchorder {
char suffix[14];
int type;
};
/* zip_searchorder defines how we search for a module in the Zip
archive: we first search for a package __init__, then for
non-package .pyc, .pyo and .py entries. The .pyc and .pyo entries
are swapped by initzipimport() if we run in optimized mode. Also,
'/' is replaced by SEP there. */
2003-03-23 09:21:03 -04:00
static struct st_zip_searchorder zip_searchorder[] = {
{"/__init__.pyc", IS_PACKAGE | IS_BYTECODE},
{"/__init__.pyo", IS_PACKAGE | IS_BYTECODE},
{"/__init__.py", IS_PACKAGE | IS_SOURCE},
{".pyc", IS_BYTECODE},
{".pyo", IS_BYTECODE},
{".py", IS_SOURCE},
{"", 0}
};
/* zipimporter object definition and support */
typedef struct _zipimporter ZipImporter;
struct _zipimporter {
PyObject_HEAD
PyObject *archive; /* pathname of the Zip archive */
PyObject *prefix; /* file prefix: "a/sub/directory/" */
PyObject *files; /* dict with file info {path: toc_entry} */
};
static PyObject *ZipImportError;
static PyObject *zip_directory_cache = NULL;
/* forward decls */
static PyObject *read_directory(char *archive);
static PyObject *get_data(char *archive, PyObject *toc_entry);
static PyObject *get_module_code(ZipImporter *self, char *fullname,
int *p_ispackage, char **p_modpath);
#define ZipImporter_Check(op) PyObject_TypeCheck(op, &ZipImporter_Type)
/* zipimporter.__init__
Split the "subdirectory" from the Zip archive path, lookup a matching
entry in sys.path_importer_cache, fetch the file directory from there
if found, or else read it from the archive. */
static int
zipimporter_init(ZipImporter *self, PyObject *args, PyObject *kwds)
{
char *path, *p, *prefix, buf[MAXPATHLEN+2];
size_t len;
if (!_PyArg_NoKeywords("zipimporter()", kwds))
return -1;
if (!PyArg_ParseTuple(args, "s:zipimporter", &path))
return -1;
len = strlen(path);
if (len == 0) {
PyErr_SetString(ZipImportError, "archive path is empty");
return -1;
}
if (len >= MAXPATHLEN) {
PyErr_SetString(ZipImportError,
"archive path too long");
return -1;
}
strcpy(buf, path);
#ifdef ALTSEP
for (p = buf; *p; p++) {
if (*p == ALTSEP)
*p = SEP;
}
#endif
path = NULL;
prefix = NULL;
for (;;) {
struct stat statbuf;
int rv;
rv = stat(buf, &statbuf);
if (rv == 0) {
/* it exists */
if (S_ISREG(statbuf.st_mode))
/* it's a file */
path = buf;
break;
}
/* back up one path element */
p = strrchr(buf, SEP);
if (prefix != NULL)
*prefix = SEP;
if (p == NULL)
break;
*p = '\0';
prefix = p;
}
if (path != NULL) {
PyObject *files;
files = PyDict_GetItemString(zip_directory_cache, path);
if (files == NULL) {
files = read_directory(buf);
if (files == NULL)
return -1;
if (PyDict_SetItemString(zip_directory_cache, path,
files) != 0)
return -1;
}
else
Py_INCREF(files);
self->files = files;
}
else {
PyErr_SetString(ZipImportError, "not a Zip file");
return -1;
}
if (prefix == NULL)
prefix = "";
else {
prefix++;
len = strlen(prefix);
if (prefix[len-1] != SEP) {
/* add trailing SEP */
prefix[len] = SEP;
prefix[len + 1] = '\0';
}
}
self->archive = PyUnicode_FromString(buf);
if (self->archive == NULL)
return -1;
self->prefix = PyUnicode_FromString(prefix);
if (self->prefix == NULL)
return -1;
return 0;
}
/* GC support. */
static int
zipimporter_traverse(PyObject *obj, visitproc visit, void *arg)
{
ZipImporter *self = (ZipImporter *)obj;
Py_VISIT(self->files);
return 0;
}
static void
zipimporter_dealloc(ZipImporter *self)
{
PyObject_GC_UnTrack(self);
Py_XDECREF(self->archive);
Py_XDECREF(self->prefix);
Py_XDECREF(self->files);
Py_TYPE(self)->tp_free((PyObject *)self);
}
static PyObject *
zipimporter_repr(ZipImporter *self)
{
char *archive = "???";
char *prefix = "";
if (self->archive != NULL && PyUnicode_Check(self->archive))
archive = _PyUnicode_AsString(self->archive);
if (self->prefix != NULL && PyUnicode_Check(self->prefix))
prefix = _PyUnicode_AsString(self->prefix);
if (prefix != NULL && *prefix)
return PyUnicode_FromFormat("<zipimporter object \"%.300s%c%.150s\">",
archive, SEP, prefix);
else
return PyUnicode_FromFormat("<zipimporter object \"%.300s\">",
archive);
}
/* return fullname.split(".")[-1] */
static char *
get_subname(char *fullname)
{
char *subname = strrchr(fullname, '.');
if (subname == NULL)
subname = fullname;
else
subname++;
return subname;
}
/* Given a (sub)modulename, write the potential file path in the
archive (without extension) to the path buffer. Return the
length of the resulting string. */
static int
make_filename(char *prefix, char *name, char *path)
{
size_t len;
char *p;
len = strlen(prefix);
/* self.prefix + name [+ SEP + "__init__"] + ".py[co]" */
if (len + strlen(name) + 13 >= MAXPATHLEN) {
PyErr_SetString(ZipImportError, "path too long");
return -1;
}
strcpy(path, prefix);
strcpy(path + len, name);
for (p = path + len; *p; p++) {
if (*p == '.')
