mirror of
https://github.com/Hopiu/linkchecker.git
synced 2026-04-17 21:01:00 +00:00
added and use ftpparse module
git-svn-id: https://linkchecker.svn.sourceforge.net/svnroot/linkchecker/trunk/linkchecker@2218 e7d03fd6-7b0d-0410-9947-9c21f3af8025
This commit is contained in:
calvin
parent
6840808fa9
commit
3e79dd7e09
8 changed files with 674 additions and 21 deletions
|
|
@ -24,6 +24,11 @@
|
|||
Changed: linkcheck/logger/__init__.py, linkcheck/configuration.py,
|
||||
linkchecker
|
||||
|
||||
* Added ftpparse module to parse FTP LIST output lines.
|
||||
Type: feature
|
||||
Added linkcheck/ftpparse/*
|
||||
Changed: setup.py, linkcheck/checker/ftpurl.py
|
||||
|
||||
2.2 "Cube" (released 25.01.2005)
|
||||
|
||||
* CSV log format changes:
|
||||
|
|
|
|||
1
Makefile
1
Makefile
|
|
@ -55,6 +55,7 @@ localbuild:
|
|||
$(MAKE) -C linkcheck/HtmlParser
|
||||
$(PYTHON) setup.py build
|
||||
cp -f build/lib.linux-i686-$(PYVER)/linkcheck/HtmlParser/htmlsax.so linkcheck/HtmlParser
|
||||
cp -f build/lib.linux-i686-$(PYVER)/linkcheck/ftpparse/_ftpparse.so linkcheck/ftpparse
|
||||
|
||||
# produce the .deb Debian package
|
||||
deb_local: cleandeb
|
||||
|
|
|
|||
|
|
@ -27,7 +27,7 @@ import linkcheck
|
|||
import urlbase
|
||||
import proxysupport
|
||||
import httpurl
|
||||
|
||||
import linkcheck.ftpparse._ftpparse as ftpparse
|
||||
|
||||
class FtpUrl (urlbase.UrlBase, proxysupport.ProxySupport):
|
||||
"""
|
||||
|
|
@ -105,7 +105,7 @@ class FtpUrl (urlbase.UrlBase, proxysupport.ProxySupport):
|
|||
except EOFError, msg:
|
||||
msg = str(msg)
|
||||
raise linkcheck.LinkCheckerError(
|
||||
_("Remote host has closed connection: %s") % msg)
|
||||
_("Remote host has closed connection: %r") % msg)
|
||||
if not self.url_connection.getwelcome():
|
||||
self.close_connection()
|
||||
raise linkcheck.LinkCheckerError(
|
||||
|
|
@ -154,27 +154,19 @@ class FtpUrl (urlbase.UrlBase, proxysupport.ProxySupport):
|
|||
Get list of filenames in directory. Subdirectories have an
|
||||
ending slash.
|
||||
"""
|
||||
# Rudimentary LIST output parsing. An entry is assumed to have
|
||||
# the following form:
|
||||
# drwxr-xr-x 8 root wheel 1024 Jan 3 1994 foo
|
||||
# Symbolic links are assumed to have the following form:
|
||||
# drwxr-xr-x 8 root wheel 1024 Jan 3 1994 foo -> bar
|
||||
files = []
|
||||
def add_entry (line):
|
||||
linkcheck.log.debug(linkcheck.LOG_CHECK, "Directory entry %r",
|
||||
line)
|
||||
parts = line.split()
|
||||
if len(parts) >= 8:
|
||||
if parts[-2] == "->":
|
||||
# symbolic link
|
||||
fname = parts[-3]
|
||||
else:
|
||||
fname = parts[-1]
|
||||
if fname not in (".", ".."):
|
||||
if line.startswith("d"):
|
||||
# a directory
|
||||
fname += "/"
|
||||
files.append(fname)
|
||||
try:
|
||||
fpo = ftpparse.parse(line)
|
||||
name = fpo.name
|
||||
if fpo.trycwd:
|
||||
name += "/"
|
||||
if fpo.trycwd or fpo.tryretr:
|
||||
files.append(name)
|
||||
except (ValueError, AttributeError), msg:
|
||||
print "XXX", msg
|
||||
self.url_connection.dir(add_entry)
|
||||
return files
|
||||
|
||||
|
|
|
|||
7
linkcheck/ftpparse/__init__.py
Normal file
7
linkcheck/ftpparse/__init__.py
Normal file
|
|
@ -0,0 +1,7 @@
|
|||
# -*- coding: iso-8859-1 -*-
|
||||
"""
|
||||
a simple wrapper for Dan Bernstein's ftpparse library. for more
|
||||
information on the library, see:
|
||||
|
||||
http://cr.yp.to/ftpparse.html
|
||||
"""
|
||||
138
linkcheck/ftpparse/_ftpparse.c
Normal file
138
linkcheck/ftpparse/_ftpparse.c
Normal file
|
|
@ -0,0 +1,138 @@
|
|||
/*
|
||||
* _ftpparse wrapper
|
||||
*
|
||||
* History:
|
||||
* 2001-01-19 fl created
|
||||
*
|
||||
* Copyright (c) 2002 by Fredrik Lundh.
