Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1 | #!/usr/local/bin/python |
Guido van Rossum | 1c9daa8 | 1995-09-18 21:52:37 +0000 | [diff] [blame] | 2 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 3 | """Support module for CGI (Common Gateway Interface) scripts. |
Guido van Rossum | 1c9daa8 | 1995-09-18 21:52:37 +0000 | [diff] [blame] | 4 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 5 | This module defines a number of utilities for use by CGI scripts |
| 6 | written in Python. |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 7 | |
| 8 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 9 | Introduction |
| 10 | ------------ |
| 11 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 12 | A CGI script is invoked by an HTTP server, usually to process user |
| 13 | input submitted through an HTML <FORM> or <ISINPUT> element. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 14 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 15 | Most often, CGI scripts live in the server's special cgi-bin |
| 16 | directory. The HTTP server places all sorts of information about the |
| 17 | request (such as the client's hostname, the requested URL, the query |
| 18 | string, and lots of other goodies) in the script's shell environment, |
| 19 | executes the script, and sends the script's output back to the client. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 20 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 21 | The script's input is connected to the client too, and sometimes the |
| 22 | form data is read this way; at other times the form data is passed via |
| 23 | the "query string" part of the URL. This module (cgi.py) is intended |
| 24 | to take care of the different cases and provide a simpler interface to |
| 25 | the Python script. It also provides a number of utilities that help |
| 26 | in debugging scripts, and the latest addition is support for file |
| 27 | uploads from a form (if your browser supports it -- Grail 0.3 and |
| 28 | Netscape 2.0 do). |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 29 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 30 | The output of a CGI script should consist of two sections, separated |
| 31 | by a blank line. The first section contains a number of headers, |
| 32 | telling the client what kind of data is following. Python code to |
| 33 | generate a minimal header section looks like this: |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 34 | |
Guido van Rossum | 243ddcd | 1996-03-07 06:33:07 +0000 | [diff] [blame] | 35 | print "Content-type: text/html" # HTML is following |
| 36 | print # blank line, end of headers |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 37 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 38 | The second section is usually HTML, which allows the client software |
| 39 | to display nicely formatted text with header, in-line images, etc. |
| 40 | Here's Python code that prints a simple piece of HTML: |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 41 | |
| 42 | print "<TITLE>CGI script output</TITLE>" |
| 43 | print "<H1>This is my first CGI script</H1>" |
| 44 | print "Hello, world!" |
| 45 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 46 | (It may not be fully legal HTML according to the letter of the |
| 47 | standard, but any browser will understand it.) |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 48 | |
| 49 | |
| 50 | Using the cgi module |
| 51 | -------------------- |
| 52 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 53 | Begin by writing "import cgi". Don't use "from cgi import *" -- the |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 54 | module defines all sorts of names for its own use or for backward |
| 55 | compatibility that you don't want in your namespace. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 56 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 57 | It's best to use the FieldStorage class. The other classes define in this |
| 58 | module are provided mostly for backward compatibility. Instantiate it |
| 59 | exactly once, without arguments. This reads the form contents from |
| 60 | standard input or the environment (depending on the value of various |
| 61 | environment variables set according to the CGI standard). Since it may |
| 62 | consume standard input, it should be instantiated only once. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 63 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 64 | The FieldStorage instance can be accessed as if it were a Python |
| 65 | dictionary. For instance, the following code (which assumes that the |
| 66 | Content-type header and blank line have already been printed) checks that |
| 67 | the fields "name" and "addr" are both set to a non-empty string: |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 68 | |
Guido van Rossum | 503e50b | 1996-05-28 22:57:20 +0000 | [diff] [blame] | 69 | form = cgi.FieldStorage() |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 70 | form_ok = 0 |
| 71 | if form.has_key("name") and form.has_key("addr"): |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 72 | if form["name"].value != "" and form["addr"].value != "": |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 73 | form_ok = 1 |
| 74 | if not form_ok: |
| 75 | print "<H1>Error</H1>" |
| 76 | print "Please fill in the name and addr fields." |
| 77 | return |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 78 | ...further form processing here... |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 79 | |
Guido van Rossum | 4032c2c | 1996-03-09 04:04:35 +0000 | [diff] [blame] | 80 | Here the fields, accessed through form[key], are themselves instances |
| 81 | of FieldStorage (or MiniFieldStorage, depending on the form encoding). |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 82 | |
Guido van Rossum | 4032c2c | 1996-03-09 04:04:35 +0000 | [diff] [blame] | 83 | If the submitted form data contains more than one field with the same |
| 84 | name, the object retrieved by form[key] is not a (Mini)FieldStorage |
| 85 | instance but a list of such instances. If you expect this possibility |
| 86 | (i.e., when your HTML form comtains multiple fields with the same |
| 87 | name), use the type() function to determine whether you have a single |
| 88 | instance or a list of instances. For example, here's code that |
| 89 | concatenates any number of username fields, separated by commas: |
| 90 | |
| 91 | username = form["username"] |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 92 | if type(username) is type([]): |
| 93 | # Multiple username fields specified |
| 94 | usernames = "" |
| 95 | for item in username: |
| 96 | if usernames: |
| 97 | # Next item -- insert comma |
| 98 | usernames = usernames + "," + item.value |
| 99 | else: |
| 100 | # First item -- don't insert comma |
| 101 | usernames = item.value |
| 102 | else: |
| 103 | # Single username field specified |
Guido van Rossum | 4032c2c | 1996-03-09 04:04:35 +0000 | [diff] [blame] | 104 | usernames = username.value |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 105 | |
| 106 | If a field represents an uploaded file, the value attribute reads the |
