Show More
@@ -1,566 +1,529 | |||
|
1 | 1 | # hgweb/hgwebdir_mod.py - Web interface for a directory of repositories. |
|
2 | 2 | # |
|
3 | 3 | # Copyright 21 May 2005 - (c) 2005 Jake Edge <jake@edge2.net> |
|
4 | 4 | # Copyright 2005, 2006 Matt Mackall <mpm@selenic.com> |
|
5 | 5 | # |
|
6 | 6 | # This software may be used and distributed according to the terms of the |
|
7 | 7 | # GNU General Public License version 2 or any later version. |
|
8 | 8 | |
|
9 | 9 | from __future__ import absolute_import |
|
10 | 10 | |
|
11 | 11 | import os |
|
12 | 12 | import re |
|
13 | 13 | import time |
|
14 | 14 | |
|
15 | 15 | from ..i18n import _ |
|
16 | 16 | |
|
17 | 17 | from .common import ( |
|
18 | 18 | ErrorResponse, |
|
19 | 19 | HTTP_NOT_FOUND, |
|
20 | 20 | HTTP_OK, |
|
21 | 21 | HTTP_SERVER_ERROR, |
|
22 | 22 | cspvalues, |
|
23 | 23 | get_contact, |
|
24 | 24 | get_mtime, |
|
25 | 25 | ismember, |
|
26 | 26 | paritygen, |
|
27 | 27 | staticfile, |
|
28 | 28 | ) |
|
29 | 29 | |
|
30 | 30 | from .. import ( |
|
31 | 31 | configitems, |
|
32 | 32 | encoding, |
|
33 | 33 | error, |
|
34 | 34 | hg, |
|
35 | 35 | profiling, |
|
36 | pycompat, | |
|
37 | 36 | scmutil, |
|
38 | 37 | templater, |
|
39 | 38 | ui as uimod, |
|
40 | 39 | util, |
|
41 | 40 | ) |
|
42 | 41 | |
|
43 | 42 | from . import ( |
|
44 | 43 | hgweb_mod, |
|
45 | 44 | request as requestmod, |
|
46 | 45 | webutil, |
|
47 | 46 | wsgicgi, |
|
48 | 47 | ) |
|
49 | 48 | from ..utils import dateutil |
|
50 | 49 | |
|
51 | 50 | def cleannames(items): |
|
52 | 51 | return [(util.pconvert(name).strip('/'), path) for name, path in items] |
|
53 | 52 | |
|
54 | 53 | def findrepos(paths): |
|
55 | 54 | repos = [] |
|
56 | 55 | for prefix, root in cleannames(paths): |
|
57 | 56 | roothead, roottail = os.path.split(root) |
|
58 | 57 | # "foo = /bar/*" or "foo = /bar/**" lets every repo /bar/N in or below |
|
59 | 58 | # /bar/ be served as as foo/N . |
|
60 | 59 | # '*' will not search inside dirs with .hg (except .hg/patches), |
|
61 | 60 | # '**' will search inside dirs with .hg (and thus also find subrepos). |
|
62 | 61 | try: |
|
63 | 62 | recurse = {'*': False, '**': True}[roottail] |
|
64 | 63 | except KeyError: |
|
65 | 64 | repos.append((prefix, root)) |
|
66 | 65 | continue |
|
67 | 66 | roothead = os.path.normpath(os.path.abspath(roothead)) |
|
68 | 67 | paths = scmutil.walkrepos(roothead, followsym=True, recurse=recurse) |
|
69 | 68 | repos.extend(urlrepos(prefix, roothead, paths)) |
|
70 | 69 | return repos |
|
71 | 70 | |
|
72 | 71 | def urlrepos(prefix, roothead, paths): |
|
73 | 72 | """yield url paths and filesystem paths from a list of repo paths |
|
74 | 73 | |
|
75 | 74 | >>> conv = lambda seq: [(v, util.pconvert(p)) for v,p in seq] |
|
76 | 75 | >>> conv(urlrepos(b'hg', b'/opt', [b'/opt/r', b'/opt/r/r', b'/opt'])) |
|
77 | 76 | [('hg/r', '/opt/r'), ('hg/r/r', '/opt/r/r'), ('hg', '/opt')] |
|
78 | 77 | >>> conv(urlrepos(b'', b'/opt', [b'/opt/r', b'/opt/r/r', b'/opt'])) |
|
79 | 78 | [('r', '/opt/r'), ('r/r', '/opt/r/r'), ('', '/opt')] |
|
80 | 79 | """ |
|
81 | 80 | for path in paths: |
|
82 | 81 | path = os.path.normpath(path) |
|
83 | 82 | yield (prefix + '/' + |
|
84 | 83 | util.pconvert(path[len(roothead):]).lstrip('/')).strip('/'), path |
|
85 | 84 | |
|
86 | def geturlcgivars(baseurl, port): | |
|
87 | """ | |
|
88 | Extract CGI variables from baseurl | |
|
89 | ||
|
90 | >>> geturlcgivars(b"http://host.org/base", b"80") | |
|
91 | ('host.org', '80', '/base') | |
|
92 | >>> geturlcgivars(b"http://host.org:8000/base", b"80") | |
|
93 | ('host.org', '8000', '/base') | |
|
94 | >>> geturlcgivars(b'/base', 8000) | |
|
95 | ('', '8000', '/base') | |
|
96 | >>> geturlcgivars(b"base", b'8000') | |
|
97 | ('', '8000', '/base') | |
|
98 | >>> geturlcgivars(b"http://host", b'8000') | |
|
99 | ('host', '8000', '/') | |
|
100 | >>> geturlcgivars(b"http://host/", b'8000') | |
|
101 | ('host', '8000', '/') | |
|
102 | """ | |
|
103 | u = util.url(baseurl) | |
|
104 | name = u.host or '' | |
|
105 | if u.port: | |
|
106 | port = u.port | |
|
107 | path = u.path or "" | |
|
108 | if not path.startswith('/'): | |
|
109 | path = '/' + path | |
|
110 | ||
|
111 | return name, pycompat.bytestr(port), path | |
|
112 | ||
|
113 | 85 | def readallowed(ui, req): |
|
114 | 86 | """Check allow_read and deny_read config options of a repo's ui object |
|
115 | 87 | to determine user permissions. By default, with neither option set (or |
|
116 | 88 | both empty), allow all users to read the repo. There are two ways a |
|
117 | 89 | user can be denied read access: (1) deny_read is not empty, and the |
|
118 | 90 | user is unauthenticated or deny_read contains user (or *), and (2) |
|
119 | 91 | allow_read is not empty and the user is not in allow_read. Return True |
|
120 | 92 | if user is allowed to read the repo, else return False.""" |
|
121 | 93 | |
|
122 | 94 | user = req.remoteuser |
|
123 | 95 | |
|
124 | 96 | deny_read = ui.configlist('web', 'deny_read', untrusted=True) |
|
125 | 97 | if deny_read and (not user or ismember(ui, user, deny_read)): |
|
126 | 98 | return False |
|
127 | 99 | |
|
128 | 100 | allow_read = ui.configlist('web', 'allow_read', untrusted=True) |
|
129 | 101 | # by default, allow reading if no allow_read option has been set |
|
130 | 102 | if not allow_read or ismember(ui, user, allow_read): |
|
131 | 103 | return True |
|
132 | 104 | |
|
133 | 105 | return False |
|
134 | 106 | |
|
135 | 107 | def archivelist(ui, nodeid, url): |
|
136 | 108 | allowed = ui.configlist('web', 'allow_archive', untrusted=True) |
|
137 | 109 | archives = [] |
|
138 | 110 | |
|
139 | 111 | for typ, spec in hgweb_mod.archivespecs.iteritems(): |
|
140 | 112 | if typ in allowed or ui.configbool('web', 'allow' + typ, |
|
141 | 113 | untrusted=True): |
|
142 | 114 | archives.append({ |
|
143 | 115 | 'type': typ, |
|
144 | 116 | 'extension': spec[2], |
|
145 | 117 | 'node': nodeid, |
|
146 | 118 | 'url': url, |
|
147 | 119 | }) |
|
148 | 120 | |
|
149 | 121 | return archives |
|
150 | 122 | |
|
151 | 123 | def rawindexentries(ui, repos, wsgireq, req, subdir=''): |
|
152 | 124 | descend = ui.configbool('web', 'descend') |
|
153 | 125 | collapse = ui.configbool('web', 'collapse') |
|
154 | 126 | seenrepos = set() |
|
155 | 127 | seendirs = set() |
|
156 | 128 | for name, path in repos: |
|
157 | 129 | |
|
158 | 130 | if not name.startswith(subdir): |
|
159 | 131 | continue |
|
160 | 132 | name = name[len(subdir):] |
|
161 | 133 | directory = False |
|
162 | 134 | |
|
163 | 135 | if '/' in name: |
|
164 | 136 | if not descend: |
|
165 | 137 | continue |
|
166 | 138 | |
|
167 | 139 | nameparts = name.split('/') |
|
168 | 140 | rootname = nameparts[0] |
|
169 | 141 | |
|
170 | 142 | if not collapse: |
|
171 | 143 | pass |
|
172 | 144 | elif rootname in seendirs: |
|
173 | 145 | continue |
|
174 | 146 | elif rootname in seenrepos: |
|
175 | 147 | pass |
|
176 | 148 | else: |
|
177 | 149 | directory = True |
|
178 | 150 | name = rootname |
|
179 | 151 | |
|
180 | 152 | # redefine the path to refer to the directory |
|
181 | 153 | discarded = '/'.join(nameparts[1:]) |
|
182 | 154 | |
|
183 | 155 | # remove name parts plus accompanying slash |
|
184 | 156 | path = path[:-len(discarded) - 1] |
|
185 | 157 | |
|
186 | 158 | try: |
|
187 | 159 | r = hg.repository(ui, path) |
|
188 | 160 | directory = False |
|
189 | 161 | except (IOError, error.RepoError): |
|
190 | 162 | pass |
|
191 | 163 | |
|
192 | 164 | parts = [name] |
|
193 | 165 | parts.insert(0, '/' + subdir.rstrip('/')) |
|
194 | 166 | if wsgireq.env['SCRIPT_NAME']: |
|
195 | 167 | parts.insert(0, wsgireq.env['SCRIPT_NAME']) |
|
196 | 168 | url = re.sub(r'/+', '/', '/'.join(parts) + '/') |
|
197 | 169 | |
|
198 | 170 | # show either a directory entry or a repository |
|
199 | 171 | if directory: |
|
200 | 172 | # get the directory's time information |
|
201 | 173 | try: |
|
202 | 174 | d = (get_mtime(path), dateutil.makedate()[1]) |
|
203 | 175 | except OSError: |
|
204 | 176 | continue |
|
205 | 177 | |