|
||||
*
|
||||
* This wrapper module can be used freely, also for commercial
|
||||
* purposes. See ftpparse.c for more information on the under-
|
||||
* lying ftpparse library.
|
||||
*/
|
||||
|
||||
#include "Python.h"
|
||||
#include "ftpparse.h"
|
||||
|
||||
/* the fpObject type is a simple wrapper for ftpparse structs */
|
||||
|
||||
typedef struct {
|
||||
PyObject_HEAD
|
||||
PyObject* string;
|
||||
struct ftpparse fp;
|
||||
} ftpparseObject;
|
||||
|
||||
staticforward PyTypeObject ftpparse_Type;
|
||||
|
||||
static PyObject *
|
||||
ftpparse_parse(PyObject* self_, PyObject* args)
|
||||
{
|
||||
ftpparseObject* self;
|
||||
int ok;
|
||||
|
||||
PyObject* str;
|
||||
if (!PyArg_ParseTuple(args, "O!", &PyString_Type, &str))
|
||||
return NULL;
|
||||
|
||||
self = PyObject_NEW(ftpparseObject, &ftpparse_Type);
|
||||
if (!self)
|
||||
return NULL;
|
||||
|
||||
ok = ftpparse(&self->fp, PyString_AS_STRING(str), PyString_GET_SIZE(str));
|
||||
|
||||
if (!ok) {
|
||||
PyObject_DEL(self);
|
||||
PyErr_SetString(PyExc_ValueError, "cannot find filename");
|
||||
return NULL;
|
||||
}
|
||||
|
||||
/* the ftpparse structure contains pointers to inside the source
|
||||
string, so we better keep a reference to it */
|
||||
Py_INCREF(str);
|
||||
self->string = str;
|
||||
|
||||
return (PyObject*) self;
|
||||
}
|
||||
|
||||
static void
|
||||
ftpparse_dealloc(ftpparseObject* self)
|
||||
{
|
||||
Py_XDECREF(self->string);
|
||||
PyMem_DEL(self);
|
||||
}
|
||||
|
||||
static PyObject*
|
||||
ftpparse_getattr(ftpparseObject* self, char* name)
|
||||
{
|
||||
/* get descriptor attribute */
|
||||
|
||||
if (strcmp(name, "name") == 0)
|
||||
/* filename */
|
||||
return Py_BuildValue("s#", self->fp.name, self->fp.namelen);
|
||||
|
||||
if (strcmp(name, "size") == 0) {
|
||||
/* size, in bytes (use sizetype to check if text/data) */
|
||||
if (self->fp.sizetype == FTPPARSE_SIZE_UNKNOWN)
|
||||
goto unknown;
|
||||
return Py_BuildValue("l", self->fp.size);
|
||||
}
|
||||
|
||||
if (strcmp(name, "mtime") == 0) {
|
||||
/* modification time (use mtimetype to check local/remote) */
|
||||
if (self->fp.mtimetype == FTPPARSE_MTIME_UNKNOWN)
|
||||
goto unknown;
|
||||
return Py_BuildValue("l", self->fp.mtime);
|
||||
}
|
||||
|
||||
if (strcmp(name, "id") == 0) {
|
||||
/* unique identifier */
|
||||
if (self->fp.idtype == FTPPARSE_ID_UNKNOWN)
|
||||
goto unknown;
|
||||
return Py_BuildValue("s#", self->fp.id, self->fp.idlen);
|
||||
}
|
||||
|
||||
if (strcmp(name, "tryretr") == 0) {
|
||||
PyObject* flag = (self->fp.flagtryretr) ? Py_True : Py_False;
|
||||
Py_INCREF(flag);
|
||||
return flag;
|
||||
}
|
||||
|
||||
if (strcmp(name, "trycwd") == 0) {
|
||||
PyObject* flag = (self->fp.flagtrycwd) ? Py_True : Py_False;
|
||||
Py_INCREF(flag);
|
||||
return flag;
|
||||
}
|
||||
|
||||