*p = SEP;
}
len += strlen(name);
assert(len < INT_MAX);
return (int)len;
}
enum zi_module_info {
MI_ERROR,
MI_NOT_FOUND,
MI_MODULE,
MI_PACKAGE
};
/* Return some information about a module. */
static enum zi_module_info
get_module_info(ZipImporter *self, char *fullname)
{
char *subname, path[MAXPATHLEN + 1];
int len;
struct st_zip_searchorder *zso;
subname = get_subname(fullname);
len = make_filename(_PyUnicode_AsString(self->prefix), subname, path);
if (len < 0)
return MI_ERROR;
for (zso = zip_searchorder; *zso->suffix; zso++) {
strcpy(path + len, zso->suffix);
if (PyDict_GetItemString(self->files, path) != NULL) {
if (zso->type & IS_PACKAGE)
return MI_PACKAGE;
else
return MI_MODULE;
}
}
return MI_NOT_FOUND;
}
/* Check whether we can satisfy the import of the module named by
'fullname'. Return self if we can, None if we can't. */
static PyObject *
zipimporter_find_module(PyObject *obj, PyObject *args)
{
ZipImporter *self = (ZipImporter *)obj;
PyObject *path = NULL;
char *fullname;
enum zi_module_info mi;
if (!PyArg_ParseTuple(args, "s|O:zipimporter.find_module",
&fullname, &path))
return NULL;
mi = get_module_info(self, fullname);
if (mi == MI_ERROR)
return NULL;
if (mi == MI_NOT_FOUND) {
Py_INCREF(Py_None);
return Py_None;
}
Py_INCREF(self);
return (PyObject *)self;
}
/* Load and return the module named by 'fullname'. */
static PyObject *
zipimporter_load_module(PyObject *obj, PyObject *args)
{
ZipImporter *self = (ZipImporter *)obj;
PyObject *code, *mod, *dict;
char *fullname, *modpath;
int ispackage;
if (!PyArg_ParseTuple(args, "s:zipimporter.load_module",
&fullname))
return NULL;
code = get_module_code(self, fullname, &ispackage, &modpath);
if (code == NULL)
return NULL;
mod = PyImport_AddModule(fullname);
if (mod == NULL) {
Py_DECREF(code);
return NULL;
}
dict = PyModule_GetDict(mod);
/* mod.__loader__ = self */
if (PyDict_SetItemString(dict, "__loader__", (PyObject *)self) != 0)
goto error;
if (ispackage) {
/* add __path__ to the module *before* the code gets
executed */
PyObject *pkgpath, *fullpath;
char *subname = get_subname(fullname);
int err;
fullpath = PyUnicode_FromFormat("%U%c%U%s",
self->archive, SEP,
self->prefix, subname);
if (fullpath == NULL)
goto error;
pkgpath = Py_BuildValue("[O]", fullpath);
Py_DECREF(fullpath);
if (pkgpath == NULL)
goto error;
err = PyDict_SetItemString(dict, "__path__", pkgpath);
Py_DECREF(pkgpath);
if (err != 0)
goto error;
}
mod = PyImport_ExecCodeModuleEx(fullname, code, modpath);
Py_DECREF(code);
if (Py_VerboseFlag)
PySys_WriteStderr("import %s # loaded from Zip %s\n",
fullname, modpath);
return mod;
error:
Py_DECREF(code);
Py_DECREF(mod);
return NULL;
}
/* Return a string matching __file__ for the named module */
static PyObject *
zipimporter_get_filename(PyObject *obj, PyObject *args)
{
ZipImporter *self = (ZipImporter *)obj;
PyObject *code;
char *fullname, *modpath;
int ispackage;
if (!PyArg_ParseTuple(args, "s:zipimporter.get_filename",
&fullname))
return NULL;
/* Deciding the filename requires working out where the code
would come from if the module was actually loaded */
code = get_module_code(self, fullname, &ispackage, &modpath);
if (code == NULL)
return NULL;
Py_DECREF(code); /* Only need the path info */
return PyUnicode_FromString(modpath);
}
/* Return a bool signifying whether the module is a package or not. */
static PyObject *
zipimporter_is_package(PyObject *obj, PyObject *args)
{
ZipImporter *self = (ZipImporter *)obj;
char *fullname;
enum zi_module_info mi;
if (!PyArg_ParseTuple(args, "s:zipimporter.is_package",
&fullname))
return NULL;
mi = get_module_info(self, fullname);
if (mi == MI_ERROR)
return NULL;
if (mi == MI_NOT_FOUND) {
PyErr_Format(ZipImportError, "can't find module '%.200s'",
fullname);
return NULL;
}
return PyBool_FromLong(mi == MI_PACKAGE);
}
static PyObject *
zipimporter_get_data(PyObject *obj, PyObject *args)
{
ZipImporter *self = (ZipImporter *)obj;
char *path;
#ifdef ALTSEP
char *p, buf[MAXPATHLEN + 1];
#endif
PyObject *toc_entry;
Py_ssize_t len;
char *archive_str;
if (!PyArg_ParseTuple(args, "s:zipimporter.get_data", &path))
return NULL;
#ifdef ALTSEP
if (strlen(path) >= MAXPATHLEN) {
PyErr_SetString(ZipImportError, "path too long");
return NULL;
}
strcpy(buf, path);
for (p = buf; *p; p++) {
if (*p == ALTSEP)
*p = SEP;
}
path = buf;
#endif
archive_str = _PyUnicode_AsStringAndSize(self->archive, &len);
if ((size_t)len < strlen(path) &&
strncmp(path, archive_str, len) == 0 &&
path[len] == SEP) {
path = path + len + 1;
}
toc_entry = PyDict_GetItemString(self->files, path);
if (toc_entry == NULL) {
PyErr_SetFromErrnoWithFilename(PyExc_IOError, path);
return NULL;
}
return get_data(archive_str, toc_entry);
}
static PyObject *
zipimporter_get_code(PyObject *obj, PyObject *args)
{
ZipImporter *self = (ZipImporter *)obj;
char *fullname;
if (!PyArg_ParseTuple(args, "s:zipimporter.get_code", &fullname))
return NULL;
return get_module_code(self, fullname, NULL, NULL);
}
static PyObject *
zipimporter_get_source(PyObject *obj, PyObject *args)
{
ZipImporter *self = (ZipImporter *)obj;
PyObject *toc_entry;
char *fullname, *subname, path[MAXPATHLEN+1];
int len;
enum zi_module_info mi;
if (!PyArg_ParseTuple(args, "s:zipimporter.get_source", &fullname))
return NULL;
mi = get_module_info(self, fullname);
if (mi == MI_ERROR)
return NULL;