| 107 | entire file in memory as a string. This may not be what you want. You can |
| 108 | test for an uploaded file by testing either the filename attribute or the |
| 109 | file attribute. You can then read the data at leasure from the file |
| 110 | attribute: |
| 111 | |
| 112 | fileitem = form["userfile"] |
| 113 | if fileitem.file: |
| 114 | # It's an uploaded file; count lines |
| 115 | linecount = 0 |
| 116 | while 1: |
| 117 | line = fileitem.file.readline() |
| 118 | if not line: break |
| 119 | linecount = linecount + 1 |
| 120 | |
Guido van Rossum | 4032c2c | 1996-03-09 04:04:35 +0000 | [diff] [blame] | 121 | The file upload draft standard entertains the possibility of uploading |
| 122 | multiple files from one field (using a recursive multipart/* |
| 123 | encoding). When this occurs, the item will be a dictionary-like |
| 124 | FieldStorage item. This can be determined by testing its type |
| 125 | attribute, which should have the value "multipart/form-data" (or |
| 126 | perhaps another string beginning with "multipart/"). It this case, it |
| 127 | can be iterated over recursively just like the top-level form object. |
| 128 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 129 | When a form is submitted in the "old" format (as the query string or as a |
| 130 | single data part of type application/x-www-form-urlencoded), the items |
| 131 | will actually be instances of the class MiniFieldStorage. In this case, |
| 132 | the list, file and filename attributes are always None. |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 133 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 134 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 135 | Old classes |
| 136 | ----------- |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 137 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 138 | These classes, present in earlier versions of the cgi module, are still |
| 139 | supported for backward compatibility. New applications should use the |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 140 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 141 | SvFormContentDict: single value form content as dictionary; assumes each |
| 142 | field name occurs in the form only once. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 143 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 144 | FormContentDict: multiple value form content as dictionary (the form |
| 145 | items are lists of values). Useful if your form contains multiple |
| 146 | fields with the same name. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 147 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 148 | Other classes (FormContent, InterpFormContentDict) are present for |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 149 | backwards compatibility with really old applications only. If you still |
| 150 | use these and would be inconvenienced when they disappeared from a next |
| 151 | version of this module, drop me a note. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 152 | |
| 153 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 154 | Functions |
| 155 | --------- |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 156 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 157 | These are useful if you want more control, or if you want to employ |
| 158 | some of the algorithms implemented in this module in other |
| 159 | circumstances. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 160 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 161 | parse(fp): parse a form into a Python dictionary. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 162 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 163 | parse_qs(qs): parse a query string (data of type |
| 164 | application/x-www-form-urlencoded). |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 165 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 166 | parse_multipart(fp, pdict): parse input of type multipart/form-data (for |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 167 | file uploads). |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 168 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 169 | parse_header(string): parse a header like Content-type into a main |
| 170 | value and a dictionary of parameters. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 171 | |
| 172 | test(): complete test program. |
| 173 | |
| 174 | print_environ(): format the shell environment in HTML. |
| 175 | |
| 176 | print_form(form): format a form in HTML. |
| 177 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 178 | print_environ_usage(): print a list of useful environment variables in |
| 179 | HTML. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 180 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 181 | escape(): convert the characters "&", "<" and ">" to HTML-safe |
| 182 | sequences. Use this if you need to display text that might contain |
| 183 | such characters in HTML. To translate URLs for inclusion in the HREF |
| 184 | attribute of an <A> tag, use urllib.quote(). |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 185 | |
| 186 | |
| 187 | Caring about security |
| 188 | --------------------- |
| 189 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 190 | There's one important rule: if you invoke an external program (e.g. |
| 191 | via the os.system() or os.popen() functions), make very sure you don't |
| 192 | pass arbitrary strings received from the client to the shell. This is |
| 193 | a well-known security hole whereby clever hackers anywhere on the web |
| 194 | can exploit a gullible CGI script to invoke arbitrary shell commands. |
| 195 | Even parts of the URL or field names cannot be trusted, since the |
| 196 | request doesn't have to come from your form! |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 197 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 198 | To be on the safe side, if you must pass a string gotten from a form |
| 199 | to a shell command, you should make sure the string contains only |
| 200 | alphanumeric characters, dashes, underscores, and periods. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 201 | |
| 202 | |
| 203 | Installing your CGI script on a Unix system |
| 204 | ------------------------------------------- |
| 205 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 206 | Read the documentation for your HTTP server and check with your local |
| 207 | system administrator to find the directory where CGI scripts should be |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 208 | installed; usually this is in a directory cgi-bin in the server tree. |
| 209 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 210 | Make sure that your script is readable and executable by "others"; the |
| 211 | Unix file mode should be 755 (use "chmod 755 filename"). Make sure |
| 212 | that the first line of the script contains "#!" starting in column 1 |
| 213 | followed by the pathname of the Python interpreter, for instance: |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 214 | |
| 215 | #!/usr/local/bin/python |