/* FIXME: what about sizetype, mtimetype, idtype? return as
|
||||
integers or strings? or use a flag tuple instead of individual
|
||||
flags? */
|
||||
|
||||
PyErr_SetString(PyExc_AttributeError, name);
|
||||
return NULL;
|
||||
|
||||
unknown:
|
||||
Py_INCREF(Py_None);
|
||||
return Py_None;
|
||||
}
|
||||
|
||||
static PyTypeObject ftpparse_Type = {
|
||||
PyObject_HEAD_INIT(NULL)
|
||||
0, "ftpparse", sizeof(ftpparseObject), 0,
|
||||
/* methods */
|
||||
(destructor) ftpparse_dealloc, /*tp_dealloc*/
|
||||
0, /*tp_print*/
|
||||
(getattrfunc) ftpparse_getattr, /*tp_getattr*/
|
||||
};
|
||||
|
||||
static PyMethodDef _functions[] = {
|
||||
{"parse", ftpparse_parse, METH_VARARGS},
|
||||
{NULL, NULL}
|
||||
};
|
||||
|
||||
PyMODINIT_FUNC
|
||||
init_ftpparse(void)
|
||||
{
|
||||
ftpparse_Type.ob_type = &PyType_Type;
|
||||
|
||||
Py_InitModule("_ftpparse", _functions);
|
||||
}
|
||||
448
linkcheck/ftpparse/ftpparse.c
Normal file
448
linkcheck/ftpparse/ftpparse.c
Normal file
|
|
@ -0,0 +1,448 @@
|
|||
/* ftpparse.c, ftpparse.h: library for parsing FTP LIST responses
|
||||
20001223
|
||||
D. J. Bernstein, djb@cr.yp.to
|
||||
http://cr.yp.to/ftpparse.html
|
||||
|
||||
Commercial use is fine, if you let me know what programs you're using this in.
|
||||
|
||||
Currently covered formats:
|
||||
EPLF.
|
||||
UNIX ls, with or without gid.
|
||||
Microsoft FTP Service.
|
||||
Windows NT FTP Server.
|
||||
VMS.
|
||||
WFTPD.
|
||||
NetPresenz (Mac).
|
||||
NetWare.
|
||||
MSDOS.
|
||||
|
||||
Definitely not covered:
|
||||
Long VMS filenames, with information split across two lines.
|
||||
NCSA Telnet FTP server. Has LIST = NLST (and bad NLST for directories).
|
||||
*/
|
||||
|
||||
#include <time.h>
|
||||
#include "ftpparse.h"
|
||||
|
||||
static long totai(long year,long month,long mday)
|
||||
{
|
||||
long result;
|
||||
if (month >= 2) month -= 2;
|
||||
else { month += 10; --year; }
|
||||
result = (mday - 1) * 10 + 5 + 306 * month;
|
||||
result /= 10;
|
||||
if (result == 365) { year -= 3; result = 1460; }
|
||||
else result += 365 * (year % 4);
|
||||
year /= 4;
|
||||
result += 1461 * (year % 25);
|
||||
year /= 25;
|
||||
if (result == 36524) { year -= 3; result = 146096; }
|
||||
else { result += 36524 * (year % 4); }
|
||||
year /= 4;
|
||||
result += 146097 * (year - 5);
|
||||
result += 11017;
|
||||
return result * 86400;
|
||||
}
|
||||
|
||||
static int flagneedbase = 1;
|
||||
static time_t base; /* time() value on this OS at the beginning of 1970 TAI */
|
||||
static long now; /* current time */
|
||||
static int flagneedcurrentyear = 1;
|
||||
static long currentyear; /* approximation to current year */
|
||||
|
||||
static void initbase(void)
|
||||
{
|
||||
struct tm *t;
|
||||
if (!flagneedbase) return;
|
||||
|
||||
base = 0;
|
||||
t = gmtime(&base);
|
||||