|
206 | 178 | # add '/' to the name to make it obvious that |
|
207 | 179 | # the entry is a directory, not a regular repository |
|
208 | 180 | row = {'contact': "", |
|
209 | 181 | 'contact_sort': "", |
|
210 | 182 | 'name': name + '/', |
|
211 | 183 | 'name_sort': name, |
|
212 | 184 | 'url': url, |
|
213 | 185 | 'description': "", |
|
214 | 186 | 'description_sort': "", |
|
215 | 187 | 'lastchange': d, |
|
216 | 188 | 'lastchange_sort': d[1] - d[0], |
|
217 | 189 | 'archives': [], |
|
218 | 190 | 'isdirectory': True, |
|
219 | 191 | 'labels': [], |
|
220 | 192 | } |
|
221 | 193 | |
|
222 | 194 | seendirs.add(name) |
|
223 | 195 | yield row |
|
224 | 196 | continue |
|
225 | 197 | |
|
226 | 198 | u = ui.copy() |
|
227 | 199 | try: |
|
228 | 200 | u.readconfig(os.path.join(path, '.hg', 'hgrc')) |
|
229 | 201 | except Exception as e: |
|
230 | 202 | u.warn(_('error reading %s/.hg/hgrc: %s\n') % (path, e)) |
|
231 | 203 | continue |
|
232 | 204 | |
|
233 | 205 | def get(section, name, default=uimod._unset): |
|
234 | 206 | return u.config(section, name, default, untrusted=True) |
|
235 | 207 | |
|
236 | 208 | if u.configbool("web", "hidden", untrusted=True): |
|
237 | 209 | continue |
|
238 | 210 | |
|
239 | 211 | if not readallowed(u, req): |
|
240 | 212 | continue |
|
241 | 213 | |
|
242 | 214 | # update time with local timezone |
|
243 | 215 | try: |
|
244 | 216 | r = hg.repository(ui, path) |
|
245 | 217 | except IOError: |
|
246 | 218 | u.warn(_('error accessing repository at %s\n') % path) |
|
247 | 219 | continue |
|
248 | 220 | except error.RepoError: |
|
249 | 221 | u.warn(_('error accessing repository at %s\n') % path) |
|
250 | 222 | continue |
|
251 | 223 | try: |
|
252 | 224 | d = (get_mtime(r.spath), dateutil.makedate()[1]) |
|
253 | 225 | except OSError: |
|
254 | 226 | continue |
|
255 | 227 | |
|
256 | 228 | contact = get_contact(get) |
|
257 | 229 | description = get("web", "description") |
|
258 | 230 | seenrepos.add(name) |
|
259 | 231 | name = get("web", "name", name) |
|
260 | 232 | row = {'contact': contact or "unknown", |
|
261 | 233 | 'contact_sort': contact.upper() or "unknown", |
|
262 | 234 | 'name': name, |
|
263 | 235 | 'name_sort': name, |
|
264 | 236 | 'url': url, |
|
265 | 237 | 'description': description or "unknown", |
|
266 | 238 | 'description_sort': description.upper() or "unknown", |
|
267 | 239 | 'lastchange': d, |
|
268 | 240 | 'lastchange_sort': d[1] - d[0], |
|
269 | 241 | 'archives': archivelist(u, "tip", url), |
|
270 | 242 | 'isdirectory': None, |
|
271 | 243 | 'labels': u.configlist('web', 'labels', untrusted=True), |
|
272 | 244 | } |
|
273 | 245 | |
|
274 | 246 | yield row |
|
275 | 247 | |
|
276 | 248 | def indexentries(ui, repos, wsgireq, req, stripecount, sortcolumn='', |
|
277 | 249 | descending=False, subdir=''): |
|
278 | 250 | |
|
279 | 251 | rows = rawindexentries(ui, repos, wsgireq, req, subdir=subdir) |
|
280 | 252 | |
|
281 | 253 | sortdefault = None, False |
|
282 | 254 | |
|
283 | 255 | if sortcolumn and sortdefault != (sortcolumn, descending): |
|
284 | 256 | sortkey = '%s_sort' % sortcolumn |
|
285 | 257 | rows = sorted(rows, key=lambda x: x[sortkey], |
|
286 | 258 | reverse=descending) |
|
287 | 259 | |
|
288 | 260 | for row, parity in zip(rows, paritygen(stripecount)): |
|
289 | 261 | row['parity'] = parity |
|
290 | 262 | yield row |
|
291 | 263 | |
|
292 | 264 | class hgwebdir(object): |
|
293 | 265 | """HTTP server for multiple repositories. |
|
294 | 266 | |
|
295 | 267 | Given a configuration, different repositories will be served depending |
|
296 | 268 | on the request path. |
|
297 | 269 | |
|
298 | 270 | Instances are typically used as WSGI applications. |
|
299 | 271 | """ |
|
300 | 272 | def __init__(self, conf, baseui=None): |
|
301 | 273 | self.conf = conf |
|
302 | 274 | self.baseui = baseui |
|
303 | 275 | self.ui = None |
|
304 | 276 | self.lastrefresh = 0 |
|
305 | 277 | self.motd = None |
|
306 | 278 | self.refresh() |
|
307 | 279 | |
|
308 | 280 | def refresh(self): |
|
309 | 281 | if self.ui: |
|
310 | 282 | refreshinterval = self.ui.configint('web', 'refreshinterval') |
|
311 | 283 | else: |
|
312 | 284 | item = configitems.coreitems['web']['refreshinterval'] |
|
313 | 285 | refreshinterval = item.default |
|
314 | 286 | |
|
315 | 287 | # refreshinterval <= 0 means to always refresh. |
|
316 | 288 | if (refreshinterval > 0 and |
|
317 | 289 | self.lastrefresh + refreshinterval > time.time()): |
|
318 | 290 | return |
|
319 | 291 | |
|
320 | 292 | if self.baseui: |
|
321 | 293 | u = self.baseui.copy() |
|
322 | 294 | else: |
|
323 | 295 | u = uimod.ui.load() |
|
324 | 296 | u.setconfig('ui', 'report_untrusted', 'off', 'hgwebdir') |
|
325 | 297 | u.setconfig('ui', 'nontty', 'true', 'hgwebdir') |
|
326 | 298 | # displaying bundling progress bar while serving feels wrong and may |
|
327 | 299 | # break some wsgi implementations. |
|
328 | 300 | u.setconfig('progress', 'disable', 'true', 'hgweb') |
|
329 | 301 | |
|
330 | 302 | if not isinstance(self.conf, (dict, list, tuple)): |
|
331 | 303 | map = {'paths': 'hgweb-paths'} |
|
332 | 304 | if not os.path.exists(self.conf): |
|
333 | 305 | raise error.Abort(_('config file %s not found!') % self.conf) |
|
334 | 306 | u.readconfig(self.conf, remap=map, trust=True) |
|
335 | 307 | paths = [] |
|
336 | 308 | for name, ignored in u.configitems('hgweb-paths'): |
|
337 | 309 | for path in u.configlist('hgweb-paths', name): |
|
338 | 310 | paths.append((name, path)) |
|
339 | 311 | elif isinstance(self.conf, (list, tuple)): |
|
340 | 312 | paths = self.conf |
|
341 | 313 | elif isinstance(self.conf, dict): |
|
342 | 314 | paths = self.conf.items() |
|
343 | 315 | |
|
344 | 316 | repos = findrepos(paths) |
|
345 | 317 | for prefix, root in u.configitems('collections'): |
|
346 | 318 | prefix = util.pconvert(prefix) |
|
347 | 319 | for path in scmutil.walkrepos(root, followsym=True): |
|
348 | 320 | repo = os.path.normpath(path) |
|
349 | 321 | name = util.pconvert(repo) |
|
350 | 322 | if name.startswith(prefix): |
|
351 | 323 | name = name[len(prefix):] |
|
352 | 324 | repos.append((name.lstrip('/'), repo)) |
|
353 | 325 | |
|
354 | 326 | self.repos = repos |
|
355 | 327 | self.ui = u |
|
356 | 328 | encoding.encoding = self.ui.config('web', 'encoding') |
|
357 | 329 | self.style = self.ui.config('web', 'style') |
|
358 | 330 | self.templatepath = self.ui.config('web', 'templates', untrusted=False) |
|
359 | 331 | self.stripecount = self.ui.config('web', 'stripes') |
|
360 | 332 | if self.stripecount: |
|
361 | 333 | self.stripecount = int(self.stripecount) |
|
362 | self._baseurl = self.ui.config('web', 'baseurl') | |
|
363 | 334 | prefix = self.ui.config('web', 'prefix') |
|
364 | 335 | if prefix.startswith('/'): |
|
365 | 336 | prefix = prefix[1:] |
|
366 | 337 | if prefix.endswith('/'): |
|
367 | 338 | prefix = prefix[:-1] |
|
368 | 339 | self.prefix = prefix |
|
369 | 340 | self.lastrefresh = time.time() |
|
370 | 341 | |
|
371 | 342 | def run(self): |
|
372 | 343 | if not encoding.environ.get('GATEWAY_INTERFACE', |
|
373 | 344 | '').startswith("CGI/1."): |
|
374 | 345 | raise RuntimeError("This function is only intended to be " |
|
375 | 346 | "called while running as a CGI script.") |
|
376 | 347 | wsgicgi.launch(self) |
|
377 | 348 | |
|
378 | 349 | def __call__(self, env, respond): |
|
379 | wsgireq = requestmod.wsgirequest(env, respond) | |
|
350 | baseurl = self.ui.config('web', 'baseurl') | |
|
351 | wsgireq = requestmod.wsgirequest(env, respond, altbaseurl=baseurl) | |
|
380 | 352 | return self.run_wsgi(wsgireq) |
|
381 | 353 | |
|
382 | 354 | def run_wsgi(self, wsgireq): |
|
383 | 355 | profile = self.ui.configbool('profiling', 'enabled') |
|
384 | 356 | with profiling.profile(self.ui, enabled=profile): |
|
385 | 357 | for r in self._runwsgi(wsgireq): |
|
386 | 358 | yield r |
|
387 | 359 | |
|
388 | 360 | def _runwsgi(self, wsgireq): |
|
389 | 361 | req = wsgireq.req |
|
390 | 362 | res = wsgireq.res |
|
391 | 363 | |
|
392 | 364 | try: |
|