| 216 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 217 | Make sure the Python interpreter exists and is executable by "others". |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 218 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 219 | Make sure that any files your script needs to read or write are |
| 220 | readable or writable, respectively, by "others" -- their mode should |
| 221 | be 644 for readable and 666 for writable. This is because, for |
| 222 | security reasons, the HTTP server executes your script as user |
| 223 | "nobody", without any special privileges. It can only read (write, |
| 224 | execute) files that everybody can read (write, execute). The current |
| 225 | directory at execution time is also different (it is usually the |
| 226 | server's cgi-bin directory) and the set of environment variables is |
| 227 | also different from what you get at login. in particular, don't count |
| 228 | on the shell's search path for executables ($PATH) or the Python |
| 229 | module search path ($PYTHONPATH) to be set to anything interesting. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 230 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 231 | If you need to load modules from a directory which is not on Python's |
| 232 | default module search path, you can change the path in your script, |
| 233 | before importing other modules, e.g.: |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 234 | |
| 235 | import sys |
| 236 | sys.path.insert(0, "/usr/home/joe/lib/python") |
| 237 | sys.path.insert(0, "/usr/local/lib/python") |
| 238 | |
| 239 | (This way, the directory inserted last will be searched first!) |
| 240 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 241 | Instructions for non-Unix systems will vary; check your HTTP server's |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 242 | documentation (it will usually have a section on CGI scripts). |
| 243 | |
| 244 | |
| 245 | Testing your CGI script |
| 246 | ----------------------- |
| 247 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 248 | Unfortunately, a CGI script will generally not run when you try it |
| 249 | from the command line, and a script that works perfectly from the |
| 250 | command line may fail mysteriously when run from the server. There's |
| 251 | one reason why you should still test your script from the command |
| 252 | line: if it contains a syntax error, the python interpreter won't |
| 253 | execute it at all, and the HTTP server will most likely send a cryptic |
| 254 | error to the client. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 255 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 256 | Assuming your script has no syntax errors, yet it does not work, you |
| 257 | have no choice but to read the next section: |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 258 | |
| 259 | |
| 260 | Debugging CGI scripts |
| 261 | --------------------- |
| 262 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 263 | First of all, check for trivial installation errors -- reading the |
| 264 | section above on installing your CGI script carefully can save you a |
| 265 | lot of time. If you wonder whether you have understood the |
| 266 | installation procedure correctly, try installing a copy of this module |
| 267 | file (cgi.py) as a CGI script. When invoked as a script, the file |
| 268 | will dump its environment and the contents of the form in HTML form. |
| 269 | Give it the right mode etc, and send it a request. If it's installed |
| 270 | in the standard cgi-bin directory, it should be possible to send it a |
| 271 | request by entering a URL into your browser of the form: |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 272 | |
| 273 | http://yourhostname/cgi-bin/cgi.py?name=Joe+Blow&addr=At+Home |
| 274 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 275 | If this gives an error of type 404, the server cannot find the script |
| 276 | -- perhaps you need to install it in a different directory. If it |
| 277 | gives another error (e.g. 500), there's an installation problem that |
| 278 | you should fix before trying to go any further. If you get a nicely |
| 279 | formatted listing of the environment and form content (in this |
| 280 | example, the fields should be listed as "addr" with value "At Home" |
| 281 | and "name" with value "Joe Blow"), the cgi.py script has been |
| 282 | installed correctly. If you follow the same procedure for your own |
| 283 | script, you should now be able to debug it. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 284 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 285 | The next step could be to call the cgi module's test() function from |
| 286 | your script: replace its main code with the single statement |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 287 | |
| 288 | cgi.test() |
| 289 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 290 | This should produce the same results as those gotten from installing |
| 291 | the cgi.py file itself. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 292 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 293 | When an ordinary Python script raises an unhandled exception |
| 294 | (e.g. because of a typo in a module name, a file that can't be opened, |
| 295 | etc.), the Python interpreter prints a nice traceback and exits. |
| 296 | While the Python interpreter will still do this when your CGI script |
| 297 | raises an exception, most likely the traceback will end up in one of |
| 298 | the HTTP server's log file, or be discarded altogether. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 299 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 300 | Fortunately, once you have managed to get your script to execute |
| 301 | *some* code, it is easy to catch exceptions and cause a traceback to |
| 302 | be printed. The test() function below in this module is an example. |
| 303 | Here are the rules: |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 304 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 305 | 1. Import the traceback module (before entering the |
| 306 | try-except!) |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 307 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 308 | 2. Make sure you finish printing the headers and the blank |
| 309 | line early |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 310 | |
| 311 | 3. Assign sys.stderr to sys.stdout |
| 312 | |
| 313 | 3. Wrap all remaining code in a try-except statement |
| 314 | |
| 315 | 4. In the except clause, call traceback.print_exc() |
| 316 | |
| 317 | For example: |
| 318 | |
| 319 | import sys |
| 320 | import traceback |
| 321 | print "Content-type: text/html" |
| 322 | print |
| 323 | sys.stderr = sys.stdout |
| 324 | try: |
| 325 | ...your code here... |
| 326 | except: |
| 327 | print "\n\n<PRE>" |
| 328 | traceback.print_exc() |
| 329 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 330 | Notes: The assignment to sys.stderr is needed because the traceback |
| 331 | prints to sys.stderr. The print "\n\n<PRE>" statement is necessary to |
| 332 | disable the word wrapping in HTML. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 333 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 334 | If you suspect that there may be a problem in importing the traceback |