if (mi == MI_NOT_FOUND) {
PyErr_Format(ZipImportError, "can't find module '%.200s'",
fullname);
return NULL;
}
subname = get_subname(fullname);
len = make_filename(_PyUnicode_AsString(self->prefix), subname, path);
if (len < 0)
return NULL;
if (mi == MI_PACKAGE) {
path[len] = SEP;
strcpy(path + len + 1, "__init__.py");
}
else
strcpy(path + len, ".py");
toc_entry = PyDict_GetItemString(self->files, path);
if (toc_entry != NULL) {
PyObject *bytes = get_data(_PyUnicode_AsString(self->archive), toc_entry);
PyObject *res = PyUnicode_FromString(PyBytes_AsString(bytes));
Py_XDECREF(bytes);
return res;
}
/* we have the module, but no source */
Py_INCREF(Py_None);
return Py_None;
}
PyDoc_STRVAR(doc_find_module,
"find_module(fullname, path=None) -> self or None.\n\
\n\
Search for a module specified by 'fullname'. 'fullname' must be the\n\
fully qualified (dotted) module name. It returns the zipimporter\n\
instance itself if the module was found, or None if it wasn't.\n\
The optional 'path' argument is ignored -- it's there for compatibility\n\
with the importer protocol.");
PyDoc_STRVAR(doc_load_module,
"load_module(fullname) -> module.\n\
\n\
Load the module specified by 'fullname'. 'fullname' must be the\n\
fully qualified (dotted) module name. It returns the imported\n\
module, or raises ZipImportError if it wasn't found.");
PyDoc_STRVAR(doc_get_data,
"get_data(pathname) -> string with file data.\n\
\n\
Return the data associated with 'pathname'. Raise IOError if\n\
the file wasn't found.");
PyDoc_STRVAR(doc_is_package,
"is_package(fullname) -> bool.\n\
\n\
Return True if the module specified by fullname is a package.\n\
Raise ZipImportError if the module couldn't be found.");
PyDoc_STRVAR(doc_get_code,
"get_code(fullname) -> code object.\n\
\n\
Return the code object for the specified module. Raise ZipImportError\n\
if the module couldn't be found.");
PyDoc_STRVAR(doc_get_source,
"get_source(fullname) -> source string.\n\
\n\
Return the source code for the specified module. Raise ZipImportError\n\
if the module couldn't be found, return None if the archive does\n\
contain the module, but has no source for it.");
PyDoc_STRVAR(doc_get_filename,
"get_filename(fullname) -> filename string.\n\
\n\
Return the filename for the specified module.");
static PyMethodDef zipimporter_methods[] = {
{"find_module", zipimporter_find_module, METH_VARARGS,
doc_find_module},
{"load_module", zipimporter_load_module, METH_VARARGS,
doc_load_module},
{"get_data", zipimporter_get_data, METH_VARARGS,
doc_get_data},
{"get_code", zipimporter_get_code, METH_VARARGS,
doc_get_code},
{"get_source", zipimporter_get_source, METH_VARARGS,
doc_get_source},
{"get_filename", zipimporter_get_filename, METH_VARARGS,
doc_get_filename},
{"is_package", zipimporter_is_package, METH_VARARGS,
doc_is_package},
{NULL, NULL} /* sentinel */
};
static PyMemberDef zipimporter_members[] = {
{"archive", T_OBJECT, offsetof(ZipImporter, archive), READONLY},
{"prefix", T_OBJECT, offsetof(ZipImporter, prefix), READONLY},
{"_files", T_OBJECT, offsetof(ZipImporter, files), READONLY},
{NULL}
};
PyDoc_STRVAR(zipimporter_doc,
"zipimporter(archivepath) -> zipimporter object\n\
\n\
Create a new zipimporter instance. 'archivepath' must be a path to\n\
Merged revisions 63066-63076,63079,63081-63085,63087-63097,63099,63101-63104 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r63066 | georg.brandl | 2008-05-11 10:56:04 -0400 (Sun, 11 May 2008) | 2 lines #2709 followup: better description of Tk's pros and cons. ........ r63067 | georg.brandl | 2008-05-11 11:05:13 -0400 (Sun, 11 May 2008) | 2 lines #1326: document and test zipimporter.archive and zipimporter.prefix. ........ r63068 | georg.brandl | 2008-05-11 11:07:39 -0400 (Sun, 11 May 2008) | 2 lines #2816: clarify error messages for EOF while scanning strings. ........ r63069 | georg.brandl | 2008-05-11 11:17:41 -0400 (Sun, 11 May 2008) | 3 lines #2787: Flush stdout after writing test name, helpful when running hanging or long-running tests. Patch by Adam Olsen. ........ r63070 | georg.brandl | 2008-05-11 11:20:16 -0400 (Sun, 11 May 2008) | 3 lines #2803: fix wrong invocation of heappush in seldom-reached code. Thanks to Matt Harden. ........ r63073 | benjamin.peterson | 2008-05-11 12:38:07 -0400 (Sun, 11 May 2008) | 2 lines broaden .bzrignore ........ r63076 | andrew.kuchling | 2008-05-11 15:15:52 -0400 (Sun, 11 May 2008) | 1 line Add message to test assertion ........ r63083 | andrew.kuchling | 2008-05-11 16:08:33 -0400 (Sun, 11 May 2008) | 1 line Try setting HOME env.var to fix test on Win32 ........ r63092 | georg.brandl | 2008-05-11 16:53:55 -0400 (Sun, 11 May 2008) | 2 lines #2809 followup: even better split docstring. ........ r63094 | georg.brandl | 2008-05-11 17:03:42 -0400 (Sun, 11 May 2008) | 4 lines - #2250: Exceptions raised during evaluation of names in rlcompleter's ``Completer.complete()`` method are now caught and ignored. ........ r63095 | georg.brandl | 2008-05-11 17:16:37 -0400 (Sun, 11 May 2008) | 2 lines Clarify os.strerror()s exception behavior. ........ r63097 | georg.brandl | 2008-05-11 17:34:10 -0400 (Sun, 11 May 2008) | 2 lines #2535: remove duplicated method. ........ r63104 | alexandre.vassalotti | 2008-05-11 19:04:27 -0400 (Sun, 11 May 2008) | 2 lines Moved the Queue module stub in lib-old. ........