393 | 365 | self.refresh() |
|
394 | 366 | |
|
395 | 367 | csp, nonce = cspvalues(self.ui) |
|
396 | 368 | if csp: |
|
397 | 369 | res.headers['Content-Security-Policy'] = csp |
|
398 | 370 | wsgireq.headers.append(('Content-Security-Policy', csp)) |
|
399 | 371 | |
|
400 | 372 | virtual = wsgireq.env.get("PATH_INFO", "").strip('/') |
|
401 | 373 | tmpl = self.templater(wsgireq, nonce) |
|
402 | 374 | ctype = tmpl('mimetype', encoding=encoding.encoding) |
|
403 | 375 | ctype = templater.stringify(ctype) |
|
404 | 376 | |
|
405 | 377 | # Global defaults. These can be overridden by any handler. |
|
406 | 378 | res.status = '200 Script output follows' |
|
407 | 379 | res.headers['Content-Type'] = ctype |
|
408 | 380 | |
|
409 | 381 | # a static file |
|
410 | 382 | if virtual.startswith('static/') or 'static' in req.qsparams: |
|
411 | 383 | if virtual.startswith('static/'): |
|
412 | 384 | fname = virtual[7:] |
|
413 | 385 | else: |
|
414 | 386 | fname = req.qsparams['static'] |
|
415 | 387 | static = self.ui.config("web", "static", None, |
|
416 | 388 | untrusted=False) |
|
417 | 389 | if not static: |
|
418 | 390 | tp = self.templatepath or templater.templatepaths() |
|
419 | 391 | if isinstance(tp, str): |
|
420 | 392 | tp = [tp] |
|
421 | 393 | static = [os.path.join(p, 'static') for p in tp] |
|
422 | 394 | |
|
423 | 395 | staticfile(static, fname, res) |
|
424 | 396 | return res.sendresponse() |
|
425 | 397 | |
|
426 | 398 | # top-level index |
|
427 | 399 | |
|
428 | 400 | repos = dict(self.repos) |
|
429 | 401 | |
|
430 | 402 | if (not virtual or virtual == 'index') and virtual not in repos: |
|
431 | 403 | wsgireq.respond(HTTP_OK, ctype) |
|
432 | 404 | return self.makeindex(wsgireq, tmpl) |
|
433 | 405 | |
|
434 | 406 | # nested indexes and hgwebs |
|
435 | 407 | |
|
436 | 408 | if virtual.endswith('/index') and virtual not in repos: |
|
437 | 409 | subdir = virtual[:-len('index')] |
|
438 | 410 | if any(r.startswith(subdir) for r in repos): |
|
439 | 411 | wsgireq.respond(HTTP_OK, ctype) |
|
440 | 412 | return self.makeindex(wsgireq, tmpl, subdir) |
|
441 | 413 | |
|
442 | 414 | def _virtualdirs(): |
|
443 | 415 | # Check the full virtual path, each parent, and the root ('') |
|
444 | 416 | if virtual != '': |
|
445 | 417 | yield virtual |
|
446 | 418 | |
|
447 | 419 | for p in util.finddirs(virtual): |
|
448 | 420 | yield p |
|
449 | 421 | |
|
450 | 422 | yield '' |
|
451 | 423 | |
|
452 | 424 | for virtualrepo in _virtualdirs(): |
|
453 | 425 | real = repos.get(virtualrepo) |
|
454 | 426 | if real: |
|
455 | 427 | # Re-parse the WSGI environment to take into account our |
|
456 | 428 | # repository path component. |
|
457 | 429 | wsgireq.req = requestmod.parserequestfromenv( |
|
458 |
wsgireq.env, wsgireq.req.bodyfh, reponame=virtualrepo |
|
|
430 | wsgireq.env, wsgireq.req.bodyfh, reponame=virtualrepo, | |
|
431 | altbaseurl=self.ui.config('web', 'baseurl')) | |
|
459 | 432 | try: |
|
460 | 433 | # ensure caller gets private copy of ui |
|
461 | 434 | repo = hg.repository(self.ui.copy(), real) |
|
462 | 435 | return hgweb_mod.hgweb(repo).run_wsgi(wsgireq) |
|
463 | 436 | except IOError as inst: |
|
464 | 437 | msg = encoding.strtolocal(inst.strerror) |
|
465 | 438 | raise ErrorResponse(HTTP_SERVER_ERROR, msg) |
|
466 | 439 | except error.RepoError as inst: |
|
467 | 440 | raise ErrorResponse(HTTP_SERVER_ERROR, bytes(inst)) |
|
468 | 441 | |
|
469 | 442 | # browse subdirectories |
|
470 | 443 | subdir = virtual + '/' |
|
471 | 444 | if [r for r in repos if r.startswith(subdir)]: |
|
472 | 445 | wsgireq.respond(HTTP_OK, ctype) |
|
473 | 446 | return self.makeindex(wsgireq, tmpl, subdir) |
|
474 | 447 | |
|
475 | 448 | # prefixes not found |
|
476 | 449 | wsgireq.respond(HTTP_NOT_FOUND, ctype) |
|
477 | 450 | return tmpl("notfound", repo=virtual) |
|
478 | 451 | |
|
479 | 452 | except ErrorResponse as err: |
|
480 | 453 | wsgireq.respond(err, ctype) |
|
481 | 454 | return tmpl('error', error=err.message or '') |
|
482 | 455 | finally: |
|
483 | 456 | tmpl = None |
|
484 | 457 | |
|
485 | 458 | def makeindex(self, wsgireq, tmpl, subdir=""): |
|
486 | 459 | req = wsgireq.req |
|
487 | 460 | |
|
488 | 461 | self.refresh() |
|
489 | 462 | sortable = ["name", "description", "contact", "lastchange"] |
|
490 | 463 | sortcolumn, descending = None, False |
|
491 | 464 | if 'sort' in req.qsparams: |
|
492 | 465 | sortcolumn = req.qsparams['sort'] |
|
493 | 466 | descending = sortcolumn.startswith('-') |
|
494 | 467 | if descending: |
|
495 | 468 | sortcolumn = sortcolumn[1:] |
|
496 | 469 | if sortcolumn not in sortable: |
|
497 | 470 | sortcolumn = "" |
|
498 | 471 | |
|
499 | 472 | sort = [("sort_%s" % column, |
|
500 | 473 | "%s%s" % ((not descending and column == sortcolumn) |
|
501 | 474 | and "-" or "", column)) |
|
502 | 475 | for column in sortable] |
|
503 | 476 | |
|
504 | 477 | self.refresh() |
|
505 | self.updatereqenv(wsgireq.env) | |
|
506 | 478 | |
|
507 | 479 | entries = indexentries(self.ui, self.repos, wsgireq, req, |
|
508 | 480 | self.stripecount, sortcolumn=sortcolumn, |
|
509 | 481 | descending=descending, subdir=subdir) |
|
510 | 482 | |
|
511 | 483 | return tmpl("index", entries=entries, subdir=subdir, |
|
512 | 484 | pathdef=hgweb_mod.makebreadcrumb('/' + subdir, self.prefix), |
|
513 | 485 | sortcolumn=sortcolumn, descending=descending, |
|
514 | 486 | **dict(sort)) |
|
515 | 487 | |
|
516 | 488 | def templater(self, wsgireq, nonce): |
|
517 | 489 | |
|
518 | 490 | def motd(**map): |
|
519 | 491 | if self.motd is not None: |
|
520 | 492 | yield self.motd |
|
521 | 493 | else: |
|
522 | 494 | yield config('web', 'motd') |
|
523 | 495 | |
|
524 | 496 | def config(section, name, default=uimod._unset, untrusted=True): |
|
525 | 497 | return self.ui.config(section, name, default, untrusted) |
|
526 | 498 | |
|
527 | self.updatereqenv(wsgireq.env) | |
|
528 | ||
|
529 | 499 | url = wsgireq.env.get('SCRIPT_NAME', '') |
|
530 | 500 | if not url.endswith('/'): |
|
531 | 501 | url += '/' |
|
532 | 502 | |
|
533 | 503 | vars = {} |
|
534 | 504 | styles, (style, mapfile) = hgweb_mod.getstyle(wsgireq.req, config, |
|
535 | 505 | self.templatepath) |
|
536 | 506 | if style == styles[0]: |
|
537 | 507 | vars['style'] = style |
|
538 | 508 | |
|
539 | 509 | sessionvars = webutil.sessionvars(vars, r'?') |
|
540 | 510 | logourl = config('web', 'logourl') |
|
541 | 511 | logoimg = config('web', 'logoimg') |
|
542 | 512 | staticurl = (config('web', 'staticurl') |
|
543 | 513 | or wsgireq.req.apppath + '/static/') |
|
544 | 514 | if not staticurl.endswith('/'): |
|
545 | 515 | staticurl += '/' |
|
546 | 516 | |
|
547 | 517 | defaults = { |
|
548 | 518 | "encoding": encoding.encoding, |
|
549 | 519 | "motd": motd, |
|
550 | 520 | "url": url, |
|
551 | 521 | "logourl": logourl, |
|
552 | 522 | "logoimg": logoimg, |
|
553 | 523 | "staticurl": staticurl, |
|
554 | 524 | "sessionvars": sessionvars, |
|
555 | 525 | "style": style, |
|
556 | 526 | "nonce": nonce, |
|
557 | 527 | } |
|
558 | 528 | tmpl = templater.templater.frommapfile(mapfile, defaults=defaults) |
|
559 | 529 | return tmpl |
|
560 | ||
|
561 | def updatereqenv(self, env): | |
|
562 | if self._baseurl is not None: | |
|
563 | name, port, path = geturlcgivars(self._baseurl, env['SERVER_PORT']) | |
|
564 | env['SERVER_NAME'] = name | |
|
565 | env['SERVER_PORT'] = port | |
|
566 | env['SCRIPT_NAME'] = path |
@@ -1,669 +1,713 | |||
|
1 | 1 | # hgweb/request.py - An http request from either CGI or the standalone server. |
|
2 | 2 | # |
|
3 | 3 | # Copyright 21 May 2005 - (c) 2005 Jake Edge <jake@edge2.net> |
|
4 | 4 | # Copyright 2005, 2006 Matt Mackall <mpm@selenic.com> |
|
5 | 5 | # |
|
6 | 6 | # This software may be used and distributed according to the terms of the |
|
7 | 7 | # GNU General Public License version 2 or any later version. |
|
8 | 8 | |
|
9 | 9 | from __future__ import absolute_import |
|
10 | 10 | |
|
11 | 11 | import errno |
|
12 | 12 | import socket |
|
13 | 13 | import wsgiref.headers as wsgiheaders |
|