| 335 | module, you can use an even more robust approach (which only uses |
| 336 | built-in modules): |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 337 | |
| 338 | import sys |
| 339 | sys.stderr = sys.stdout |
| 340 | print "Content-type: text/plain" |
| 341 | print |
| 342 | ...your code here... |
| 343 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 344 | This relies on the Python interpreter to print the traceback. The |
| 345 | content type of the output is set to plain text, which disables all |
| 346 | HTML processing. If your script works, the raw HTML will be displayed |
| 347 | by your client. If it raises an exception, most likely after the |
| 348 | first two lines have been printed, a traceback will be displayed. |
| 349 | Because no HTML interpretation is going on, the traceback will |
| 350 | readable. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 351 | |
| 352 | Good luck! |
| 353 | |
| 354 | |
| 355 | Common problems and solutions |
| 356 | ----------------------------- |
| 357 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 358 | - Most HTTP servers buffer the output from CGI scripts until the |
| 359 | script is completed. This means that it is not possible to display a |
| 360 | progress report on the client's display while the script is running. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 361 | |
| 362 | - Check the installation instructions above. |
| 363 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 364 | - Check the HTTP server's log files. ("tail -f logfile" in a separate |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 365 | window may be useful!) |
| 366 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 367 | - Always check a script for syntax errors first, by doing something |
| 368 | like "python script.py". |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 369 | |
| 370 | - When using any of the debugging techniques, don't forget to add |
| 371 | "import sys" to the top of the script. |
| 372 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 373 | - When invoking external programs, make sure they can be found. |
| 374 | Usually, this means using absolute path names -- $PATH is usually not |
| 375 | set to a very useful value in a CGI script. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 376 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 377 | - When reading or writing external files, make sure they can be read |
| 378 | or written by every user on the system. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 379 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 380 | - Don't try to give a CGI script a set-uid mode. This doesn't work on |
| 381 | most systems, and is a security liability as well. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 382 | |
| 383 | |
| 384 | History |
| 385 | ------- |
| 386 | |
Guido van Rossum | 391b4e6 | 1996-03-06 19:11:33 +0000 | [diff] [blame] | 387 | Michael McLay started this module. Steve Majewski changed the |
| 388 | interface to SvFormContentDict and FormContentDict. The multipart |
| 389 | parsing was inspired by code submitted by Andreas Paepcke. Guido van |
| 390 | Rossum rewrote, reformatted and documented the module and is currently |
| 391 | responsible for its maintenance. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 392 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 393 | |
| 394 | XXX The module is getting pretty heavy with all those docstrings. |
| 395 | Perhaps there should be a slimmed version that doesn't contain all those |
| 396 | backwards compatible and debugging classes and functions? |
| 397 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 398 | """ |
| 399 | |
Guido van Rossum | fa59cd7 | 1996-05-28 22:58:15 +0000 | [diff] [blame] | 400 | __version__ = "2.0b1" |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 401 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 402 | |
| 403 | # Imports |
| 404 | # ======= |
| 405 | |
| 406 | import string |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 407 | import sys |
| 408 | import os |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 409 | |
| 410 | |
| 411 | # A shorthand for os.environ |
| 412 | environ = os.environ |
| 413 | |
| 414 | |
| 415 | # Parsing functions |
| 416 | # ================= |
| 417 | |
| 418 | def parse(fp=None): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 419 | """Parse a query in the environment or from a file (default stdin)""" |
| 420 | if not fp: |
| 421 | fp = sys.stdin |
| 422 | if not environ.has_key('REQUEST_METHOD'): |
| 423 | environ['REQUEST_METHOD'] = 'GET' # For testing stand-alone |
| 424 | if environ['REQUEST_METHOD'] == 'POST': |
| 425 | ctype, pdict = parse_header(environ['CONTENT_TYPE']) |
| 426 | if ctype == 'multipart/form-data': |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 427 | return parse_multipart(fp, pdict) |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 428 | elif ctype == 'application/x-www-form-urlencoded': |
| 429 | clength = string.atoi(environ['CONTENT_LENGTH']) |
| 430 | qs = fp.read(clength) |
Guido van Rossum | 1c9daa8 | 1995-09-18 21:52:37 +0000 | [diff] [blame] | 431 | else: |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 432 | qs = '' # Unknown content-type |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 433 | environ['QUERY_STRING'] = qs # XXX Shouldn't, really |
| 434 | elif environ.has_key('QUERY_STRING'): |
| 435 | qs = environ['QUERY_STRING'] |
| 436 | else: |
| 437 | if sys.argv[1:]: |
| 438 | qs = sys.argv[1] |
| 439 | else: |
| 440 | qs = "" |
| 441 | environ['QUERY_STRING'] = qs # XXX Shouldn't, really |
| 442 | return parse_qs(qs) |
Guido van Rossum | e780877 | 1995-08-07 20:12:09 +0000 | [diff] [blame] | 443 | |
| 444 | |
| 445 | def parse_qs(qs): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 446 | """Parse a query given as a string argument""" |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 447 | import urllib, regsub |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 448 | name_value_pairs = string.splitfields(qs, '&') |
| 449 | dict = {} |
| 450 | for name_value in name_value_pairs: |
| 451 | nv = string.splitfields(name_value, '=') |
| 452 | if len(nv) != 2: |
| 453 | continue |
| 454 | name = nv[0] |
| 455 | value = urllib.unquote(regsub.gsub('+', ' ', nv[1])) |
| 456 | if len(value): |
| 457 | if dict.has_key (name): |
| 458 | dict[name].append(value) |
| 459 | else: |
| 460 | dict[name] = [value] |
| 461 | return dict |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 462 | |
| 463 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 464 | def parse_multipart(fp, pdict): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 465 | """Parse multipart input. |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 466 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 467 | Arguments: |
| 468 | fp : input file |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 469 | pdict: dictionary containing other parameters of conten-type header |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 470 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 471 | Returns a dictionary just like parse_qs(): keys are the field names, each |