2008-05-15 21:41:41 -03:00
a zipfile, or to a specific path inside a zipfile. For example, it can be\n\
'/tmp/myimport.zip', or '/tmp/myimport.zip/mydirectory', if mydirectory is a\n\
valid directory inside the archive.\n\
\n\
'ZipImportError is raised if 'archivepath' doesn't point to a valid Zip\n\
archive.\n\
\n\
The 'archive' attribute of zipimporter objects contains the name of the\n\
zipfile targeted.");
#define DEFERRED_ADDRESS(ADDR) 0
static PyTypeObject ZipImporter_Type = {
PyVarObject_HEAD_INIT(DEFERRED_ADDRESS(&PyType_Type), 0)
"zipimport.zipimporter",
sizeof(ZipImporter),
0, /* tp_itemsize */
(destructor)zipimporter_dealloc, /* tp_dealloc */
0, /* tp_print */
0, /* tp_getattr */
0, /* tp_setattr */
0, /* tp_reserved */
(reprfunc)zipimporter_repr, /* tp_repr */
0, /* tp_as_number */
0, /* tp_as_sequence */
0, /* tp_as_mapping */
0, /* tp_hash */
0, /* tp_call */
0, /* tp_str */
PyObject_GenericGetAttr, /* tp_getattro */
0, /* tp_setattro */
0, /* tp_as_buffer */
Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE |
Py_TPFLAGS_HAVE_GC, /* tp_flags */
zipimporter_doc, /* tp_doc */
zipimporter_traverse, /* tp_traverse */
0, /* tp_clear */
0, /* tp_richcompare */
0, /* tp_weaklistoffset */
0, /* tp_iter */
0, /* tp_iternext */
zipimporter_methods, /* tp_methods */
zipimporter_members, /* tp_members */
0, /* tp_getset */
0, /* tp_base */
0, /* tp_dict */
0, /* tp_descr_get */
0, /* tp_descr_set */
0, /* tp_dictoffset */
(initproc)zipimporter_init, /* tp_init */
PyType_GenericAlloc, /* tp_alloc */
PyType_GenericNew, /* tp_new */
PyObject_GC_Del, /* tp_free */
};
/* implementation */
/* Given a buffer, return the long that is represented by the first
4 bytes, encoded as little endian. This partially reimplements
marshal.c:r_long() */
static long
get_long(unsigned char *buf) {
long x;
x = buf[0];
x |= (long)buf[1] << 8;
x |= (long)buf[2] << 16;
x |= (long)buf[3] << 24;
#if SIZEOF_LONG > 4
/* Sign extension for 64-bit machines */
x |= -(x & 0x80000000L);
#endif
return x;
}
/*
read_directory(archive) -> files dict (new reference)
Given a path to a Zip archive, build a dict, mapping file names
(local to the archive, using SEP as a separator) to toc entries.
A toc_entry is a tuple:
2005-11-11 15:34:56 -04:00
(__file__, # value to use for __file__, available for all files
compress, # compression kind; 0 for uncompressed
data_size, # size of compressed data on disk
file_size, # size of decompressed data
file_offset, # offset of file header from start of archive
time, # mod time of file (in dos format)
date, # mod data of file (in dos format)
crc, # crc checksum of the data
)
Directories can be recognized by the trailing SEP in the name,
data_size and file_offset are 0.