base = -(totai(t->tm_year + 1900,t->tm_mon,t->tm_mday) + t->tm_hour * 3600 + t->tm_min * 60 + t->tm_sec);
|
||||
/* assumes the right time_t, counting seconds. */
|
||||
/* base may be slightly off if time_t counts non-leap seconds. */
|
||||
flagneedbase = 0;
|
||||
}
|
||||
|
||||
static void initnow(void)
|
||||
{
|
||||
long day;
|
||||
long year;
|
||||
|
||||
initbase();
|
||||
now = time((time_t *) 0) - base;
|
||||
|
||||
if (flagneedcurrentyear) {
|
||||
day = now / 86400;
|
||||
if ((now % 86400) < 0) --day;
|
||||
day -= 11017;
|
||||
year = 5 + day / 146097;
|
||||
day = day % 146097;
|
||||
if (day < 0) { day += 146097; --year; }
|
||||
year *= 4;
|
||||
if (day == 146096) { year += 3; day = 36524; }
|
||||
else { year += day / 36524; day %= 36524; }
|
||||
year *= 25;
|
||||
year += day / 1461;
|
||||
day %= 1461;
|
||||
year *= 4;
|
||||
if (day == 1460) { year += 3; day = 365; }
|
||||
else { year += day / 365; day %= 365; }
|
||||
day *= 10;
|
||||
if ((day + 5) / 306 >= 10) ++year;
|
||||
currentyear = year;
|
||||
flagneedcurrentyear = 0;
|
||||
}
|
||||
}
|
||||
|
||||
/* UNIX ls does not show the year for dates in the last six months. */
|
||||
/* So we have to guess the year. */
|
||||
/* Apparently NetWare uses ``twelve months'' instead of ``six months''; ugh. */
|
||||
/* Some versions of ls also fail to show the year for future dates. */
|
||||
static long guesstai(long month,long mday)
|
||||
{
|
||||
long year;
|
||||
long t;
|
||||
|
||||
initnow();
|
||||
|
||||
for (year = currentyear - 1;year < currentyear + 100;++year) {
|
||||
t = totai(year,month,mday);
|
||||
if (now - t < 350 * 86400)
|
||||
return t;
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
|
||||
static int check(char *buf,char *monthname)
|
||||
{
|
||||
if ((buf[0] != monthname[0]) && (buf[0] != monthname[0] - 32)) return 0;
|
||||
if ((buf[1] != monthname[1]) && (buf[1] != monthname[1] - 32)) return 0;
|
||||
if ((buf[2] != monthname[2]) && (buf[2] != monthname[2] - 32)) return 0;
|
||||
return 1;
|
||||
}
|
||||
|
||||
static char *months[12] = {
|
||||
"jan","feb","mar","apr","may","jun","jul","aug","sep","oct","nov","dec"
|
||||
} ;
|
||||
|
||||
static int getmonth(char *buf,int len)
|
||||
{
|
||||
int i;
|
||||
if (len == 3)
|
||||
for (i = 0;i < 12;++i)
|
||||
if (check(buf,months[i])) return i;
|
||||
return -1;
|
||||
}
|
||||
|
||||
static long getlong(char *buf,int len)
|
||||
{
|
||||
long u = 0;
|
||||
while (len-- > 0)
|
||||
u = u * 10 + (*buf++ - '0');
|
||||
return u;
|
||||
}
|
||||
|
||||
int ftpparse(struct ftpparse *fp,char *buf,int len)
|
||||
{
|
||||
int i;
|
||||
int j;
|
||||
int state;
|
||||
long size = 0;
|
||||
long year;
|
||||
long month = 0;
|
||||
long mday = 0;
|
||||
long hour;
|
||||
long minute;
|
||||
|
||||
fp->name = 0;
|
||||
fp->namelen = 0;
|
||||
fp->flagtrycwd = 0;
|
||||