| 472 | value is a list of values for that field. This is easy to use but not |
| 473 | much good if you are expecting megabytes to be uploaded -- in that case, |
| 474 | use the FieldStorage class instead which is much more flexible. Note |
| 475 | that content-type is the raw, unparsed contents of the content-type |
| 476 | header. |
| 477 | |
| 478 | XXX This does not parse nested multipart parts -- use FieldStorage for |
| 479 | that. |
| 480 | |
| 481 | XXX This should really be subsumed by FieldStorage altogether -- no |
| 482 | point in having two implementations of the same parsing algorithm. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 483 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 484 | """ |
| 485 | import mimetools |
| 486 | if pdict.has_key('boundary'): |
| 487 | boundary = pdict['boundary'] |
| 488 | else: |
| 489 | boundary = "" |
| 490 | nextpart = "--" + boundary |
| 491 | lastpart = "--" + boundary + "--" |
| 492 | partdict = {} |
| 493 | terminator = "" |
| 494 | |
| 495 | while terminator != lastpart: |
| 496 | bytes = -1 |
| 497 | data = None |
| 498 | if terminator: |
| 499 | # At start of next part. Read headers first. |
| 500 | headers = mimetools.Message(fp) |
| 501 | clength = headers.getheader('content-length') |
| 502 | if clength: |
| 503 | try: |
| 504 | bytes = string.atoi(clength) |
| 505 | except string.atoi_error: |
| 506 | pass |
| 507 | if bytes > 0: |
| 508 | data = fp.read(bytes) |
| 509 | else: |
| 510 | data = "" |
| 511 | # Read lines until end of part. |
| 512 | lines = [] |
| 513 | while 1: |
| 514 | line = fp.readline() |
| 515 | if not line: |
| 516 | terminator = lastpart # End outer loop |
| 517 | break |
| 518 | if line[:2] == "--": |
| 519 | terminator = string.strip(line) |
| 520 | if terminator in (nextpart, lastpart): |
| 521 | break |
| 522 | if line[-2:] == '\r\n': |
| 523 | line = line[:-2] |
| 524 | elif line[-1:] == '\n': |
| 525 | line = line[:-1] |
| 526 | lines.append(line) |
| 527 | # Done with part. |
| 528 | if data is None: |
| 529 | continue |
| 530 | if bytes < 0: |
| 531 | data = string.joinfields(lines, "\n") |
| 532 | line = headers['content-disposition'] |
| 533 | if not line: |
| 534 | continue |
| 535 | key, params = parse_header(line) |
| 536 | if key != 'form-data': |
| 537 | continue |
| 538 | if params.has_key('name'): |
| 539 | name = params['name'] |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 540 | else: |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 541 | continue |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 542 | if partdict.has_key(name): |
| 543 | partdict[name].append(data) |
| 544 | else: |
| 545 | partdict[name] = [data] |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 546 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 547 | return partdict |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 548 | |
| 549 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 550 | def parse_header(line): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 551 | """Parse a Content-type like header. |
| 552 | |
| 553 | Return the main content-type and a dictionary of options. |
| 554 | |
| 555 | """ |
| 556 | plist = map(string.strip, string.splitfields(line, ';')) |
| 557 | key = string.lower(plist[0]) |
| 558 | del plist[0] |
| 559 | pdict = {} |
| 560 | for p in plist: |
| 561 | i = string.find(p, '=') |
| 562 | if i >= 0: |
| 563 | name = string.lower(string.strip(p[:i])) |
| 564 | value = string.strip(p[i+1:]) |
| 565 | if len(value) >= 2 and value[0] == value[-1] == '"': |
| 566 | value = value[1:-1] |
| 567 | pdict[name] = value |
| 568 | return key, pdict |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 569 | |
| 570 | |
Guido van Rossum | 243ddcd | 1996-03-07 06:33:07 +0000 | [diff] [blame] | 571 | # Classes for field storage |
| 572 | # ========================= |
| 573 | |
| 574 | class MiniFieldStorage: |
| 575 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 576 | """Like FieldStorage, for use when no file uploads are possible.""" |
Guido van Rossum | 243ddcd | 1996-03-07 06:33:07 +0000 | [diff] [blame] | 577 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 578 | # Dummy attributes |
| 579 | filename = None |
| 580 | list = None |
| 581 | type = None |
Guido van Rossum | 4032c2c | 1996-03-09 04:04:35 +0000 | [diff] [blame] | 582 | type_options = {} |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 583 | disposition = None |
| 584 | disposition_options = {} |
| 585 | headers = {} |
Guido van Rossum | 243ddcd | 1996-03-07 06:33:07 +0000 | [diff] [blame] | 586 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 587 | def __init__(self, name, value): |
| 588 | """Constructor from field name and value.""" |
| 589 | from StringIO import StringIO |
| 590 | self.name = name |
| 591 | self.value = value |
| 592 | self.file = StringIO(value) |
| 593 | |
| 594 | def __repr__(self): |
| 595 | """Return printable representation.""" |
| 596 | return "MiniFieldStorage(%s, %s)" % (`self.name`, `self.value`) |
Guido van Rossum | 243ddcd | 1996-03-07 06:33:07 +0000 | [diff] [blame] | 597 | |
| 598 | |
| 599 | class FieldStorage: |
| 600 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 601 | """Store a sequence of fields, reading multipart/form-data. |
Guido van Rossum | 243ddcd | 1996-03-07 06:33:07 +0000 | [diff] [blame] | 602 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 603 | This class provides naming, typing, files stored on disk, and |
| 604 | more. At the top level, it is accessible like a dictionary, whose |
| 605 | keys are the field names. (Note: None can occur as a field name.) |
| 606 | The items are either a Python list (if there's multiple values) or |
| 607 | another FieldStorage or MiniFieldStorage object. If it's a single |
| 608 | object, it has the following attributes: |
Guido van Rossum | 243ddcd | 1996-03-07 06:33:07 +0000 | [diff] [blame] | 609 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 610 | name: the field name, if specified; otherwise None |
Guido van Rossum | 243ddcd | 1996-03-07 06:33:07 +0000 | [diff] [blame] | 611 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 612 | filename: the filename, if specified; otherwise None; this is the |
| 613 | client side filename, *not* the file name on which it is |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 614 | stored (that's a temporary file you don't deal with) |
Guido van Rossum | 243ddcd | 1996-03-07 06:33:07 +0000 | [diff] [blame] | 615 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 616 | value: the value as a *string*; for file uploads, this |
| 617 | transparently reads the file every time you request the value |
| 618 | |
| 619 | file: the file(-like) object from which you can read the data; |
| 620 | None if the data is stored a simple string |
| 621 | |
| 622 | type: the content-type, or None if not specified |
| 623 | |
| 624 | type_options: dictionary of options specified on the content-type |
| 625 | line |
| 626 | |
| 627 | disposition: content-disposition, or None if not specified |
| 628 | |
| 629 | disposition_options: dictionary of corresponding options |
| 630 | |
| 631 | headers: a dictionary(-like) object (sometimes rfc822.Message or a |