14 | 14 | #import wsgiref.validate |
|
15 | 15 | |
|
16 | 16 | from .common import ( |
|
17 | 17 | ErrorResponse, |
|
18 | 18 | statusmessage, |
|
19 | 19 | ) |
|
20 | 20 | |
|
21 | 21 | from ..thirdparty import ( |
|
22 | 22 | attr, |
|
23 | 23 | ) |
|
24 | 24 | from .. import ( |
|
25 | 25 | error, |
|
26 | 26 | pycompat, |
|
27 | 27 | util, |
|
28 | 28 | ) |
|
29 | 29 | |
|
30 | 30 | class multidict(object): |
|
31 | 31 | """A dict like object that can store multiple values for a key. |
|
32 | 32 | |
|
33 | 33 | Used to store parsed request parameters. |
|
34 | 34 | |
|
35 | 35 | This is inspired by WebOb's class of the same name. |
|
36 | 36 | """ |
|
37 | 37 | def __init__(self): |
|
38 | 38 | # Stores (key, value) 2-tuples. This isn't the most efficient. But we |
|
39 | 39 | # don't rely on parameters that much, so it shouldn't be a perf issue. |
|
40 | 40 | # we can always add dict for fast lookups. |
|
41 | 41 | self._items = [] |
|
42 | 42 | |
|
43 | 43 | def __getitem__(self, key): |
|
44 | 44 | """Returns the last set value for a key.""" |
|
45 | 45 | for k, v in reversed(self._items): |
|
46 | 46 | if k == key: |
|
47 | 47 | return v |
|
48 | 48 | |
|
49 | 49 | raise KeyError(key) |
|
50 | 50 | |
|
51 | 51 | def __setitem__(self, key, value): |
|
52 | 52 | """Replace a values for a key with a new value.""" |
|
53 | 53 | try: |
|
54 | 54 | del self[key] |
|
55 | 55 | except KeyError: |
|
56 | 56 | pass |
|
57 | 57 | |
|
58 | 58 | self._items.append((key, value)) |
|
59 | 59 | |
|
60 | 60 | def __delitem__(self, key): |
|
61 | 61 | """Delete all values for a key.""" |
|
62 | 62 | oldlen = len(self._items) |
|
63 | 63 | |
|
64 | 64 | self._items[:] = [(k, v) for k, v in self._items if k != key] |
|
65 | 65 | |
|
66 | 66 | if oldlen == len(self._items): |
|
67 | 67 | raise KeyError(key) |
|
68 | 68 | |
|
69 | 69 | def __contains__(self, key): |
|
70 | 70 | return any(k == key for k, v in self._items) |
|
71 | 71 | |
|
72 | 72 | def __len__(self): |
|
73 | 73 | return len(self._items) |
|
74 | 74 | |
|
75 | 75 | def get(self, key, default=None): |
|
76 | 76 | try: |
|
77 | 77 | return self.__getitem__(key) |
|
78 | 78 | except KeyError: |
|
79 | 79 | return default |
|
80 | 80 | |
|
81 | 81 | def add(self, key, value): |
|
82 | 82 | """Add a new value for a key. Does not replace existing values.""" |
|
83 | 83 | self._items.append((key, value)) |
|
84 | 84 | |
|
85 | 85 | def getall(self, key): |
|
86 | 86 | """Obtains all values for a key.""" |
|
87 | 87 | return [v for k, v in self._items if k == key] |
|
88 | 88 | |
|
89 | 89 | def getone(self, key): |
|
90 | 90 | """Obtain a single value for a key. |
|
91 | 91 | |
|
92 | 92 | Raises KeyError if key not defined or it has multiple values set. |
|
93 | 93 | """ |
|
94 | 94 | vals = self.getall(key) |
|
95 | 95 | |
|
96 | 96 | if not vals: |
|
97 | 97 | raise KeyError(key) |
|
98 | 98 | |
|
99 | 99 | if len(vals) > 1: |
|
100 | 100 | raise KeyError('multiple values for %r' % key) |
|
101 | 101 | |
|
102 | 102 | return vals[0] |
|
103 | 103 | |
|
104 | 104 | def asdictoflists(self): |
|
105 | 105 | d = {} |
|
106 | 106 | for k, v in self._items: |
|
107 | 107 | if k in d: |
|
108 | 108 | d[k].append(v) |
|
109 | 109 | else: |
|
110 | 110 | d[k] = [v] |
|
111 | 111 | |
|
112 | 112 | return d |
|
113 | 113 | |
|
114 | 114 | @attr.s(frozen=True) |
|
115 | 115 | class parsedrequest(object): |
|
116 | 116 | """Represents a parsed WSGI request. |
|
117 | 117 | |
|
118 | 118 | Contains both parsed parameters as well as a handle on the input stream. |
|
119 | 119 | """ |
|
120 | 120 | |
|
121 | 121 | # Request method. |
|
122 | 122 | method = attr.ib() |
|
123 | 123 | # Full URL for this request. |
|
124 | 124 | url = attr.ib() |
|
125 | 125 | # URL without any path components. Just <proto>://<host><port>. |
|
126 | 126 | baseurl = attr.ib() |
|
127 | 127 | # Advertised URL. Like ``url`` and ``baseurl`` but uses SERVER_NAME instead |
|
128 | 128 | # of HTTP: Host header for hostname. This is likely what clients used. |
|
129 | 129 | advertisedurl = attr.ib() |
|
130 | 130 | advertisedbaseurl = attr.ib() |
|
131 | 131 | # URL scheme (part before ``://``). e.g. ``http`` or ``https``. |
|
132 | 132 | urlscheme = attr.ib() |
|
133 | 133 | # Value of REMOTE_USER, if set, or None. |
|
134 | 134 | remoteuser = attr.ib() |
|
135 | 135 | # Value of REMOTE_HOST, if set, or None. |
|
136 | 136 | remotehost = attr.ib() |
|
137 | 137 | # Relative WSGI application path. If defined, will begin with a |
|
138 | 138 | # ``/``. |
|
139 | 139 | apppath = attr.ib() |
|
140 | 140 | # List of path parts to be used for dispatch. |
|
141 | 141 | dispatchparts = attr.ib() |
|
142 | 142 | # URL path component (no query string) used for dispatch. Can be |
|
143 | 143 | # ``None`` to signal no path component given to the request, an |
|
144 | 144 | # empty string to signal a request to the application's root URL, |
|
145 | 145 | # or a string not beginning with ``/`` containing the requested |
|
146 | 146 | # path under the application. |
|
147 | 147 | dispatchpath = attr.ib() |
|
148 | 148 | # The name of the repository being accessed. |
|
149 | 149 | reponame = attr.ib() |
|
150 | 150 | # Raw query string (part after "?" in URL). |
|
151 | 151 | querystring = attr.ib() |
|
152 | 152 | # multidict of query string parameters. |
|
153 | 153 | qsparams = attr.ib() |
|
154 | 154 | # wsgiref.headers.Headers instance. Operates like a dict with case |
|
155 | 155 | # insensitive keys. |
|
156 | 156 | headers = attr.ib() |
|
157 | 157 | # Request body input stream. |
|
158 | 158 | bodyfh = attr.ib() |
|
159 | 159 | |
|
160 | def parserequestfromenv(env, bodyfh, reponame=None): | |
|
160 | def parserequestfromenv(env, bodyfh, reponame=None, altbaseurl=None): | |
|
161 | 161 | """Parse URL components from environment variables. |
|
162 | 162 | |
|
163 | 163 | WSGI defines request attributes via environment variables. This function |
|
164 | 164 | parses the environment variables into a data structure. |
|
165 | 165 | |
|
166 | 166 | If ``reponame`` is defined, the leading path components matching that |
|
167 | 167 | string are effectively shifted from ``PATH_INFO`` to ``SCRIPT_NAME``. |
|
168 | 168 | This simulates the world view of a WSGI application that processes |
|
169 | 169 | requests from the base URL of a repo. |
|
170 | ||
|
171 | If ``altbaseurl`` (typically comes from ``web.baseurl`` config option) | |
|
172 | is defined, it is used - instead of the WSGI environment variables - for | |
|
173 | constructing URL components up to and including the WSGI application path. | |
|
174 | For example, if the current WSGI application is at ``/repo`` and a request | |
|
175 | is made to ``/rev/@`` with this argument set to | |
|
176 | ``http://myserver:9000/prefix``, the URL and path components will resolve as | |
|
177 | if the request were to ``http://myserver:9000/prefix/rev/@``. In other | |
|
178 | words, ``wsgi.url_scheme``, ``SERVER_NAME``, ``SERVER_PORT``, and | |
|
179 | ``SCRIPT_NAME`` are all effectively replaced by components from this URL. | |
|
170 | 180 | """ |
|
171 |
# PEP |
|
|
181 | # PEP 3333 defines the WSGI spec and is a useful reference for this code. | |
|
172 | 182 | |
|
173 | 183 | # We first validate that the incoming object conforms with the WSGI spec. |
|
174 | 184 | # We only want to be dealing with spec-conforming WSGI implementations. |
|
175 | 185 | # TODO enable this once we fix internal violations. |
|
176 | 186 | #wsgiref.validate.check_environ(env) |
|
177 | 187 | |
|
178 | 188 | # PEP-0333 states that environment keys and values are native strings |
|
179 | 189 | # (bytes on Python 2 and str on Python 3). The code points for the Unicode |
|
180 | 190 | # strings on Python 3 must be between \00000-\000FF. We deal with bytes |
|
181 | 191 | # in Mercurial, so mass convert string keys and values to bytes. |
|
182 | 192 | if pycompat.ispy3: |
|
183 | 193 | env = {k.encode('latin-1'): v for k, v in env.iteritems()} |
|