fp->flagtryretr = 0;
|
||||
fp->sizetype = FTPPARSE_SIZE_UNKNOWN;
|
||||
fp->size = 0;
|
||||
fp->mtimetype = FTPPARSE_MTIME_UNKNOWN;
|
||||
fp->mtime = 0;
|
||||
fp->idtype = FTPPARSE_ID_UNKNOWN;
|
||||
fp->id = 0;
|
||||
fp->idlen = 0;
|
||||
|
||||
if (len < 2) /* an empty name in EPLF, with no info, could be 2 chars */
|
||||
return 0;
|
||||
|
||||
switch(*buf) {
|
||||
/* see http://pobox.com/~djb/proto/eplf.txt */
|
||||
/* "+i8388621.29609,m824255902,/,\tdev" */
|
||||
/* "+i8388621.44468,m839956783,r,s10376,\tRFCEPLF" */
|
||||
case '+':
|
||||
i = 1;
|
||||
for (j = 1;j < len;++j) {
|
||||
if (buf[j] == 9) {
|
||||
fp->name = buf + j + 1;
|
||||
fp->namelen = len - j - 1;
|
||||
return 1;
|
||||
}
|
||||
if (buf[j] == ',') {
|
||||
switch(buf[i]) {
|
||||
case '/':
|
||||
fp->flagtrycwd = 1;
|
||||
break;
|
||||
case 'r':
|
||||
fp->flagtryretr = 1;
|
||||
break;
|
||||
case 's':
|
||||
fp->sizetype = FTPPARSE_SIZE_BINARY;
|
||||
fp->size = getlong(buf + i + 1,j - i - 1);
|
||||
break;
|
||||
case 'm':
|
||||
fp->mtimetype = FTPPARSE_MTIME_LOCAL;
|
||||
initbase();
|
||||
fp->mtime = base + getlong(buf + i + 1,j - i - 1);
|
||||
break;
|
||||
case 'i':
|
||||
fp->idtype = FTPPARSE_ID_FULL;
|
||||
fp->id = buf + i + 1;
|
||||
fp->idlen = j - i - 1;
|
||||
}
|
||||
i = j + 1;
|
||||
}
|
||||
}
|
||||
return 0;
|
||||
|
||||
/* UNIX-style listing, without inum and without blocks */
|
||||
/* "-rw-r--r-- 1 root other 531 Jan 29 03:26 README" */
|
||||
/* "dr-xr-xr-x 2 root other 512 Apr 8 1994 etc" */
|
||||
/* "dr-xr-xr-x 2 root 512 Apr 8 1994 etc" */
|
||||
/* "lrwxrwxrwx 1 root other 7 Jan 25 00:17 bin -> usr/bin" */
|
||||
/* Also produced by Microsoft's FTP servers for Windows: */
|
||||
/* "---------- 1 owner group 1803128 Jul 10 10:18 ls-lR.Z" */
|
||||
/* "d--------- 1 owner group 0 May 9 19:45 Softlib" */
|
||||
/* Also WFTPD for MSDOS: */
|
||||
/* "-rwxrwxrwx 1 noone nogroup 322 Aug 19 1996 message.ftp" */
|
||||
/* Also NetWare: */
|
||||
/* "d [R----F--] supervisor 512 Jan 16 18:53 login" */
|
||||
/* "- [R----F--] rhesus 214059 Oct 20 15:27 cx.exe" */
|
||||
/* Also NetPresenz for the Mac: */
|
||||
/* "-------r-- 326 1391972 1392298 Nov 22 1995 MegaPhone.sit" */
|
||||
/* "drwxrwxr-x folder 2 May 10 1996 network" */
|
||||
case 'b':
|
||||
case 'c':
|
||||
case 'd':
|
||||
case 'l':
|
||||
case 'p':
|
||||
case 's':
|
||||
case '-':
|
||||
|
||||
if (*buf == 'd') fp->flagtrycwd = 1;
|
||||
if (*buf == '-') fp->flagtryretr = 1;
|
||||
if (*buf == 'l') fp->flagtrycwd = fp->flagtryretr = 1;
|
||||
|
||||
state = 1;
|
||||
i = 0;
|
||||
for (j = 1;j < len;++j)
|
||||
if ((buf[j] == ' ') && (buf[j - 1] != ' ')) {
|
||||
switch(state) {
|
||||
case 1: /* skipping perm */
|
||||
state = 2;
|
||||
break;
|
||||
case 2: /* skipping nlink */
|
||||
state = 3;
|
||||
if ((j - i == 6) && (buf[i] == 'f')) /* for NetPresenz */