| 632 | subclass thereof) containing *all* headers |
| 633 | |
| 634 | The class is subclassable, mostly for the purpose of overriding |
| 635 | the make_file() method, which is called internally to come up with |
| 636 | a file open for reading and writing. This makes it possible to |
| 637 | override the default choice of storing all files in a temporary |
| 638 | directory and unlinking them as soon as they have been opened. |
| 639 | |
| 640 | """ |
| 641 | |
| 642 | def __init__(self, fp=None, headers=None, outerboundary=""): |
| 643 | """Constructor. Read multipart/* until last part. |
| 644 | |
| 645 | Arguments, all optional: |
| 646 | |
| 647 | fp : file pointer; default: sys.stdin |
| 648 | |
| 649 | headers : header dictionary-like object; default: |
| 650 | taken from environ as per CGI spec |
| 651 | |
| 652 | outerboundary : optional terminating multipart boundary |
| 653 | (for internal use only) |
| 654 | |
| 655 | """ |
| 656 | method = None |
| 657 | if environ.has_key('REQUEST_METHOD'): |
| 658 | method = string.upper(environ['REQUEST_METHOD']) |
| 659 | if not fp and method == 'GET': |
| 660 | qs = None |
| 661 | if environ.has_key('QUERY_STRING'): |
| 662 | qs = environ['QUERY_STRING'] |
| 663 | from StringIO import StringIO |
| 664 | fp = StringIO(qs or "") |
| 665 | if headers is None: |
| 666 | headers = {'content-type': |
| 667 | "application/x-www-form-urlencoded"} |
| 668 | if headers is None: |
| 669 | headers = {} |
| 670 | if environ.has_key('CONTENT_TYPE'): |
| 671 | headers['content-type'] = environ['CONTENT_TYPE'] |
| 672 | if environ.has_key('CONTENT_LENGTH'): |
| 673 | headers['content-length'] = environ['CONTENT_LENGTH'] |
| 674 | self.fp = fp or sys.stdin |
| 675 | self.headers = headers |
| 676 | self.outerboundary = outerboundary |
| 677 | |
| 678 | # Process content-disposition header |
| 679 | cdisp, pdict = "", {} |
| 680 | if self.headers.has_key('content-disposition'): |
| 681 | cdisp, pdict = parse_header(self.headers['content-disposition']) |
| 682 | self.disposition = cdisp |
| 683 | self.disposition_options = pdict |
| 684 | self.name = None |
| 685 | if pdict.has_key('name'): |
| 686 | self.name = pdict['name'] |
| 687 | self.filename = None |
| 688 | if pdict.has_key('filename'): |
| 689 | self.filename = pdict['filename'] |
| 690 | |
| 691 | # Process content-type header |
| 692 | ctype, pdict = "text/plain", {} |
| 693 | if self.headers.has_key('content-type'): |
| 694 | ctype, pdict = parse_header(self.headers['content-type']) |
| 695 | self.type = ctype |
| 696 | self.type_options = pdict |
| 697 | self.innerboundary = "" |
| 698 | if pdict.has_key('boundary'): |
| 699 | self.innerboundary = pdict['boundary'] |
| 700 | clen = -1 |
| 701 | if self.headers.has_key('content-length'): |
| 702 | try: |
| 703 | clen = string.atoi(self.headers['content-length']) |
| 704 | except: |
| 705 | pass |
| 706 | self.length = clen |
| 707 | |
| 708 | self.list = self.file = None |
| 709 | self.done = 0 |
| 710 | self.lines = [] |
| 711 | if ctype == 'application/x-www-form-urlencoded': |
| 712 | self.read_urlencoded() |
| 713 | elif ctype[:10] == 'multipart/': |
| 714 | self.read_multi() |
| 715 | else: |
| 716 | self.read_single() |
| 717 | |
| 718 | def __repr__(self): |
| 719 | """Return a printable representation.""" |
| 720 | return "FieldStorage(%s, %s, %s)" % ( |
| 721 | `self.name`, `self.filename`, `self.value`) |
| 722 | |
| 723 | def __getattr__(self, name): |
| 724 | if name != 'value': |
| 725 | raise AttributeError, name |
| 726 | if self.file: |
| 727 | self.file.seek(0) |
| 728 | value = self.file.read() |
| 729 | self.file.seek(0) |
| 730 | elif self.list is not None: |
| 731 | value = self.list |
| 732 | else: |
| 733 | value = None |
| 734 | return value |
| 735 | |
| 736 | def __getitem__(self, key): |
| 737 | """Dictionary style indexing.""" |
| 738 | if self.list is None: |
| 739 | raise TypeError, "not indexable" |
| 740 | found = [] |
| 741 | for item in self.list: |
| 742 | if item.name == key: found.append(item) |
| 743 | if not found: |
| 744 | raise KeyError, key |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 745 | if len(found) == 1: |
| 746 | return found[0] |
| 747 | else: |
| 748 | return found |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 749 | |
| 750 | def keys(self): |
| 751 | """Dictionary style keys() method.""" |
| 752 | if self.list is None: |
| 753 | raise TypeError, "not indexable" |
| 754 | keys = [] |
| 755 | for item in self.list: |
| 756 | if item.name not in keys: keys.append(item.name) |
| 757 | return keys |
| 758 | |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 759 | def has_key(self, key): |
| 760 | """Dictionary style has_key() method.""" |
| 761 | if self.list is None: |
| 762 | raise TypeError, "not indexable" |
| 763 | for item in self.list: |
| 764 | if item.name == key: return 1 |
| 765 | return 0 |
| 766 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 767 | def read_urlencoded(self): |
| 768 | """Internal: read data in query string format.""" |
| 769 | qs = self.fp.read(self.length) |
| 770 | dict = parse_qs(qs) |
| 771 | self.list = [] |
| 772 | for key, valuelist in dict.items(): |
| 773 | for value in valuelist: |
| 774 | self.list.append(MiniFieldStorage(key, value)) |
| 775 | self.skip_lines() |
| 776 | |
| 777 | def read_multi(self): |
| 778 | """Internal: read a part that is itself multipart.""" |
| 779 | import rfc822 |
| 780 | self.list = [] |
| 781 | part = self.__class__(self.fp, {}, self.innerboundary) |
| 782 | # Throw first part away |
| 783 | while not part.done: |
| 784 | headers = rfc822.Message(self.fp) |
| 785 | part = self.__class__(self.fp, headers, self.innerboundary) |
| 786 | self.list.append(part) |
| 787 | self.skip_lines() |
| 788 | |
| 789 | def read_single(self): |
| 790 | """Internal: read an atomic part.""" |
| 791 | if self.length >= 0: |
| 792 | self.read_binary() |
| 793 | self.skip_lines() |
| 794 | else: |
| 795 | self.read_lines() |
| 796 | self.file.seek(0) |
| 797 | |
| 798 | bufsize = 8*1024 # I/O buffering size for copy to file |
| 799 | |
| 800 | def read_binary(self): |
| 801 | """Internal: read binary data.""" |
| 802 | self.file = self.make_file('b') |
| 803 | todo = self.length |
| 804 | if todo >= 0: |
| 805 | while todo > 0: |
| 806 | data = self.fp.read(min(todo, self.bufsize)) |
| 807 | if not data: |
| 808 | self.done = -1 |
| 809 | break |
| 810 | self.file.write(data) |
| 811 | todo = todo - len(data) |
| 812 | |
| 813 | def read_lines(self): |
| 814 | """Internal: read lines until EOF or outerboundary.""" |
| 815 | self.file = self.make_file('') |
| 816 | if self.outerboundary: |
| 817 | self.read_lines_to_outerboundary() |
| 818 | else: |
| 819 | self.read_lines_to_eof() |
| 820 | |
| 821 | def read_lines_to_eof(self): |
| 822 | """Internal: read lines until EOF.""" |
| 823 | while 1: |
| 824 | line = self.fp.readline() |
| 825 | if not line: |
| 826 | self.done = -1 |
| 827 | break |
| 828 | self.lines.append(line) |
| 829 | if line[-2:] == '\r\n': |
| 830 | line = line[:-2] + '\n' |
| 831 | self.file.write(line) |
| 832 | |
| 833 | def read_lines_to_outerboundary(self): |
| 834 | """Internal: read lines until outerboundary.""" |