184 | 194 | env = {k: v.encode('latin-1') if isinstance(v, str) else v |
|
185 | 195 | for k, v in env.iteritems()} |
|
186 | 196 | |
|
197 | if altbaseurl: | |
|
198 | altbaseurl = util.url(altbaseurl) | |
|
199 | ||
|
187 | 200 | # https://www.python.org/dev/peps/pep-0333/#environ-variables defines |
|
188 | 201 | # the environment variables. |
|
189 | 202 | # https://www.python.org/dev/peps/pep-0333/#url-reconstruction defines |
|
190 | 203 | # how URLs are reconstructed. |
|
191 | 204 | fullurl = env['wsgi.url_scheme'] + '://' |
|
192 | advertisedfullurl = fullurl | |
|
205 | ||
|
206 | if altbaseurl and altbaseurl.scheme: | |
|
207 | advertisedfullurl = altbaseurl.scheme + '://' | |
|
208 | else: | |
|
209 | advertisedfullurl = fullurl | |
|
193 | 210 | |
|
194 | def addport(s): | |
|
195 | if env['wsgi.url_scheme'] == 'https': | |
|
196 |
if |
|
|
197 |
s += ':' + |
|
|
211 | def addport(s, port): | |
|
212 | if s.startswith('https://'): | |
|
213 | if port != '443': | |
|
214 | s += ':' + port | |
|
198 | 215 | else: |
|
199 |
if |
|
|
200 |
s += ':' + |
|
|
216 | if port != '80': | |
|
217 | s += ':' + port | |
|
201 | 218 | |
|
202 | 219 | return s |
|
203 | 220 | |
|
204 | 221 | if env.get('HTTP_HOST'): |
|
205 | 222 | fullurl += env['HTTP_HOST'] |
|
206 | 223 | else: |
|
207 | 224 | fullurl += env['SERVER_NAME'] |
|
208 | fullurl = addport(fullurl) | |
|
225 | fullurl = addport(fullurl, env['SERVER_PORT']) | |
|
226 | ||
|
227 | if altbaseurl and altbaseurl.host: | |
|
228 | advertisedfullurl += altbaseurl.host | |
|
209 | 229 | |
|
210 | advertisedfullurl += env['SERVER_NAME'] | |
|
211 | advertisedfullurl = addport(advertisedfullurl) | |
|
230 | if altbaseurl.port: | |
|
231 | port = altbaseurl.port | |
|
232 | elif altbaseurl.scheme == 'http' and not altbaseurl.port: | |
|
233 | port = '80' | |
|
234 | elif altbaseurl.scheme == 'https' and not altbaseurl.port: | |
|
235 | port = '443' | |
|
236 | else: | |
|
237 | port = env['SERVER_PORT'] | |
|
238 | ||
|
239 | advertisedfullurl = addport(advertisedfullurl, port) | |
|
240 | else: | |
|
241 | advertisedfullurl += env['SERVER_NAME'] | |
|
242 | advertisedfullurl = addport(advertisedfullurl, env['SERVER_PORT']) | |
|
212 | 243 | |
|
213 | 244 | baseurl = fullurl |
|
214 | 245 | advertisedbaseurl = advertisedfullurl |
|
215 | 246 | |
|
216 | 247 | fullurl += util.urlreq.quote(env.get('SCRIPT_NAME', '')) |
|
217 | advertisedfullurl += util.urlreq.quote(env.get('SCRIPT_NAME', '')) | |
|
218 | 248 | fullurl += util.urlreq.quote(env.get('PATH_INFO', '')) |
|
249 | ||
|
250 | if altbaseurl: | |
|
251 | path = altbaseurl.path or '' | |
|
252 | if path and not path.startswith('/'): | |
|
253 | path = '/' + path | |
|
254 | advertisedfullurl += util.urlreq.quote(path) | |
|
255 | else: | |
|
256 | advertisedfullurl += util.urlreq.quote(env.get('SCRIPT_NAME', '')) | |
|
257 | ||
|
219 | 258 | advertisedfullurl += util.urlreq.quote(env.get('PATH_INFO', '')) |
|
220 | 259 | |
|
221 | 260 | if env.get('QUERY_STRING'): |
|
222 | 261 | fullurl += '?' + env['QUERY_STRING'] |
|
223 | 262 | advertisedfullurl += '?' + env['QUERY_STRING'] |
|
224 | 263 | |
|
225 | 264 | # If ``reponame`` is defined, that must be a prefix on PATH_INFO |
|
226 | 265 | # that represents the repository being dispatched to. When computing |
|
227 | 266 | # the dispatch info, we ignore these leading path components. |
|
228 | 267 | |
|
229 | apppath = env.get('SCRIPT_NAME', '') | |
|
268 | if altbaseurl: | |
|
269 | apppath = altbaseurl.path or '' | |
|
270 | if apppath and not apppath.startswith('/'): | |
|
271 | apppath = '/' + apppath | |
|
272 | else: | |
|
273 | apppath = env.get('SCRIPT_NAME', '') | |
|
230 | 274 | |
|
231 | 275 | if reponame: |
|
232 | 276 | repoprefix = '/' + reponame.strip('/') |
|
233 | 277 | |
|
234 | 278 | if not env.get('PATH_INFO'): |
|
235 | 279 | raise error.ProgrammingError('reponame requires PATH_INFO') |
|
236 | 280 | |
|
237 | 281 | if not env['PATH_INFO'].startswith(repoprefix): |
|
238 | 282 | raise error.ProgrammingError('PATH_INFO does not begin with repo ' |
|
239 | 283 | 'name: %s (%s)' % (env['PATH_INFO'], |
|
240 | 284 | reponame)) |
|
241 | 285 | |
|
242 | 286 | dispatchpath = env['PATH_INFO'][len(repoprefix):] |
|
243 | 287 | |
|
244 | 288 | if dispatchpath and not dispatchpath.startswith('/'): |
|
245 | 289 | raise error.ProgrammingError('reponame prefix of PATH_INFO does ' |
|
246 | 290 | 'not end at path delimiter: %s (%s)' % |
|
247 | 291 | (env['PATH_INFO'], reponame)) |
|
248 | 292 | |
|
249 | 293 | apppath = apppath.rstrip('/') + repoprefix |
|
250 | 294 | dispatchparts = dispatchpath.strip('/').split('/') |
|
251 | 295 | dispatchpath = '/'.join(dispatchparts) |
|
252 | 296 | |
|
253 | 297 | elif 'PATH_INFO' in env: |
|
254 | 298 | if env['PATH_INFO'].strip('/'): |
|
255 | 299 | dispatchparts = env['PATH_INFO'].strip('/').split('/') |
|
256 | 300 | dispatchpath = '/'.join(dispatchparts) |
|
257 | 301 | else: |
|
258 | 302 | dispatchparts = [] |
|
259 | 303 | dispatchpath = '' |
|
260 | 304 | else: |
|
261 | 305 | dispatchparts = [] |
|
262 | 306 | dispatchpath = None |
|
263 | 307 | |
|
264 | 308 | querystring = env.get('QUERY_STRING', '') |
|
265 | 309 | |
|
266 | 310 | # We store as a list so we have ordering information. We also store as |
|
267 | 311 | # a dict to facilitate fast lookup. |
|
268 | 312 | qsparams = multidict() |
|
269 | 313 | for k, v in util.urlreq.parseqsl(querystring, keep_blank_values=True): |
|
270 | 314 | qsparams.add(k, v) |
|
271 | 315 | |
|
272 | 316 | # HTTP_* keys contain HTTP request headers. The Headers structure should |
|
273 | 317 | # perform case normalization for us. We just rewrite underscore to dash |
|
274 | 318 | # so keys match what likely went over the wire. |
|
275 | 319 | headers = [] |
|
276 | 320 | for k, v in env.iteritems(): |
|
277 | 321 | if k.startswith('HTTP_'): |
|
278 | 322 | headers.append((k[len('HTTP_'):].replace('_', '-'), v)) |
|
279 | 323 | |
|
280 | 324 | headers = wsgiheaders.Headers(headers) |
|
281 | 325 | |
|
282 | 326 | # This is kind of a lie because the HTTP header wasn't explicitly |
|
283 | 327 | # sent. But for all intents and purposes it should be OK to lie about |
|
284 | 328 | # this, since a consumer will either either value to determine how many |
|
285 | 329 | # bytes are available to read. |
|
286 | 330 | if 'CONTENT_LENGTH' in env and 'HTTP_CONTENT_LENGTH' not in env: |
|
287 | 331 | headers['Content-Length'] = env['CONTENT_LENGTH'] |
|
288 | 332 | |
|
289 | 333 | # TODO do this once we remove wsgirequest.inp, otherwise we could have |
|
290 | 334 | # multiple readers from the underlying input stream. |
|
291 | 335 | #bodyfh = env['wsgi.input'] |
|
292 | 336 | #if 'Content-Length' in headers: |
|
293 | 337 | # bodyfh = util.cappedreader(bodyfh, int(headers['Content-Length'])) |
|
294 | 338 | |
|
295 | 339 | return parsedrequest(method=env['REQUEST_METHOD'], |
|
296 | 340 | url=fullurl, baseurl=baseurl, |
|
297 | 341 | advertisedurl=advertisedfullurl, |
|
298 | 342 | advertisedbaseurl=advertisedbaseurl, |
|
299 | 343 | urlscheme=env['wsgi.url_scheme'], |
|
300 | 344 | remoteuser=env.get('REMOTE_USER'), |
|
301 | 345 | remotehost=env.get('REMOTE_HOST'), |
|
302 | 346 | apppath=apppath, |
|
303 | 347 | dispatchparts=dispatchparts, dispatchpath=dispatchpath, |
|
304 | 348 | reponame=reponame, |
|
305 | 349 | querystring=querystring, |
|
306 | 350 | qsparams=qsparams, |
|
307 | 351 | headers=headers, |
|
308 | 352 | bodyfh=bodyfh) |
|
309 | 353 | |
|
310 | 354 | class offsettrackingwriter(object): |
|
311 | 355 | """A file object like object that is append only and tracks write count. |
|
312 | 356 | |
|
313 | 357 | Instances are bound to a callable. This callable is called with data |
|
314 | 358 | whenever a ``write()`` is attempted. |
|
315 | 359 | |
|
316 | 360 | Instances track the amount of written data so they can answer ``tell()`` |
|
317 | 361 | requests. |
|
318 | 362 | |
|
319 | 363 | The intent of this class is to wrap the ``write()`` function returned by |
|