*/
static PyObject *
read_directory(char *archive)
{
PyObject *files = NULL;
FILE *fp;
long compress, crc, data_size, file_size, file_offset, date, time;
long header_offset, name_size, header_size, header_position;
long i, l, count;
size_t length;
char path[MAXPATHLEN + 5];
char name[MAXPATHLEN + 5];
char *p, endof_central_dir[22];
long arc_offset; /* offset from beginning of file to start of zip-archive */
if (strlen(archive) > MAXPATHLEN) {
PyErr_SetString(PyExc_OverflowError,
"Zip path name is too long");
return NULL;
}
strcpy(path, archive);
fp = fopen(archive, "rb");
if (fp == NULL) {
PyErr_Format(ZipImportError, "can't open Zip file: "
"'%.200s'", archive);
return NULL;
}
fseek(fp, -22, SEEK_END);
header_position = ftell(fp);
if (fread(endof_central_dir, 1, 22, fp) != 22) {
fclose(fp);
PyErr_Format(ZipImportError, "can't read Zip file: "
"'%.200s'", archive);
return NULL;
}
if (get_long((unsigned char *)endof_central_dir) != 0x06054B50) {
/* Bad: End of Central Dir signature */
fclose(fp);
PyErr_Format(ZipImportError, "not a Zip file: "
"'%.200s'", archive);
return NULL;
}
header_size = get_long((unsigned char *)endof_central_dir + 12);
header_offset = get_long((unsigned char *)endof_central_dir + 16);
arc_offset = header_position - header_offset - header_size;
header_offset += arc_offset;
files = PyDict_New();
if (files == NULL)
goto error;
length = (long)strlen(path);
path[length] = SEP;
/* Start of Central Directory */
count = 0;
for (;;) {
PyObject *t;
int err;
fseek(fp, header_offset, 0); /* Start of file header */
l = PyMarshal_ReadLongFromFile(fp);
if (l != 0x02014B50)
break; /* Bad: Central Dir File Header */
fseek(fp, header_offset + 10, 0);
compress = PyMarshal_ReadShortFromFile(fp);
time = PyMarshal_ReadShortFromFile(fp);
date = PyMarshal_ReadShortFromFile(fp);
crc = PyMarshal_ReadLongFromFile(fp);
data_size = PyMarshal_ReadLongFromFile(fp);
file_size = PyMarshal_ReadLongFromFile(fp);
name_size = PyMarshal_ReadShortFromFile(fp);
header_size = 46 + name_size +
PyMarshal_ReadShortFromFile(fp) +
PyMarshal_ReadShortFromFile(fp);
fseek(fp, header_offset + 42, 0);
file_offset = PyMarshal_ReadLongFromFile(fp) + arc_offset;
if (name_size > MAXPATHLEN)
name_size = MAXPATHLEN;
p = name;
for (i = 0; i < name_size; i++) {
*p = (char)getc(fp);
if (*p == '/')
*p = SEP;
p++;
}
*p = 0; /* Add terminating null byte */
header_offset += header_size;
strncpy(path + length + 1, name, MAXPATHLEN - length - 1);
t = Py_BuildValue("siiiiiii", path, compress, data_size,
file_size, file_offset, time, date, crc);
if (t == NULL)
goto error;
err = PyDict_SetItemString(files, name, t);
Py_DECREF(t);
if (err != 0)
goto error;
count++;
}
fclose(fp);
if (Py_VerboseFlag)
PySys_WriteStderr("# zipimport: found %ld names in %s\n",
count, archive);
return files;
error:
fclose(fp);
Py_XDECREF(files);
return NULL;
}
/* Return the zlib.decompress function object, or NULL if zlib couldn't
be imported. The function is cached when found, so subsequent calls
don't import zlib again. Returns a *borrowed* reference.
XXX This makes zlib.decompress immortal. */
static PyObject *
get_decompress_func(void)
{
static PyObject *decompress = NULL;
if (decompress == NULL) {
PyObject *zlib;
static int importing_zlib = 0;
if (importing_zlib != 0)
/* Someone has a zlib.py[co] in their Zip file;
let's avoid a stack overflow. */
return NULL;
importing_zlib = 1;
zlib = PyImport_ImportModuleNoBlock("zlib");
importing_zlib = 0;
if (zlib != NULL) {
decompress = PyObject_GetAttrString(zlib,
"decompress");
Py_DECREF(zlib);
}
else
PyErr_Clear();
if (Py_VerboseFlag)
PySys_WriteStderr("# zipimport: zlib %s\n",
zlib != NULL ? "available": "UNAVAILABLE");
}
return decompress;
}
/* Given a path to a Zip file and a toc_entry, return the (uncompressed)
data as a new reference. */
static PyObject *
get_data(char *archive, PyObject *toc_entry)
{
PyObject *raw_data, *data = NULL, *decompress;
char *buf;
FILE *fp;
int err;
Py_ssize_t bytes_read = 0;
long l;
char *datapath;
long compress, data_size, file_size, file_offset, bytes_size;
long time, date, crc;
if (!PyArg_ParseTuple(toc_entry, "slllllll", &datapath, &compress,
&data_size, &file_size, &file_offset, &time,
&date, &crc)) {
return NULL;
}
fp = fopen(archive, "rb");
if (!fp) {
PyErr_Format(PyExc_IOError,
"zipimport: can not open file %s", archive);
return NULL;
}
/* Check to make sure the local file header is correct */
fseek(fp, file_offset, 0);
l = PyMarshal_ReadLongFromFile(fp);
if (l != 0x04034B50) {
/* Bad: Local File Header */
PyErr_Format(ZipImportError,
"bad local file header in %s",
archive);
fclose(fp);
return NULL;
}
fseek(fp, file_offset + 26, 0);
l = 30 + PyMarshal_ReadShortFromFile(fp) +
PyMarshal_ReadShortFromFile(fp); /* local header size */
file_offset += l; /* Start of file data */
bytes_size = compress == 0 ? data_size : data_size + 1;
if (bytes_size == 0)
bytes_size++;
raw_data = PyBytes_FromStringAndSize((char *)NULL, bytes_size);
if (raw_data == NULL) {
fclose(fp);
return NULL;
}
buf = PyBytes_AsString(raw_data);
err = fseek(fp, file_offset, 0);
if (err == 0)
bytes_read = fread(buf, 1, data_size, fp);
fclose(fp);
if (err || bytes_read != data_size) {
PyErr_SetString(PyExc_IOError,
"zipimport: can't read data");
Py_DECREF(raw_data);
return NULL;
}
if (compress != 0) {
buf[data_size] = 'Z'; /* saw this in zipfile.py */
data_size++;
}
buf[data_size] = '\0';
if (compress == 0) { /* data is not compressed */
data = PyBytes_FromStringAndSize(buf, data_size);
Py_DECREF(raw_data);
return data;
}
/* Decompress with zlib */
decompress = get_decompress_func();
if (decompress == NULL) {
PyErr_SetString(ZipImportError,
"can't decompress data; "
"zlib not available");
goto error;
}
data = PyObject_CallFunction(decompress, "Oi", raw_data, -15);
error:
Py_DECREF(raw_data);
return data;
}
/* Lenient date/time comparison function. The precision of the mtime
in the archive is lower than the mtime stored in a .pyc: we
must allow a difference of at most one second. */
static int
eq_mtime(time_t t1, time_t t2)
{
time_t d = t1 - t2;
if (d < 0)
d = -d;
/* dostime only stores even seconds, so be lenient */
return d <= 1;
}
/* Given the contents of a .py[co] file in a buffer, unmarshal the data
and return the code object. Return None if it the magic word doesn't
match (we do this instead of raising an exception as we fall back
to .py if available and we don't want to mask other errors).