| 835 | next = "--" + self.outerboundary |
| 836 | last = next + "--" |
| 837 | delim = "" |
| 838 | while 1: |
| 839 | line = self.fp.readline() |
| 840 | if not line: |
| 841 | self.done = -1 |
| 842 | break |
| 843 | self.lines.append(line) |
| 844 | if line[:2] == "--": |
| 845 | strippedline = string.strip(line) |
| 846 | if strippedline == next: |
| 847 | break |
| 848 | if strippedline == last: |
| 849 | self.done = 1 |
| 850 | break |
| 851 | if line[-2:] == "\r\n": |
| 852 | line = line[:-2] |
| 853 | elif line[-1] == "\n": |
| 854 | line = line[:-1] |
| 855 | self.file.write(delim + line) |
| 856 | delim = "\n" |
| 857 | |
| 858 | def skip_lines(self): |
| 859 | """Internal: skip lines until outer boundary if defined.""" |
| 860 | if not self.outerboundary or self.done: |
| 861 | return |
| 862 | next = "--" + self.outerboundary |
| 863 | last = next + "--" |
| 864 | while 1: |
| 865 | line = self.fp.readline() |
| 866 | if not line: |
| 867 | self.done = -1 |
| 868 | break |
| 869 | self.lines.append(line) |
| 870 | if line[:2] == "--": |
| 871 | strippedline = string.strip(line) |
| 872 | if strippedline == next: |
| 873 | break |
| 874 | if strippedline == last: |
| 875 | self.done = 1 |
| 876 | break |
| 877 | |
| 878 | def make_file(self, binary): |
| 879 | """Overridable: return a readable & writable file. |
| 880 | |
| 881 | The file will be used as follows: |
| 882 | - data is written to it |
| 883 | - seek(0) |
| 884 | - data is read from it |
| 885 | |
| 886 | The 'binary' argument is 'b' if the file should be created in |
| 887 | binary mode (on non-Unix systems), '' otherwise. |
| 888 | |
Guido van Rossum | 4032c2c | 1996-03-09 04:04:35 +0000 | [diff] [blame] | 889 | This version opens a temporary file for reading and writing, |
| 890 | and immediately deletes (unlinks) it. The trick (on Unix!) is |
| 891 | that the file can still be used, but it can't be opened by |
| 892 | another process, and it will automatically be deleted when it |
| 893 | is closed or when the current process terminates. |
| 894 | |
| 895 | If you want a more permanent file, you derive a class which |
| 896 | overrides this method. If you want a visible temporary file |
| 897 | that is nevertheless automatically deleted when the script |
| 898 | terminates, try defining a __del__ method in a derived class |
| 899 | which unlinks the temporary files you have created. |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 900 | |
| 901 | """ |
Guido van Rossum | 4032c2c | 1996-03-09 04:04:35 +0000 | [diff] [blame] | 902 | import tempfile |
| 903 | tfn = tempfile.mktemp() |
| 904 | f = open(tfn, "w%s+" % binary) |
| 905 | os.unlink(tfn) |
| 906 | return f |
Guido van Rossum | 243ddcd | 1996-03-07 06:33:07 +0000 | [diff] [blame] | 907 | |
| 908 | |
Guido van Rossum | 4032c2c | 1996-03-09 04:04:35 +0000 | [diff] [blame] | 909 | # Backwards Compatibility Classes |
| 910 | # =============================== |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 911 | |
| 912 | class FormContentDict: |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 913 | """Basic (multiple values per field) form content as dictionary. |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 914 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 915 | form = FormContentDict() |
| 916 | |
| 917 | form[key] -> [value, value, ...] |
| 918 | form.has_key(key) -> Boolean |
| 919 | form.keys() -> [key, key, ...] |
| 920 | form.values() -> [[val, val, ...], [val, val, ...], ...] |
| 921 | form.items() -> [(key, [val, val, ...]), (key, [val, val, ...]), ...] |
| 922 | form.dict == {key: [val, val, ...], ...} |
| 923 | |
| 924 | """ |
| 925 | def __init__( self ): |
| 926 | self.dict = parse() |
| 927 | self.query_string = environ['QUERY_STRING'] |
| 928 | def __getitem__(self,key): |
| 929 | return self.dict[key] |
| 930 | def keys(self): |
| 931 | return self.dict.keys() |
| 932 | def has_key(self, key): |
| 933 | return self.dict.has_key(key) |
| 934 | def values(self): |
| 935 | return self.dict.values() |
| 936 | def items(self): |
| 937 | return self.dict.items() |
| 938 | def __len__( self ): |
| 939 | return len(self.dict) |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 940 | |
| 941 | |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 942 | class SvFormContentDict(FormContentDict): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 943 | """Strict single-value expecting form content as dictionary. |
| 944 | |
| 945 | IF you only expect a single value for each field, then form[key] |
| 946 | will return that single value. It will raise an IndexError if |
| 947 | that expectation is not true. IF you expect a field to have |
| 948 | possible multiple values, than you can use form.getlist(key) to |
| 949 | get all of the values. values() and items() are a compromise: |
| 950 | they return single strings where there is a single value, and |
| 951 | lists of strings otherwise. |
| 952 | |
| 953 | """ |
| 954 | def __getitem__(self, key): |
| 955 | if len(self.dict[key]) > 1: |
| 956 | raise IndexError, 'expecting a single value' |
| 957 | return self.dict[key][0] |
| 958 | def getlist(self, key): |
| 959 | return self.dict[key] |
| 960 | def values(self): |
| 961 | lis = [] |
| 962 | for each in self.dict.values(): |
| 963 | if len( each ) == 1 : |
| 964 | lis.append(each[0]) |
| 965 | else: lis.append(each) |
| 966 | return lis |
| 967 | def items(self): |
| 968 | lis = [] |
| 969 | for key,value in self.dict.items(): |
| 970 | if len(value) == 1 : |
| 971 | lis.append((key, value[0])) |
| 972 | else: lis.append((key, value)) |
| 973 | return lis |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 974 | |
| 975 | |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 976 | class InterpFormContentDict(SvFormContentDict): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 977 | """This class is present for backwards compatibility only.""" |
| 978 | def __getitem__( self, key ): |
| 979 | v = SvFormContentDict.__getitem__( self, key ) |
| 980 | if v[0] in string.digits+'+-.' : |
| 981 | try: return string.atoi( v ) |
| 982 | except ValueError: |
| 983 | try: return string.atof( v ) |
| 984 | except ValueError: pass |
| 985 | return string.strip(v) |
| 986 | def values( self ): |
| 987 | lis = [] |
| 988 | for key in self.keys(): |
| 989 | try: |
| 990 | lis.append( self[key] ) |
| 991 | except IndexError: |
| 992 | lis.append( self.dict[key] ) |
| 993 | return lis |
| 994 | def items( self ): |
| 995 | lis = [] |
| 996 | for key in self.keys(): |
| 997 | try: |
| 998 | lis.append( (key, self[key]) ) |
| 999 | except IndexError: |
| 1000 | lis.append( (key, self.dict[key]) ) |
| 1001 | return lis |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1002 | |
| 1003 | |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1004 | class FormContent(FormContentDict): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1005 | """This class is present for backwards compatibility only.""" |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 1006 | def values(self, key): |
