| 1 | import urllib.parse |
| 2 | from . import proto |
| 3 | |
| 4 | __all__ = ["formdata"] |
| 5 | |
| 6 | def formparse(req): |
| 7 | buf = {} |
| 8 | buf.update(urllib.parse.parse_qsl(req.query)) |
| 9 | if req.ihead.get("Content-Type") == "application/x-www-form-urlencoded": |
| 10 | try: |
| 11 | rbody = req.input.read(2 ** 20) |
| 12 | except IOError as exc: |
| 13 | return exc |
| 14 | if len(rbody) >= 2 ** 20: |
| 15 | return ValueError("x-www-form-urlencoded data is absurdly long") |
| 16 | buf.update(urllib.parse.parse_qsl(rbody.decode("latin1"))) |
| 17 | return buf |
| 18 | |
| 19 | class badmultipart(Exception): |
| 20 | pass |
| 21 | |
| 22 | class formpart(object): |
| 23 | def __init__(self, form): |
| 24 | self.form = form |
| 25 | self.buf = b"" |
| 26 | self.eof = False |
| 27 | self.head = {} |
| 28 | |
| 29 | def parsehead(self): |
| 30 | pass |
| 31 | |
| 32 | def fillbuf(self, sz): |
| 33 | req = self.form.req |
| 34 | mboundary = b"\r\n--" + self.form.boundary + b"\r\n" |
| 35 | lboundary = b"\r\n--" + self.form.boundary + b"--\r\n" |
| 36 | while not self.eof: |
| 37 | p = self.form.buf.find(mboundary) |
| 38 | if p >= 0: |
| 39 | self.buf += self.form.buf[:p] |
| 40 | self.form.buf = self.form.buf[p + len(mboundary):] |
| 41 | self.eof = True |
| 42 | break |
| 43 | p = self.form.buf.find(lboundary) |
| 44 | if p >= 0: |
| 45 | self.buf += self.form.buf[:p] |
| 46 | self.form.buf = self.form.buf[p + len(lboundary):] |
| 47 | self.eof = True |
| 48 | self.form.eof = True |
| 49 | break |
| 50 | self.buf += self.form.buf[:-len(lboundary)] |
| 51 | self.form.buf = self.form.buf[-len(lboundary):] |
| 52 | if sz >= 0 and len(self.buf) >= sz: |
| 53 | break |
| 54 | while len(self.form.buf) <= len(lboundary): |
| 55 | ret = req.input.read(8192) |
| 56 | if ret == b"": |
| 57 | raise badmultipart("Missing last multipart boundary") |
| 58 | self.form.buf += ret |
| 59 | |
| 60 | def read(self, limit=-1): |
| 61 | self.fillbuf(limit) |
| 62 | if limit >= 0: |
| 63 | ret = self.buf[:limit] |
| 64 | self.buf = self.buf[limit:] |
| 65 | else: |
| 66 | ret = self.buf |
| 67 | self.buf = b"" |
| 68 | return ret |
| 69 | |
| 70 | def readline(self, limit=-1): |
| 71 | last = 0 |
| 72 | while True: |
| 73 | p = self.buf.find(b'\n', last) |
| 74 | if p < 0: |
| 75 | if self.eof: |
| 76 | ret = self.buf |
| 77 | self.buf = b"" |
| 78 | return ret |
| 79 | last = len(self.buf) |
| 80 | self.fillbuf(last + 128) |
| 81 | else: |
| 82 | ret = self.buf[:p + 1] |
| 83 | self.buf = self.buf[p + 1:] |
| 84 | return ret |
| 85 | |
| 86 | def close(self): |
| 87 | while True: |
| 88 | if self.read(8192) == b"": |
| 89 | break |
| 90 | |
| 91 | def __enter__(self): |
| 92 | return self |
| 93 | |
| 94 | def __exit__(self, *excinfo): |
| 95 | self.close() |
| 96 | return False |
| 97 | |
| 98 | def parsehead(self, charset): |
| 99 | def headline(): |
| 100 | ln = self.readline(256) |
| 101 | if ln[-1] != ord(b'\n'): |
| 102 | raise badmultipart("Too long header line in part") |
| 103 | try: |
| 104 | return ln.decode(charset).rstrip() |
| 105 | except UnicodeError: |
| 106 | raise badmultipart("Form part header is not in assumed charset") |
| 107 | |
| 108 | ln = headline() |
| 109 | while True: |
| 110 | if ln == "": |
| 111 | break |
| 112 | buf = ln |
| 113 | while True: |
| 114 | ln = headline() |
| 115 | if not ln[1:].isspace(): |
| 116 | break |
| 117 | buf += ln.lstrip() |
| 118 | p = buf.find(':') |
| 119 | if p < 0: |
| 120 | raise badmultipart("Malformed multipart header line") |
| 121 | self.head[buf[:p].strip().lower()] = buf[p + 1:].lstrip() |
| 122 | |
| 123 | val, par = proto.pmimehead(self.head.get("content-disposition", "")) |
| 124 | if val != "form-data": |
| 125 | raise badmultipart("Unexpected Content-Disposition in form part: %r" % val) |
| 126 | if not "name" in par: |
| 127 | raise badmultipart("Missing name in form part") |
| 128 | self.name = par["name"] |
| 129 | self.filename = par.get("filename") |
| 130 | val, par = proto.pmimehead(self.head.get("content-type", "")) |
| 131 | self.ctype = val |
| 132 | self.charset = par.get("charset") |
| 133 | encoding = self.head.get("content-transfer-encoding", "binary") |
| 134 | if encoding != "binary": |
| 135 | raise badmultipart("Form part uses unexpected transfer encoding: %r" % encoding) |
| 136 | |
| 137 | class multipart(object): |
| 138 | def __init__(self, req, charset): |
| 139 | val, par = proto.pmimehead(req.ihead.get("Content-Type", "")) |
| 140 | if req.method != "POST" or val != "multipart/form-data": |
| 141 | raise badmultipart("Request is not a multipart form") |
| 142 | if "boundary" not in par: |
| 143 | raise badmultipart("Multipart form lacks boundary") |
| 144 | try: |
| 145 | self.boundary = par["boundary"].encode("us-ascii") |
| 146 | except UnicodeError: |
| 147 | raise badmultipart("Multipart boundary must be ASCII string") |
| 148 | self.req = req |
| 149 | self.buf = b"\r\n" |
| 150 | self.eof = False |
| 151 | self.headcs = charset |
| 152 | self.lastpart = formpart(self) |
| 153 | self.lastpart.close() |
| 154 | |
| 155 | def __iter__(self): |
| 156 | return self |
| 157 | |
| 158 | def __next__(self): |
| 159 | if not self.lastpart.eof: |
| 160 | raise RuntimeError("All form parts must be read entirely") |
| 161 | if self.eof: |
| 162 | raise StopIteration() |
| 163 | self.lastpart = formpart(self) |
| 164 | self.lastpart.parsehead(self.headcs) |
| 165 | return self.lastpart |
| 166 | |
| 167 | def formdata(req, onerror=Exception): |
| 168 | data = req.item(formparse) |
| 169 | if isinstance(data, Exception): |
| 170 | if onerror is Exception: |
| 171 | raise data |
| 172 | return onerror |
| 173 | return data |