320 | 364 | a WSGI ``start_response()`` function. Since ``write()`` is a callable and |
|
321 | 365 | not a file object, it doesn't implement other file object methods. |
|
322 | 366 | """ |
|
323 | 367 | def __init__(self, writefn): |
|
324 | 368 | self._write = writefn |
|
325 | 369 | self._offset = 0 |
|
326 | 370 | |
|
327 | 371 | def write(self, s): |
|
328 | 372 | res = self._write(s) |
|
329 | 373 | # Some Python objects don't report the number of bytes written. |
|
330 | 374 | if res is None: |
|
331 | 375 | self._offset += len(s) |
|
332 | 376 | else: |
|
333 | 377 | self._offset += res |
|
334 | 378 | |
|
335 | 379 | def flush(self): |
|
336 | 380 | pass |
|
337 | 381 | |
|
338 | 382 | def tell(self): |
|
339 | 383 | return self._offset |
|
340 | 384 | |
|
341 | 385 | class wsgiresponse(object): |
|
342 | 386 | """Represents a response to a WSGI request. |
|
343 | 387 | |
|
344 | 388 | A response consists of a status line, headers, and a body. |
|
345 | 389 | |
|
346 | 390 | Consumers must populate the ``status`` and ``headers`` fields and |
|
347 | 391 | make a call to a ``setbody*()`` method before the response can be |
|
348 | 392 | issued. |
|
349 | 393 | |
|
350 | 394 | When it is time to start sending the response over the wire, |
|
351 | 395 | ``sendresponse()`` is called. It handles emitting the header portion |
|
352 | 396 | of the response message. It then yields chunks of body data to be |
|
353 | 397 | written to the peer. Typically, the WSGI application itself calls |
|
354 | 398 | and returns the value from ``sendresponse()``. |
|
355 | 399 | """ |
|
356 | 400 | |
|
357 | 401 | def __init__(self, req, startresponse): |
|
358 | 402 | """Create an empty response tied to a specific request. |
|
359 | 403 | |
|
360 | 404 | ``req`` is a ``parsedrequest``. ``startresponse`` is the |
|
361 | 405 | ``start_response`` function passed to the WSGI application. |
|
362 | 406 | """ |
|
363 | 407 | self._req = req |
|
364 | 408 | self._startresponse = startresponse |
|
365 | 409 | |
|
366 | 410 | self.status = None |
|
367 | 411 | self.headers = wsgiheaders.Headers([]) |
|
368 | 412 | |
|
369 | 413 | self._bodybytes = None |
|
370 | 414 | self._bodygen = None |
|
371 | 415 | self._bodywillwrite = False |
|
372 | 416 | self._started = False |
|
373 | 417 | self._bodywritefn = None |
|
374 | 418 | |
|
375 | 419 | def _verifybody(self): |
|
376 | 420 | if (self._bodybytes is not None or self._bodygen is not None |
|
377 | 421 | or self._bodywillwrite): |
|
378 | 422 | raise error.ProgrammingError('cannot define body multiple times') |
|
379 | 423 | |
|
380 | 424 | def setbodybytes(self, b): |
|
381 | 425 | """Define the response body as static bytes. |
|
382 | 426 | |
|
383 | 427 | The empty string signals that there is no response body. |
|
384 | 428 | """ |
|
385 | 429 | self._verifybody() |
|
386 | 430 | self._bodybytes = b |
|
387 | 431 | self.headers['Content-Length'] = '%d' % len(b) |
|
388 | 432 | |
|
389 | 433 | def setbodygen(self, gen): |
|
390 | 434 | """Define the response body as a generator of bytes.""" |
|
391 | 435 | self._verifybody() |
|
392 | 436 | self._bodygen = gen |
|
393 | 437 | |
|
394 | 438 | def setbodywillwrite(self): |
|
395 | 439 | """Signal an intent to use write() to emit the response body. |
|
396 | 440 | |
|
397 | 441 | **This is the least preferred way to send a body.** |
|
398 | 442 | |
|
399 | 443 | It is preferred for WSGI applications to emit a generator of chunks |
|
400 | 444 | constituting the response body. However, some consumers can't emit |
|
401 | 445 | data this way. So, WSGI provides a way to obtain a ``write(data)`` |
|
402 | 446 | function that can be used to synchronously perform an unbuffered |
|
403 | 447 | write. |
|
404 | 448 | |
|
405 | 449 | Calling this function signals an intent to produce the body in this |
|
406 | 450 | manner. |
|
407 | 451 | """ |
|
408 | 452 | self._verifybody() |
|
409 | 453 | self._bodywillwrite = True |
|
410 | 454 | |
|
411 | 455 | def sendresponse(self): |
|
412 | 456 | """Send the generated response to the client. |
|
413 | 457 | |
|
414 | 458 | Before this is called, ``status`` must be set and one of |
|
415 | 459 | ``setbodybytes()`` or ``setbodygen()`` must be called. |
|
416 | 460 | |
|
417 | 461 | Calling this method multiple times is not allowed. |
|
418 | 462 | """ |
|
419 | 463 | if self._started: |
|
420 | 464 | raise error.ProgrammingError('sendresponse() called multiple times') |
|
421 | 465 | |
|
422 | 466 | self._started = True |
|
423 | 467 | |
|
424 | 468 | if not self.status: |
|
425 | 469 | raise error.ProgrammingError('status line not defined') |
|
426 | 470 | |
|
427 | 471 | if (self._bodybytes is None and self._bodygen is None |
|
428 | 472 | and not self._bodywillwrite): |
|
429 | 473 | raise error.ProgrammingError('response body not defined') |
|
430 | 474 | |
|
431 | 475 | # RFC 7232 Section 4.1 states that a 304 MUST generate one of |
|
432 | 476 | # {Cache-Control, Content-Location, Date, ETag, Expires, Vary} |
|
433 | 477 | # and SHOULD NOT generate other headers unless they could be used |
|
434 | 478 | # to guide cache updates. Furthermore, RFC 7230 Section 3.3.2 |
|
435 | 479 | # states that no response body can be issued. Content-Length can |
|
436 | 480 | # be sent. But if it is present, it should be the size of the response |
|
437 | 481 | # that wasn't transferred. |
|
438 | 482 | if self.status.startswith('304 '): |
|
439 | 483 | # setbodybytes('') will set C-L to 0. This doesn't conform with the |
|
440 | 484 | # spec. So remove it. |
|
441 | 485 | if self.headers.get('Content-Length') == '0': |
|
442 | 486 | del self.headers['Content-Length'] |
|
443 | 487 | |
|
444 | 488 | # Strictly speaking, this is too strict. But until it causes |
|
445 | 489 | # problems, let's be strict. |
|
446 | 490 | badheaders = {k for k in self.headers.keys() |
|
447 | 491 | if k.lower() not in ('date', 'etag', 'expires', |
|
448 | 492 | 'cache-control', |
|
449 | 493 | 'content-location', |
|
450 | 494 | 'vary')} |
|
451 | 495 | if badheaders: |
|
452 | 496 | raise error.ProgrammingError( |
|
453 | 497 | 'illegal header on 304 response: %s' % |
|
454 | 498 | ', '.join(sorted(badheaders))) |
|
455 | 499 | |
|
456 | 500 | if self._bodygen is not None or self._bodywillwrite: |
|
457 | 501 | raise error.ProgrammingError("must use setbodybytes('') with " |
|
458 | 502 | "304 responses") |
|
459 | 503 | |
|
460 | 504 | # Various HTTP clients (notably httplib) won't read the HTTP response |
|
461 | 505 | # until the HTTP request has been sent in full. If servers (us) send a |
|
462 | 506 | # response before the HTTP request has been fully sent, the connection |
|
463 | 507 | # may deadlock because neither end is reading. |
|
464 | 508 | # |
|
465 | 509 | # We work around this by "draining" the request data before |
|
466 | 510 | # sending any response in some conditions. |
|
467 | 511 | drain = False |
|
468 | 512 | close = False |
|
469 | 513 | |
|
470 | 514 | # If the client sent Expect: 100-continue, we assume it is smart enough |
|
471 | 515 | # to deal with the server sending a response before reading the request. |
|
472 | 516 | # (httplib doesn't do this.) |
|
473 | 517 | if self._req.headers.get('Expect', '').lower() == '100-continue': |
|
474 | 518 | pass |
|
475 | 519 | # Only tend to request methods that have bodies. Strictly speaking, |
|
476 | 520 | # we should sniff for a body. But this is fine for our existing |
|
477 | 521 | # WSGI applications. |
|
478 | 522 | elif self._req.method not in ('POST', 'PUT'): |
|
479 | 523 | pass |
|
480 | 524 | else: |
|
481 | 525 | # If we don't know how much data to read, there's no guarantee |
|
482 | 526 | # that we can drain the request responsibly. The WSGI |
|
483 | 527 | # specification only says that servers *should* ensure the |
|
484 | 528 | # input stream doesn't overrun the actual request. So there's |
|
485 | 529 | # no guarantee that reading until EOF won't corrupt the stream |
|
486 | 530 | # state. |
|