| 1007 | if self.dict.has_key(key) :return self.dict[key] |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1008 | else: return None |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 1009 | def indexed_value(self, key, location): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1010 | if self.dict.has_key(key): |
| 1011 | if len (self.dict[key]) > location: |
| 1012 | return self.dict[key][location] |
| 1013 | else: return None |
| 1014 | else: return None |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 1015 | def value(self, key): |
| 1016 | if self.dict.has_key(key): return self.dict[key][0] |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1017 | else: return None |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 1018 | def length(self, key): |
| 1019 | return len(self.dict[key]) |
| 1020 | def stripped(self, key): |
| 1021 | if self.dict.has_key(key): return string.strip(self.dict[key][0]) |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1022 | else: return None |
| 1023 | def pars(self): |
| 1024 | return self.dict |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1025 | |
| 1026 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 1027 | # Test/debug code |
| 1028 | # =============== |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1029 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 1030 | def test(): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1031 | """Robust test CGI script, usable as main program. |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1032 | |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1033 | Write minimal HTTP headers and dump all information provided to |
| 1034 | the script in HTML form. |
| 1035 | |
| 1036 | """ |
| 1037 | import traceback |
| 1038 | print "Content-type: text/html" |
| 1039 | print |
| 1040 | sys.stderr = sys.stdout |
| 1041 | try: |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 1042 | form = FieldStorage() # Replace with other classes to test those |
| 1043 | print_form(form) |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1044 | print_environ() |
| 1045 | print_directory() |
Guido van Rossum | a8738a5 | 1996-03-14 21:30:28 +0000 | [diff] [blame] | 1046 | print_arguments() |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1047 | print_environ_usage() |
| 1048 | except: |
| 1049 | print "\n\n<PRE>" # Turn off HTML word wrap |
| 1050 | traceback.print_exc() |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1051 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 1052 | def print_environ(): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1053 | """Dump the shell environment as HTML.""" |
| 1054 | keys = environ.keys() |
| 1055 | keys.sort() |
| 1056 | print |
Guido van Rossum | 503e50b | 1996-05-28 22:57:20 +0000 | [diff] [blame] | 1057 | print "<H3>Shell Environment:</H3>" |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1058 | print "<DL>" |
| 1059 | for key in keys: |
| 1060 | print "<DT>", escape(key), "<DD>", escape(environ[key]) |
| 1061 | print "</DL>" |
| 1062 | print |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 1063 | |
| 1064 | def print_form(form): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1065 | """Dump the contents of a form as HTML.""" |
| 1066 | keys = form.keys() |
| 1067 | keys.sort() |
| 1068 | print |
Guido van Rossum | 503e50b | 1996-05-28 22:57:20 +0000 | [diff] [blame] | 1069 | print "<H3>Form Contents:</H3>" |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1070 | print "<DL>" |
| 1071 | for key in keys: |
| 1072 | print "<DT>" + escape(key) + ":", |
| 1073 | value = form[key] |
| 1074 | print "<i>" + escape(`type(value)`) + "</i>" |
| 1075 | print "<DD>" + escape(`value`) |
| 1076 | print "</DL>" |
| 1077 | print |
| 1078 | |
| 1079 | def print_directory(): |
| 1080 | """Dump the current directory as HTML.""" |
| 1081 | print |
| 1082 | print "<H3>Current Working Directory:</H3>" |
| 1083 | try: |
| 1084 | pwd = os.getcwd() |
| 1085 | except os.error, msg: |
| 1086 | print "os.error:", escape(str(msg)) |
| 1087 | else: |
| 1088 | print escape(pwd) |
| 1089 | print |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1090 | |
Guido van Rossum | a8738a5 | 1996-03-14 21:30:28 +0000 | [diff] [blame] | 1091 | def print_arguments(): |
| 1092 | print |
Guido van Rossum | 503e50b | 1996-05-28 22:57:20 +0000 | [diff] [blame] | 1093 | print "<H3>Command Line Arguments:</H3>" |
Guido van Rossum | a8738a5 | 1996-03-14 21:30:28 +0000 | [diff] [blame] | 1094 | print |
| 1095 | print sys.argv |
| 1096 | print |
| 1097 | |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1098 | def print_environ_usage(): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1099 | """Dump a list of environment variables used by CGI as HTML.""" |
| 1100 | print """ |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 1101 | <H3>These environment variables could have been set:</H3> |
| 1102 | <UL> |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1103 | <LI>AUTH_TYPE |
| 1104 | <LI>CONTENT_LENGTH |
| 1105 | <LI>CONTENT_TYPE |
| 1106 | <LI>DATE_GMT |
| 1107 | <LI>DATE_LOCAL |
| 1108 | <LI>DOCUMENT_NAME |
| 1109 | <LI>DOCUMENT_ROOT |
| 1110 | <LI>DOCUMENT_URI |
| 1111 | <LI>GATEWAY_INTERFACE |
| 1112 | <LI>LAST_MODIFIED |
| 1113 | <LI>PATH |
| 1114 | <LI>PATH_INFO |
| 1115 | <LI>PATH_TRANSLATED |
| 1116 | <LI>QUERY_STRING |
| 1117 | <LI>REMOTE_ADDR |
| 1118 | <LI>REMOTE_HOST |
| 1119 | <LI>REMOTE_IDENT |
| 1120 | <LI>REMOTE_USER |
| 1121 | <LI>REQUEST_METHOD |
| 1122 | <LI>SCRIPT_NAME |
| 1123 | <LI>SERVER_NAME |
| 1124 | <LI>SERVER_PORT |
| 1125 | <LI>SERVER_PROTOCOL |
| 1126 | <LI>SERVER_ROOT |
| 1127 | <LI>SERVER_SOFTWARE |
| 1128 | </UL> |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1129 | In addition, HTTP headers sent by the server may be passed in the |
| 1130 | environment as well. Here are some common variable names: |
| 1131 | <UL> |
| 1132 | <LI>HTTP_ACCEPT |
| 1133 | <LI>HTTP_CONNECTION |
| 1134 | <LI>HTTP_HOST |
| 1135 | <LI>HTTP_PRAGMA |
| 1136 | <LI>HTTP_REFERER |
| 1137 | <LI>HTTP_USER_AGENT |
| 1138 | </UL> |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1139 | """ |
| 1140 | |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1141 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 1142 | # Utilities |
| 1143 | # ========= |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1144 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 1145 | def escape(s): |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1146 | """Replace special characters '&', '<' and '>' by SGML entities.""" |
Guido van Rossum | 0147db0 | 1996-03-09 03:16:04 +0000 | [diff] [blame] | 1147 | import regsub |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1148 | s = regsub.gsub("&", "&", s) # Must be done first! |
| 1149 | s = regsub.gsub("<", "<", s) |
| 1150 | s = regsub.gsub(">", ">", s) |
| 1151 | return s |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1152 | |
Guido van Rossum | 9a22de1 | 1995-01-12 12:29:47 +0000 | [diff] [blame] | 1153 | |
Guido van Rossum | 7275561 | 1996-03-06 07:20:06 +0000 | [diff] [blame] | 1154 | # Invoke mainline |
| 1155 | # =============== |
| 1156 | |
| 1157 | # Call test() when this file is run as a script (not imported as a module) |
| 1158 | if __name__ == '__main__': |
Guido van Rossum | 7aee384 | 1996-03-07 18:00:44 +0000 | [diff] [blame] | 1159 | test() |