Returns a new reference. */
static PyObject *
unmarshal_code(char *pathname, PyObject *data, time_t mtime)
{
PyObject *code;
char *buf = PyBytes_AsString(data);
Py_ssize_t size = PyBytes_Size(data);
if (size <= 9) {
PyErr_SetString(ZipImportError,
"bad pyc data");
return NULL;
}
if (get_long((unsigned char *)buf) != PyImport_GetMagicNumber()) {
if (Py_VerboseFlag)
PySys_WriteStderr("# %s has bad magic\n",
pathname);
Py_INCREF(Py_None);
return Py_None; /* signal caller to try alternative */
}
if (mtime != 0 && !eq_mtime(get_long((unsigned char *)buf + 4),
mtime)) {
if (Py_VerboseFlag)
PySys_WriteStderr("# %s has bad mtime\n",
pathname);
Py_INCREF(Py_None);
return Py_None; /* signal caller to try alternative */
}
code = PyMarshal_ReadObjectFromString(buf + 8, size - 8);
if (code == NULL)
return NULL;
if (!PyCode_Check(code)) {
Py_DECREF(code);
PyErr_Format(PyExc_TypeError,
"compiled module %.200s is not a code object",
pathname);
return NULL;
}
return code;
}
/* Replace any occurances of "\r\n?" in the input string with "\n".
This converts DOS and Mac line endings to Unix line endings.
Also append a trailing "\n" to be compatible with
PyParser_SimpleParseFile(). Returns a new reference. */
static PyObject *
normalize_line_endings(PyObject *source)
{
char *buf, *q, *p = PyBytes_AsString(source);
PyObject *fixed_source;
int len = 0;
if (!p) {
return PyBytes_FromStringAndSize("\n\0", 2);
}
/* one char extra for trailing \n and one for terminating \0 */
buf = (char *)PyMem_Malloc(PyBytes_Size(source) + 2);
if (buf == NULL) {
PyErr_SetString(PyExc_MemoryError,
"zipimport: no memory to allocate "
"source buffer");
return NULL;
}
/* replace "\r\n?" by "\n" */
for (q = buf; *p != '\0'; p++) {
if (*p == '\r') {
*q++ = '\n';
if (*(p + 1) == '\n')
p++;
}
else
*q++ = *p;
len++;
}
*q++ = '\n'; /* add trailing \n */
*q = '\0';
fixed_source = PyBytes_FromStringAndSize(buf, len + 2);
PyMem_Free(buf);
return fixed_source;
}
/* Given a string buffer containing Python source code, compile it
return and return a code object as a new reference. */
static PyObject *
compile_source(char *pathname, PyObject *source)
{
PyObject *code, *fixed_source;
fixed_source = normalize_line_endings(source);
if (fixed_source == NULL)
return NULL;
code = Py_CompileString(PyBytes_AsString(fixed_source), pathname,
Py_file_input);
Py_DECREF(fixed_source);
return code;
}
/* Convert the date/time values found in the Zip archive to a value
that's compatible with the time stamp stored in .pyc files. */
2003-03-23 09:21:03 -04:00
static time_t
parse_dostime(int dostime, int dosdate)
{
struct tm stm;
memset((void *) &stm, '\0', sizeof(stm));
Merged revisions 59985-60000,60002,60005-60007,60009-60042 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r59987 | raymond.hettinger | 2008-01-15 21:52:42 +0100 (Tue, 15 Jan 2008) | 1 line Refactor if/elif chain for clarity and speed. Remove dependency on subclasses having to implement _empty and _full. ........ r59988 | raymond.hettinger | 2008-01-15 22:22:47 +0100 (Tue, 15 Jan 2008) | 1 line Fix-up half-written paragraph in the docs ........ r59989 | amaury.forgeotdarc | 2008-01-15 22:25:11 +0100 (Tue, 15 Jan 2008) | 3 lines test_doctest fails since r59984. Not sure if these are the correct values, but save_stdout has to be set before its usage... ........ r59992 | andrew.kuchling | 2008-01-16 01:32:03 +0100 (Wed, 16 Jan 2008) | 1 line Docstring typos ........ r59993 | andrew.kuchling | 2008-01-16 04:17:25 +0100 (Wed, 16 Jan 2008) | 1 line Add PEP 3141 section ........ r59998 | andrew.kuchling | 2008-01-16 14:01:51 +0100 (Wed, 16 Jan 2008) | 1 line Markup fix ........ r59999 | georg.brandl | 2008-01-16 17:56:29 +0100 (Wed, 16 Jan 2008) | 2 lines Fix MSDN library URL. (#1854) ........ r60006 | georg.brandl | 2008-01-16 21:27:56 +0100 (Wed, 16 Jan 2008) | 3 lines Add Python-specific content to Doc dir. Update configuration file to work with the newest Sphinx. ........ r60007 | georg.brandl | 2008-01-16 21:29:00 +0100 (Wed, 16 Jan 2008) | 2 lines Doc build should work with 2.4 now. ........ r60009 | raymond.hettinger | 2008-01-17 00:38:16 +0100 (Thu, 17 Jan 2008) | 1 line Minor wordsmithing. ........ r60010 | raymond.hettinger | 2008-01-17 00:40:45 +0100 (Thu, 17 Jan 2008) | 1 line Add queues will alternative fetch orders (priority based and stack based). ........ r60011 | raymond.hettinger | 2008-01-17 00:49:35 +0100 (Thu, 17 Jan 2008) | 1 line Add news entry. ........ r60013 | raymond.hettinger | 2008-01-17 04:02:14 +0100 (Thu, 17 Jan 2008) | 1 line Make starmap() match its pure python definition and accept any itertable input (not just tuples). ........ r60015 | gregory.p.smith | 2008-01-17 08:43:20 +0100 (Thu, 17 Jan 2008) | 3 lines Comply with RFC 3207. Fixes issue 829951 - http://bugs.python.org/issue829951 ........ r60018 | gregory.p.smith | 2008-01-17 09:03:17 +0100 (Thu, 17 Jan 2008) | 2 lines entry for r60015 ........ r60019 | raymond.hettinger | 2008-01-17 09:07:05 +0100 (Thu, 17 Jan 2008) | 1 line Note versionadded. ........ r60020 | gregory.p.smith | 2008-01-17 09:35:49 +0100 (Thu, 17 Jan 2008) | 8 lines Fixes (accepts patch) issue1339 - http://bugs.python.org/issue1339 - Factor out the duplication of EHLO/HELO in login() and sendmail() to a new function, ehlo_or_helo_if_needed(). - Use ehlo_or_helo_if_needed() in starttls() - Check for the starttls exception in starttls() in the same way as login() checks for the auth extension. Contributed by Bill Fenner. ........ r60021 | andrew.kuchling | 2008-01-17 13:00:15 +0100 (Thu, 17 Jan 2008) | 1 line Revise 3141 section a bit; add some Windows items ........ r60022 | brett.cannon | 2008-01-17 19:45:10 +0100 (Thu, 17 Jan 2008) | 2 lines Fix a function pointer declaration to silence the compiler. ........ r60024 | raymond.hettinger | 2008-01-17 20:31:38 +0100 (Thu, 17 Jan 2008) | 1 line Issue #1861: Add read-only attribute listing upcoming events in the order they will be run. ........ r60025 | andrew.kuchling | 2008-01-17 20:49:24 +0100 (Thu, 17 Jan 2008) | 1 line Correction from Jordan Lewis: halfdelay() uses tenths of a second, not milliseconds ........ r60026 | raymond.hettinger | 2008-01-17 23:27:49 +0100 (Thu, 17 Jan 2008) | 1 line Add advice on choosing between scheduler and threading.Timer(). ........ r60028 | christian.heimes | 2008-01-18 00:01:44 +0100 (Fri, 18 Jan 2008) | 2 lines Updated new property syntax. An elaborate example for subclassing and the getter was missing. Added comment about VS 2008 and PGO builds. ........ r60029 | raymond.hettinger | 2008-01-18 00:32:01 +0100 (Fri, 18 Jan 2008) | 1 line Fix-up Timer() example. ........ r60030 | raymond.hettinger | 2008-01-18 00:56:56 +0100 (Fri, 18 Jan 2008) | 1 line Fix markup ........ r60031 | raymond.hettinger | 2008-01-18 01:10:42 +0100 (Fri, 18 Jan 2008) | 1 line clearcache() needs to remove the dict as well as clear it. ........ r60033 | andrew.kuchling | 2008-01-18 03:26:16 +0100 (Fri, 18 Jan 2008) | 1 line Bump verson ........ r60034 | andrew.kuchling | 2008-01-18 03:42:52 +0100 (Fri, 18 Jan 2008) | 1 line Typo fix ........ r60035 | christian.heimes | 2008-01-18 08:30:20 +0100 (Fri, 18 Jan 2008) | 3 lines Coverity issue CID #197 var_decl: Declared variable "stm" without initializer ninit_use_in_call: Using uninitialized value "stm" (field "stm".tm_zone uninitialized) in call to function "mktime" ........ r60036 | christian.heimes | 2008-01-18 08:45:30 +0100 (Fri, 18 Jan 2008) | 11 lines Coverity issue CID #167 Event alloc_fn: Called allocation function "metacompile" [model] Event var_assign: Assigned variable "gr" to storage returned from "metacompile" gr = metacompile(n); Event pass_arg: Variable "gr" not freed or pointed-to in function "maketables" [model] g = maketables(gr); translatelabels(g); addfirstsets(g); Event leaked_storage: Returned without freeing storage "gr" return g; ........ r60038 | christian.heimes | 2008-01-18 09:04:57 +0100 (Fri, 18 Jan 2008) | 3 lines Coverity issue CID #182 size_error: Allocating 1 bytes to pointer "children", which needs at least 4 bytes ........ r60041 | christian.heimes | 2008-01-18 09:47:59 +0100 (Fri, 18 Jan 2008) | 4 lines Coverity issue CID #169 local_ptr_assign_local: Assigning address of stack variable "namebuf" to pointer "filename" out_of_scope: Variable "namebuf" goes out of scope use_invalid: Used "filename" pointing to out-of-scope variable "namebuf" ........ r60042 | christian.heimes | 2008-01-18 09:53:45 +0100 (Fri, 18 Jan 2008) | 2 lines Coverity CID #168 leaked_storage: Returned without freeing storage "fp" ........