487 | 531 | if not isinstance(self._req.bodyfh, util.cappedreader): |
|
488 | 532 | close = True |
|
489 | 533 | else: |
|
490 | 534 | # We /could/ only drain certain HTTP response codes. But 200 and |
|
491 | 535 | # non-200 wire protocol responses both require draining. Since |
|
492 | 536 | # we have a capped reader in place for all situations where we |
|
493 | 537 | # drain, it is safe to read from that stream. We'll either do |
|
494 | 538 | # a drain or no-op if we're already at EOF. |
|
495 | 539 | drain = True |
|
496 | 540 | |
|
497 | 541 | if close: |
|
498 | 542 | self.headers['Connection'] = 'Close' |
|
499 | 543 | |
|
500 | 544 | if drain: |
|
501 | 545 | assert isinstance(self._req.bodyfh, util.cappedreader) |
|
502 | 546 | while True: |
|
503 | 547 | chunk = self._req.bodyfh.read(32768) |
|
504 | 548 | if not chunk: |
|
505 | 549 | break |
|
506 | 550 | |
|
507 | 551 | write = self._startresponse(pycompat.sysstr(self.status), |
|
508 | 552 | self.headers.items()) |
|
509 | 553 | |
|
510 | 554 | if self._bodybytes: |
|
511 | 555 | yield self._bodybytes |
|
512 | 556 | elif self._bodygen: |
|
513 | 557 | for chunk in self._bodygen: |
|
514 | 558 | yield chunk |
|
515 | 559 | elif self._bodywillwrite: |
|
516 | 560 | self._bodywritefn = write |
|
517 | 561 | else: |
|
518 | 562 | error.ProgrammingError('do not know how to send body') |
|
519 | 563 | |
|
520 | 564 | def getbodyfile(self): |
|
521 | 565 | """Obtain a file object like object representing the response body. |
|
522 | 566 | |
|
523 | 567 | For this to work, you must call ``setbodywillwrite()`` and then |
|
524 | 568 | ``sendresponse()`` first. ``sendresponse()`` is a generator and the |
|
525 | 569 | function won't run to completion unless the generator is advanced. The |
|
526 | 570 | generator yields not items. The easiest way to consume it is with |
|
527 | 571 | ``list(res.sendresponse())``, which should resolve to an empty list - |
|
528 | 572 | ``[]``. |
|
529 | 573 | """ |
|
530 | 574 | if not self._bodywillwrite: |
|
531 | 575 | raise error.ProgrammingError('must call setbodywillwrite() first') |
|
532 | 576 | |
|
533 | 577 | if not self._started: |
|
534 | 578 | raise error.ProgrammingError('must call sendresponse() first; did ' |
|
535 | 579 | 'you remember to consume it since it ' |
|
536 | 580 | 'is a generator?') |
|
537 | 581 | |
|
538 | 582 | assert self._bodywritefn |
|
539 | 583 | return offsettrackingwriter(self._bodywritefn) |
|
540 | 584 | |
|
541 | 585 | class wsgirequest(object): |
|
542 | 586 | """Higher-level API for a WSGI request. |
|
543 | 587 | |
|
544 | 588 | WSGI applications are invoked with 2 arguments. They are used to |
|
545 | 589 | instantiate instances of this class, which provides higher-level APIs |
|
546 | 590 | for obtaining request parameters, writing HTTP output, etc. |
|
547 | 591 | """ |
|
548 | def __init__(self, wsgienv, start_response): | |
|
592 | def __init__(self, wsgienv, start_response, altbaseurl=None): | |
|
549 | 593 | version = wsgienv[r'wsgi.version'] |
|
550 | 594 | if (version < (1, 0)) or (version >= (2, 0)): |
|
551 | 595 | raise RuntimeError("Unknown and unsupported WSGI version %d.%d" |
|
552 | 596 | % version) |
|
553 | 597 | |
|
554 | 598 | inp = wsgienv[r'wsgi.input'] |
|
555 | 599 | |
|
556 | 600 | if r'HTTP_CONTENT_LENGTH' in wsgienv: |
|
557 | 601 | inp = util.cappedreader(inp, int(wsgienv[r'HTTP_CONTENT_LENGTH'])) |
|
558 | 602 | elif r'CONTENT_LENGTH' in wsgienv: |
|
559 | 603 | inp = util.cappedreader(inp, int(wsgienv[r'CONTENT_LENGTH'])) |
|
560 | 604 | |
|
561 | 605 | self.err = wsgienv[r'wsgi.errors'] |
|
562 | 606 | self.threaded = wsgienv[r'wsgi.multithread'] |
|
563 | 607 | self.multiprocess = wsgienv[r'wsgi.multiprocess'] |
|
564 | 608 | self.run_once = wsgienv[r'wsgi.run_once'] |
|
565 | 609 | self.env = wsgienv |
|
566 | self.req = parserequestfromenv(wsgienv, inp) | |
|
610 | self.req = parserequestfromenv(wsgienv, inp, altbaseurl=altbaseurl) | |
|
567 | 611 | self.res = wsgiresponse(self.req, start_response) |
|
568 | 612 | self._start_response = start_response |
|
569 | 613 | self.server_write = None |
|
570 | 614 | self.headers = [] |
|
571 | 615 | |
|
572 | 616 | def respond(self, status, type, filename=None, body=None): |
|
573 | 617 | if not isinstance(type, str): |
|
574 | 618 | type = pycompat.sysstr(type) |
|
575 | 619 | if self._start_response is not None: |
|
576 | 620 | self.headers.append((r'Content-Type', type)) |
|
577 | 621 | if filename: |
|
578 | 622 | filename = (filename.rpartition('/')[-1] |
|
579 | 623 | .replace('\\', '\\\\').replace('"', '\\"')) |
|
580 | 624 | self.headers.append(('Content-Disposition', |
|
581 | 625 | 'inline; filename="%s"' % filename)) |
|
582 | 626 | if body is not None: |
|
583 | 627 | self.headers.append((r'Content-Length', str(len(body)))) |
|
584 | 628 | |
|
585 | 629 | for k, v in self.headers: |
|
586 | 630 | if not isinstance(v, str): |
|
587 | 631 | raise TypeError('header value must be string: %r' % (v,)) |
|
588 | 632 | |
|
589 | 633 | if isinstance(status, ErrorResponse): |
|
590 | 634 | self.headers.extend(status.headers) |
|
591 | 635 | status = statusmessage(status.code, pycompat.bytestr(status)) |
|
592 | 636 | elif status == 200: |
|
593 | 637 | status = '200 Script output follows' |
|
594 | 638 | elif isinstance(status, int): |
|
595 | 639 | status = statusmessage(status) |
|
596 | 640 | |
|
597 | 641 | # Various HTTP clients (notably httplib) won't read the HTTP |
|
598 | 642 | # response until the HTTP request has been sent in full. If servers |
|
599 | 643 | # (us) send a response before the HTTP request has been fully sent, |
|
600 | 644 | # the connection may deadlock because neither end is reading. |
|
601 | 645 | # |
|
602 | 646 | # We work around this by "draining" the request data before |
|
603 | 647 | # sending any response in some conditions. |
|
604 | 648 | drain = False |
|
605 | 649 | close = False |
|
606 | 650 | |
|
607 | 651 | # If the client sent Expect: 100-continue, we assume it is smart |
|
608 | 652 | # enough to deal with the server sending a response before reading |
|
609 | 653 | # the request. (httplib doesn't do this.) |
|
610 | 654 | if self.env.get(r'HTTP_EXPECT', r'').lower() == r'100-continue': |
|
611 | 655 | pass |
|
612 | 656 | # Only tend to request methods that have bodies. Strictly speaking, |
|
613 | 657 | # we should sniff for a body. But this is fine for our existing |
|
614 | 658 | # WSGI applications. |
|
615 | 659 | elif self.env[r'REQUEST_METHOD'] not in (r'POST', r'PUT'): |
|
616 | 660 | pass |
|
617 | 661 | else: |
|
618 | 662 | # If we don't know how much data to read, there's no guarantee |
|
619 | 663 | # that we can drain the request responsibly. The WSGI |
|
620 | 664 | # specification only says that servers *should* ensure the |
|
621 | 665 | # input stream doesn't overrun the actual request. So there's |
|
622 | 666 | # no guarantee that reading until EOF won't corrupt the stream |
|
623 | 667 | # state. |
|
624 | 668 | if not isinstance(self.req.bodyfh, util.cappedreader): |
|
625 | 669 | close = True |
|
626 | 670 | else: |
|
627 | 671 | # We /could/ only drain certain HTTP response codes. But 200 |
|
628 | 672 | # and non-200 wire protocol responses both require draining. |
|
629 | 673 | # Since we have a capped reader in place for all situations |
|
630 | 674 | # where we drain, it is safe to read from that stream. We'll |
|
631 | 675 | # either do a drain or no-op if we're already at EOF. |
|
632 | 676 | drain = True |
|
633 | 677 | |
|
634 | 678 | if close: |
|
635 | 679 | self.headers.append((r'Connection', r'Close')) |
|
636 | 680 | |
|
637 | 681 | if drain: |
|
638 | 682 | assert isinstance(self.req.bodyfh, util.cappedreader) |
|
639 | 683 | while True: |
|
640 | 684 | chunk = self.req.bodyfh.read(32768) |
|
641 | 685 | if not chunk: |
|
642 | 686 | break |
|
643 | 687 | |
|
644 | 688 | self.server_write = self._start_response( |
|
645 | 689 | pycompat.sysstr(status), self.headers) |
|
646 | 690 | self._start_response = None |
|