|
||||
state = 4;
|
||||
break;
|
||||
case 3: /* skipping uid */
|
||||
state = 4;
|
||||
break;
|
||||
case 4: /* getting tentative size */
|
||||
size = getlong(buf + i,j - i);
|
||||
state = 5;
|
||||
break;
|
||||
case 5: /* searching for month, otherwise getting tentative size */
|
||||
month = getmonth(buf + i,j - i);
|
||||
if (month >= 0)
|
||||
state = 6;
|
||||
else
|
||||
size = getlong(buf + i,j - i);
|
||||
break;
|
||||
case 6: /* have size and month */
|
||||
mday = getlong(buf + i,j - i);
|
||||
state = 7;
|
||||
break;
|
||||
case 7: /* have size, month, mday */
|
||||
if ((j - i == 4) && (buf[i + 1] == ':')) {
|
||||
hour = getlong(buf + i,1);
|
||||
minute = getlong(buf + i + 2,2);
|
||||
fp->mtimetype = FTPPARSE_MTIME_REMOTEMINUTE;
|
||||
initbase();
|
||||
fp->mtime = base + guesstai(month,mday) + hour * 3600 + minute * 60;
|
||||
} else if ((j - i == 5) && (buf[i + 2] == ':')) {
|
||||
hour = getlong(buf + i,2);
|
||||
minute = getlong(buf + i + 3,2);
|
||||
fp->mtimetype = FTPPARSE_MTIME_REMOTEMINUTE;
|
||||
initbase();
|
||||
fp->mtime = base + guesstai(month,mday) + hour * 3600 + minute * 60;
|
||||
}
|
||||
else if (j - i >= 4) {
|
||||
year = getlong(buf + i,j - i);
|
||||
fp->mtimetype = FTPPARSE_MTIME_REMOTEDAY;
|
||||
initbase();
|
||||
fp->mtime = base + totai(year,month,mday);
|
||||
}
|
||||
else
|
||||
return 0;
|
||||
fp->name = buf + j + 1;
|
||||
fp->namelen = len - j - 1;
|
||||
state = 8;
|
||||
break;
|
||||
case 8: /* twiddling thumbs */
|
||||
break;
|
||||
}
|
||||
i = j + 1;
|
||||
while ((i < len) && (buf[i] == ' ')) ++i;
|
||||
}
|
||||
|
||||
if (state != 8)
|
||||
return 0;
|
||||
|
||||
fp->size = size;
|
||||
fp->sizetype = FTPPARSE_SIZE_BINARY;
|
||||
|
||||
if (*buf == 'l')
|
||||
for (i = 0;i + 3 < fp->namelen;++i)
|
||||
if (fp->name[i] == ' ')
|
||||
if (fp->name[i + 1] == '-')
|
||||
if (fp->name[i + 2] == '>')
|
||||
if (fp->name[i + 3] == ' ') {
|
||||
fp->namelen = i;
|
||||
break;
|
||||
}
|
||||
|
||||
/* eliminate extra NetWare spaces */
|
||||
if ((buf[1] == ' ') || (buf[1] == '['))
|
||||
if (fp->namelen > 3)
|
||||
if (fp->name[0] == ' ')
|
||||
if (fp->name[1] == ' ')
|
||||
if (fp->name[2] == ' ') {
|
||||
fp->name += 3;
|
||||
fp->namelen -= 3;
|
||||
}
|
||||
|
||||
return 1;
|
||||
}
|
||||
|
||||
/* MultiNet (some spaces removed from examples) */
|
||||
/* "00README.TXT;1 2 30-DEC-1996 17:44 [SYSTEM] (RWED,RWED,RE,RE)" */
|
||||
/* "CORE.DIR;1 1 8-SEP-1996 16:09 [SYSTEM] (RWE,RWE,RE,RE)" */
|
||||
/* and non-MutliNet VMS: */
|
||||
/* "CII-MANUAL.TEX;1 213/216 29-JAN-1996 03:33:12 [ANONYMOU,ANONYMOUS] (RWED,RWED,,)" */
|
||||
for (i = 0;i < len;++i)
|
||||
if (buf[i] == ';')
|
||||
break;
|
||||
if (i < len) {
|
||||
fp->name = buf;
|
||||
fp->namelen = i;
|
||||
if (i > 4)
|
||||
if (buf[i - 4] == '.')