2008-01-18 05:56:22 -04:00
stm.tm_sec = (dostime & 0x1f) * 2;
stm.tm_min = (dostime >> 5) & 0x3f;
stm.tm_hour = (dostime >> 11) & 0x1f;
stm.tm_mday = dosdate & 0x1f;
stm.tm_mon = ((dosdate >> 5) & 0x0f) - 1;
stm.tm_year = ((dosdate >> 9) & 0x7f) + 80;
stm.tm_isdst = -1; /* wday/yday is ignored */
return mktime(&stm);
}
/* Given a path to a .pyc or .pyo file in the archive, return the
modifictaion time of the matching .py file, or 0 if no source
is available. */
static time_t
get_mtime_of_source(ZipImporter *self, char *path)
{
PyObject *toc_entry;
time_t mtime = 0;
Py_ssize_t lastchar = strlen(path) - 1;
char savechar = path[lastchar];
path[lastchar] = '\0'; /* strip 'c' or 'o' from *.py[co] */
toc_entry = PyDict_GetItemString(self->files, path);
if (toc_entry != NULL && PyTuple_Check(toc_entry) &&
PyTuple_Size(toc_entry) == 8) {
/* fetch the time stamp of the .py file for comparison
with an embedded pyc time stamp */
int time, date;
time = PyLong_AsLong(PyTuple_GetItem(toc_entry, 5));
date = PyLong_AsLong(PyTuple_GetItem(toc_entry, 6));
mtime = parse_dostime(time, date);
}
path[lastchar] = savechar;
return mtime;
}
/* Return the code object for the module named by 'fullname' from the
Zip archive as a new reference. */
static PyObject *
get_code_from_data(ZipImporter *self, int ispackage, int isbytecode,
time_t mtime, PyObject *toc_entry)
{
PyObject *data, *code;
char *modpath;
char *archive = _PyUnicode_AsString(self->archive);
if (archive == NULL)
return NULL;
data = get_data(archive, toc_entry);
if (data == NULL)
return NULL;
modpath = _PyUnicode_AsString(PyTuple_GetItem(toc_entry, 0));
if (isbytecode) {
code = unmarshal_code(modpath, data, mtime);
}
else {
code = compile_source(modpath, data);
}
Py_DECREF(data);
return code;
}
/* Get the code object assoiciated with the module specified by
'fullname'. */
static PyObject *
get_module_code(ZipImporter *self, char *fullname,
int *p_ispackage, char **p_modpath)
{
PyObject *toc_entry;
char *subname, path[MAXPATHLEN + 1];
int len;
struct st_zip_searchorder *zso;
subname = get_subname(fullname);
len = make_filename(_PyUnicode_AsString(self->prefix), subname, path);
if (len < 0)
return NULL;
for (zso = zip_searchorder; *zso->suffix; zso++) {
PyObject *code = NULL;
strcpy(path + len, zso->suffix);
if (Py_VerboseFlag > 1)
PySys_WriteStderr("# trying %s%c%s\n",
_PyUnicode_AsString(self->archive),
(int)SEP, path);
toc_entry = PyDict_GetItemString(self->files, path);
if (toc_entry != NULL) {
time_t mtime = 0;
int ispackage = zso->type & IS_PACKAGE;
int isbytecode = zso->type & IS_BYTECODE;
if (isbytecode)
mtime = get_mtime_of_source(self, path);
if (p_ispackage != NULL)
*p_ispackage = ispackage;
code = get_code_from_data(self, ispackage,
isbytecode, mtime,
toc_entry);
if (code == Py_None) {
/* bad magic number or non-matching mtime
in byte code, try next */
Py_DECREF(code);
continue;
}
if (code != NULL && p_modpath != NULL)
*p_modpath = _PyUnicode_AsString(
PyTuple_GetItem(toc_entry, 0));
return code;
}
}
PyErr_Format(ZipImportError, "can't find module '%.200s'", fullname);
return NULL;
}
/* Module init */
PyDoc_STRVAR(zipimport_doc,
"zipimport provides support for importing Python modules from Zip archives.\n\
\n\
This module exports three objects:\n\
- zipimporter: a class; its constructor takes a path to a Zip archive.\n\
- ZipImportError: exception raised by zipimporter objects. It's a\n\
subclass of ImportError, so it can be caught as ImportError, too.\n\
- _zip_directory_cache: a dict, mapping archive paths to zip directory\n\
info dicts, as used in zipimporter._files.\n\
\n\
It is usually not needed to use the zipimport module explicitly; it is\n\
used by the builtin import mechanism for sys.path items that are paths\n\
to Zip archives.");
static struct PyModuleDef zipimportmodule = {
PyModuleDef_HEAD_INIT,
"zipimport",
zipimport_doc,
-1,
NULL,
NULL,
NULL,
NULL,
NULL
};
PyMODINIT_FUNC
PyInit_zipimport(void)
{
PyObject *mod;
if (PyType_Ready(&ZipImporter_Type) < 0)
return NULL;
/* Correct directory separator */
zip_searchorder[0].suffix[0] = SEP;
zip_searchorder[1].suffix[0] = SEP;
zip_searchorder[2].suffix[0] = SEP;
if (Py_OptimizeFlag) {
/* Reverse *.pyc and *.pyo */
struct st_zip_searchorder tmp;
tmp = zip_searchorder[0];
zip_searchorder[0] = zip_searchorder[1];
zip_searchorder[1] = tmp;
tmp = zip_searchorder[3];
zip_searchorder[3] = zip_searchorder[4];
zip_searchorder[4] = tmp;
}
mod = PyModule_Create(&zipimportmodule);
if (mod == NULL)
return NULL;
ZipImportError = PyErr_NewException("zipimport.ZipImportError",
PyExc_ImportError, NULL);
if (ZipImportError == NULL)
return NULL;
Py_INCREF(ZipImportError);
if (PyModule_AddObject(mod, "ZipImportError",
ZipImportError) < 0)
return NULL;
Py_INCREF(&ZipImporter_Type);
if (PyModule_AddObject(mod, "zipimporter",
(PyObject *)&ZipImporter_Type) < 0)
return NULL;
zip_directory_cache = PyDict_New();
if (zip_directory_cache == NULL)
return NULL;
Py_INCREF(zip_directory_cache);
if (PyModule_AddObject(mod, "_zip_directory_cache",
zip_directory_cache) < 0)
return NULL;
return mod;
}