|
||||
if (buf[i - 3] == 'D')
|
||||
if (buf[i - 2] == 'I')
|
||||
if (buf[i - 1] == 'R') {
|
||||
fp->namelen -= 4;
|
||||
fp->flagtrycwd = 1;
|
||||
}
|
||||
if (!fp->flagtrycwd)
|
||||
fp->flagtryretr = 1;
|
||||
while (buf[i] != ' ') if (++i == len) return 0;
|
||||
while (buf[i] == ' ') if (++i == len) return 0;
|
||||
while (buf[i] != ' ') if (++i == len) return 0;
|
||||
while (buf[i] == ' ') if (++i == len) return 0;
|
||||
j = i;
|
||||
while (buf[j] != '-') if (++j == len) return 0;
|
||||
mday = getlong(buf + i,j - i);
|
||||
while (buf[j] == '-') if (++j == len) return 0;
|
||||
i = j;
|
||||
while (buf[j] != '-') if (++j == len) return 0;
|
||||
month = getmonth(buf + i,j - i);
|
||||
if (month < 0) return 0;
|
||||
while (buf[j] == '-') if (++j == len) return 0;
|
||||
i = j;
|
||||
while (buf[j] != ' ') if (++j == len) return 0;
|
||||
year = getlong(buf + i,j - i);
|
||||
while (buf[j] == ' ') if (++j == len) return 0;
|
||||
i = j;
|
||||
while (buf[j] != ':') if (++j == len) return 0;
|
||||
hour = getlong(buf + i,j - i);
|
||||
while (buf[j] == ':') if (++j == len) return 0;
|
||||
i = j;
|
||||
while ((buf[j] != ':') && (buf[j] != ' ')) if (++j == len) return 0;
|
||||
minute = getlong(buf + i,j - i);
|
||||
|
||||
fp->mtimetype = FTPPARSE_MTIME_REMOTEMINUTE;
|
||||
initbase();
|
||||
fp->mtime = base + totai(year,month,mday) + hour * 3600 + minute * 60;
|
||||
|
||||
return 1;
|
||||
}
|
||||
|
||||
/* MSDOS format */
|
||||
/* 04-27-00 09:09PM <DIR> licensed */
|
||||
/* 07-18-00 10:16AM <DIR> pub */
|
||||
/* 04-14-00 03:47PM 589 readme.htm */
|
||||
if ((*buf >= '0') && (*buf <= '9')) {
|
||||
i = 0;
|
||||
j = 0;
|
||||
while (buf[j] != '-') if (++j == len) return 0;
|
||||
month = getlong(buf + i,j - i) - 1;
|
||||
while (buf[j] == '-') if (++j == len) return 0;
|
||||
i = j;
|
||||
while (buf[j] != '-') if (++j == len) return 0;
|
||||
mday = getlong(buf + i,j - i);
|
||||
while (buf[j] == '-') if (++j == len) return 0;
|
||||
i = j;
|
||||
while (buf[j] != ' ') if (++j == len) return 0;
|
||||
year = getlong(buf + i,j - i);
|
||||
if (year < 50) year += 2000;
|
||||
if (year < 1000) year += 1900;
|
||||
while (buf[j] == ' ') if (++j == len) return 0;
|
||||
i = j;
|
||||
while (buf[j] != ':') if (++j == len) return 0;
|
||||
hour = getlong(buf + i,j - i);
|
||||
while (buf[j] == ':') if (++j == len) return 0;
|
||||
i = j;
|
||||
while ((buf[j] != 'A') && (buf[j] != 'P')) if (++j == len) return 0;
|
||||
minute = getlong(buf + i,j - i);
|
||||
if (hour == 12) hour = 0;
|
||||
if (buf[j] == 'A') if (++j == len) return 0;
|
||||
if (buf[j] == 'P') { hour += 12; if (++j == len) return 0; }
|
||||
if (buf[j] == 'M') if (++j == len) return 0;
|
||||
|
||||
while (buf[j] == ' ') if (++j == len) return 0;
|
||||
if (buf[j] == '<') {
|
||||
fp->flagtrycwd = 1;
|
||||
while (buf[j] != ' ') if (++j == len) return 0;
|
||||
}
|
||||
else {
|
||||
i = j;
|
||||
while (buf[j] != ' ') if (++j == len) return 0;
|
||||
fp->size = getlong(buf + i,j - i);
|
||||
fp->sizetype = FTPPARSE_SIZE_BINARY;
|
||||
fp->flagtryretr = 1;
|
||||
}
|
||||
while (buf[j] == ' ') if (++j == len) return 0;
|
||||
|
||||
fp->name = buf + j;
|
||||
fp->namelen = len - j;
|
||||
|
||||
fp->mtimetype = FTPPARSE_MTIME_REMOTEMINUTE;
|
||||
initbase();
|
||||
fp->mtime = base + totai(year,month,mday) + hour * 3600 + minute * 60;
|
||||
|
||||
return 1;
|
||||
}
|
||||
|
||||
/* Some useless lines, safely ignored: */
|
||||
/* "Total of 11 Files, 10966 Blocks." (VMS) */
|
||||
/* "total 14786" (UNIX) */
|
||||
/* "DISK$ANONFTP:[ANONYMOUS]" (VMS) */
|
||||
/* "Directory DISK$PCSA:[ANONYM]" (VMS) */
|
||||
|
||||
return 0;
|
||||
}
|
||||
51
linkcheck/ftpparse/ftpparse.h
Normal file
51
linkcheck/ftpparse/ftpparse.h
Normal file
|
|
@ -0,0 +1,51 @@
|
|||
#ifndef FTPPARSE_H
|
||||
#define FTPPARSE_H
|
||||
|
||||
/*
|
||||
ftpparse(&fp,buf,len) tries to parse one line of LIST output.
|
||||
|
||||
The line is an array of len characters stored in buf.
|
||||
It should not include the terminating CR LF; so buf[len] is typically CR.
|
||||
|
||||
If ftpparse() can't find a filename, it returns 0.
|
||||
|
||||
If ftpparse() can find a filename, it fills in fp and returns 1.
|
||||
fp is a struct ftpparse, defined below.
|
||||
The name is an array of fp.namelen characters stored in fp.name;
|
||||
fp.name points somewhere within buf.
|
||||
*/
|
||||
|
||||
struct ftpparse {
|
||||
char *name; /* not necessarily 0-terminated */
|
||||
int namelen;
|
||||
int flagtrycwd; /* 0 if cwd is definitely pointless, 1 otherwise */
|
||||
int flagtryretr; /* 0 if retr is definitely pointless, 1 otherwise */
|
||||
int sizetype;
|
||||
long size; /* number of octets */
|
||||
int mtimetype;
|
||||
time_t mtime; /* modification time */
|
||||
int idtype;
|
||||
char *id; /* not necessarily 0-terminated */
|
||||
int idlen;
|
||||
} ;
|
||||
|
||||
#define FTPPARSE_SIZE_UNKNOWN 0
|
||||
#define FTPPARSE_SIZE_BINARY 1 /* size is the number of octets in TYPE I */
|
||||
#define FTPPARSE_SIZE_ASCII 2 /* size is the number of octets in TYPE A */
|
||||
|
||||
#define FTPPARSE_MTIME_UNKNOWN 0
|
||||
#define FTPPARSE_MTIME_LOCAL 1 /* time is correct */
|
||||
#define FTPPARSE_MTIME_REMOTEMINUTE 2 /* time zone and secs are unknown */
|
||||
#define FTPPARSE_MTIME_REMOTEDAY 3 /* time zone and time of day are unknown */
|
||||
/*
|
||||
When a time zone is unknown, it is assumed to be GMT. You may want
|
||||
to use localtime() for LOCAL times, along with an indication that the
|
||||
time is correct in the local time zone, and gmtime() for REMOTE* times.
|
||||
*/
|
||||
|
||||
#define FTPPARSE_ID_UNKNOWN 0
|
||||
#define FTPPARSE_ID_FULL 1 /* unique identifier for files on this FTP server */
|
||||
|
||||
extern int ftpparse(struct ftpparse *,char *,int);
|
||||
|
||||
#endif
|
||||
15
setup.py
15
setup.py
|
|
@ -99,7 +99,7 @@ class MyInstall (install, object):
|
|||
outs.append(self.distribution.get_conf_filename(self.install_lib))
|
||||
return outs
|
||||
|
||||
# compatibility bugfix for Python << 2.5
|
||||
# compatibility bugfix for Python << 2.5, << 2.4.1, << 2.3.5
|
||||
def dump_dirs (self, msg):
|
||||
if DEBUG:
|
||||
from distutils.fancy_getopt import longopt_xlate
|
||||
|
|
@ -316,7 +316,18 @@ o a (Fast)CGI web interface (requires HTTP server)
|
|||
define_macros = define_macros,
|
||||
include_dirs = include_dirs + \
|
||||
[normpath("linkcheck/HtmlParser")],
|
||||
)],
|
||||
),
|
||||
Extension("linkcheck.ftpparse._ftpparse",
|
||||
["linkcheck/ftpparse/_ftpparse.c",
|
||||
"linkcheck/ftpparse/ftpparse.c"],
|
||||
extra_compile_args = extra_compile_args,
|
||||
library_dirs = library_dirs,
|
||||
libraries = libraries,
|
||||
define_macros = define_macros,
|
||||
include_dirs = include_dirs + \
|
||||
[normpath("linkcheck/ftpparse")],
|
||||
),
|
||||
],
|
||||
scripts = scripts,
|
||||
data_files = data_files,
|
||||
classifiers = [
|
||||
|
|
|
|||
Loading…
Reference in a new issue