647 | 691 | self.headers = [] |
|
648 | 692 | if body is not None: |
|
649 | 693 | self.write(body) |
|
650 | 694 | self.server_write = None |
|
651 | 695 | |
|
652 | 696 | def write(self, thing): |
|
653 | 697 | if thing: |
|
654 | 698 | try: |
|
655 | 699 | self.server_write(thing) |
|
656 | 700 | except socket.error as inst: |
|
657 | 701 | if inst[0] != errno.ECONNRESET: |
|
658 | 702 | raise |
|
659 | 703 | |
|
660 | 704 | def flush(self): |
|
661 | 705 | return None |
|
662 | 706 | |
|
663 | 707 | def wsgiapplication(app_maker): |
|
664 | 708 | '''For compatibility with old CGI scripts. A plain hgweb() or hgwebdir() |
|
665 | 709 | can and should now be used as a WSGI application.''' |
|
666 | 710 | application = app_maker() |
|
667 | 711 | def run_wsgi(env, respond): |
|
668 | 712 | return application(env, respond) |
|
669 | 713 | return run_wsgi |
@@ -1,247 +1,416 | |||
|
1 | 1 | from __future__ import absolute_import, print_function |
|
2 | 2 | |
|
3 | 3 | import unittest |
|
4 | 4 | |
|
5 | 5 | from mercurial.hgweb import ( |
|
6 | 6 | request as requestmod, |
|
7 | 7 | ) |
|
8 | 8 | from mercurial import ( |
|
9 | 9 | error, |
|
10 | 10 | ) |
|
11 | 11 | |
|
12 | 12 | DEFAULT_ENV = { |
|
13 | 13 | r'REQUEST_METHOD': r'GET', |
|
14 | 14 | r'SERVER_NAME': r'testserver', |
|
15 | 15 | r'SERVER_PORT': r'80', |
|
16 | 16 | r'SERVER_PROTOCOL': r'http', |
|
17 | 17 | r'wsgi.version': (1, 0), |
|
18 | 18 | r'wsgi.url_scheme': r'http', |
|
19 | 19 | r'wsgi.input': None, |
|
20 | 20 | r'wsgi.errors': None, |
|
21 | 21 | r'wsgi.multithread': False, |
|
22 | 22 | r'wsgi.multiprocess': True, |
|
23 | 23 | r'wsgi.run_once': False, |
|
24 | 24 | } |
|
25 | 25 | |
|
26 | def parse(env, bodyfh=None, reponame=None, extra=None): | |
|
26 | def parse(env, bodyfh=None, reponame=None, altbaseurl=None, extra=None): | |
|
27 | 27 | env = dict(env) |
|
28 | 28 | env.update(extra or {}) |
|
29 | 29 | |
|
30 |
return requestmod.parserequestfromenv(env, bodyfh, reponame=reponame |
|
|
30 | return requestmod.parserequestfromenv(env, bodyfh, reponame=reponame, | |
|
31 | altbaseurl=altbaseurl) | |
|
31 | 32 | |
|
32 | 33 | class ParseRequestTests(unittest.TestCase): |
|
33 | 34 | def testdefault(self): |
|
34 | 35 | r = parse(DEFAULT_ENV) |
|
35 | 36 | self.assertEqual(r.url, b'http://testserver') |
|
36 | 37 | self.assertEqual(r.baseurl, b'http://testserver') |
|
37 | 38 | self.assertEqual(r.advertisedurl, b'http://testserver') |
|
38 | 39 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
39 | 40 | self.assertEqual(r.urlscheme, b'http') |
|
40 | 41 | self.assertEqual(r.method, b'GET') |
|
41 | 42 | self.assertIsNone(r.remoteuser) |
|
42 | 43 | self.assertIsNone(r.remotehost) |
|
43 | 44 | self.assertEqual(r.apppath, b'') |
|
44 | 45 | self.assertEqual(r.dispatchparts, []) |
|
45 | 46 | self.assertIsNone(r.dispatchpath) |
|
46 | 47 | self.assertIsNone(r.reponame) |
|
47 | 48 | self.assertEqual(r.querystring, b'') |
|
48 | 49 | self.assertEqual(len(r.qsparams), 0) |
|
49 | 50 | self.assertEqual(len(r.headers), 0) |
|
50 | 51 | |
|
51 | 52 | def testcustomport(self): |
|
52 | 53 | r = parse(DEFAULT_ENV, extra={ |
|
53 | 54 | r'SERVER_PORT': r'8000', |
|
54 | 55 | }) |
|
55 | 56 | |
|
56 | 57 | self.assertEqual(r.url, b'http://testserver:8000') |
|
57 | 58 | self.assertEqual(r.baseurl, b'http://testserver:8000') |
|
58 | 59 | self.assertEqual(r.advertisedurl, b'http://testserver:8000') |
|
59 | 60 | self.assertEqual(r.advertisedbaseurl, b'http://testserver:8000') |
|
60 | 61 | |
|
61 | 62 | r = parse(DEFAULT_ENV, extra={ |
|
62 | 63 | r'SERVER_PORT': r'4000', |
|
63 | 64 | r'wsgi.url_scheme': r'https', |
|
64 | 65 | }) |
|
65 | 66 | |
|
66 | 67 | self.assertEqual(r.url, b'https://testserver:4000') |
|
67 | 68 | self.assertEqual(r.baseurl, b'https://testserver:4000') |
|
68 | 69 | self.assertEqual(r.advertisedurl, b'https://testserver:4000') |
|
69 | 70 | self.assertEqual(r.advertisedbaseurl, b'https://testserver:4000') |
|
70 | 71 | |
|
71 | 72 | def testhttphost(self): |
|
72 | 73 | r = parse(DEFAULT_ENV, extra={ |
|
73 | 74 | r'HTTP_HOST': r'altserver', |
|
74 | 75 | }) |
|
75 | 76 | |
|
76 | 77 | self.assertEqual(r.url, b'http://altserver') |
|
77 | 78 | self.assertEqual(r.baseurl, b'http://altserver') |
|
78 | 79 | self.assertEqual(r.advertisedurl, b'http://testserver') |
|
79 | 80 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
80 | 81 | |
|
81 | 82 | def testscriptname(self): |
|
82 | 83 | r = parse(DEFAULT_ENV, extra={ |
|
83 | 84 | r'SCRIPT_NAME': r'', |
|
84 | 85 | }) |
|
85 | 86 | |
|
86 | 87 | self.assertEqual(r.url, b'http://testserver') |
|
87 | 88 | self.assertEqual(r.baseurl, b'http://testserver') |
|
88 | 89 | self.assertEqual(r.advertisedurl, b'http://testserver') |
|
89 | 90 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
90 | 91 | self.assertEqual(r.apppath, b'') |
|
91 | 92 | self.assertEqual(r.dispatchparts, []) |
|
92 | 93 | self.assertIsNone(r.dispatchpath) |
|
93 | 94 | |
|
94 | 95 | r = parse(DEFAULT_ENV, extra={ |
|
95 | 96 | r'SCRIPT_NAME': r'/script', |
|
96 | 97 | }) |
|
97 | 98 | |
|
98 | 99 | self.assertEqual(r.url, b'http://testserver/script') |
|
99 | 100 | self.assertEqual(r.baseurl, b'http://testserver') |
|
100 | 101 | self.assertEqual(r.advertisedurl, b'http://testserver/script') |
|
101 | 102 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
102 | 103 | self.assertEqual(r.apppath, b'/script') |
|
103 | 104 | self.assertEqual(r.dispatchparts, []) |
|
104 | 105 | self.assertIsNone(r.dispatchpath) |
|
105 | 106 | |
|
106 | 107 | r = parse(DEFAULT_ENV, extra={ |
|
107 | 108 | r'SCRIPT_NAME': r'/multiple words', |
|
108 | 109 | }) |
|
109 | 110 | |
|
110 | 111 | self.assertEqual(r.url, b'http://testserver/multiple%20words') |
|
111 | 112 | self.assertEqual(r.baseurl, b'http://testserver') |
|
112 | 113 | self.assertEqual(r.advertisedurl, b'http://testserver/multiple%20words') |
|
113 | 114 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
114 | 115 | self.assertEqual(r.apppath, b'/multiple words') |
|
115 | 116 | self.assertEqual(r.dispatchparts, []) |
|
116 | 117 | self.assertIsNone(r.dispatchpath) |
|
117 | 118 | |
|
118 | 119 | def testpathinfo(self): |
|
119 | 120 | r = parse(DEFAULT_ENV, extra={ |
|
120 | 121 | r'PATH_INFO': r'', |
|
121 | 122 | }) |
|
122 | 123 | |
|
123 | 124 | self.assertEqual(r.url, b'http://testserver') |
|
124 | 125 | self.assertEqual(r.baseurl, b'http://testserver') |
|
125 | 126 | self.assertEqual(r.advertisedurl, b'http://testserver') |
|
126 | 127 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
127 | 128 | self.assertEqual(r.apppath, b'') |
|
128 | 129 | self.assertEqual(r.dispatchparts, []) |
|
129 | 130 | self.assertEqual(r.dispatchpath, b'') |
|
130 | 131 | |
|
131 | 132 | r = parse(DEFAULT_ENV, extra={ |
|
132 | 133 | r'PATH_INFO': r'/pathinfo', |
|
133 | 134 | }) |
|
134 | 135 | |
|
135 | 136 | self.assertEqual(r.url, b'http://testserver/pathinfo') |
|
136 | 137 | self.assertEqual(r.baseurl, b'http://testserver') |
|
137 | 138 | self.assertEqual(r.advertisedurl, b'http://testserver/pathinfo') |
|
138 | 139 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
139 | 140 | self.assertEqual(r.apppath, b'') |
|
140 | 141 | self.assertEqual(r.dispatchparts, [b'pathinfo']) |
|
141 | 142 | self.assertEqual(r.dispatchpath, b'pathinfo') |
|
142 | 143 | |
|
143 | 144 | r = parse(DEFAULT_ENV, extra={ |
|
144 | 145 | r'PATH_INFO': r'/one/two/', |
|
145 | 146 | }) |
|
146 | 147 | |
|
147 | 148 | self.assertEqual(r.url, b'http://testserver/one/two/') |
|
148 | 149 | self.assertEqual(r.baseurl, b'http://testserver') |
|
149 | 150 | self.assertEqual(r.advertisedurl, b'http://testserver/one/two/') |
|
150 | 151 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
151 | 152 | self.assertEqual(r.apppath, b'') |
|
152 | 153 | self.assertEqual(r.dispatchparts, [b'one', b'two']) |
|
153 | 154 | self.assertEqual(r.dispatchpath, b'one/two') |
|
154 | 155 | |
|
155 | 156 | def testscriptandpathinfo(self): |
|
156 | 157 | r = parse(DEFAULT_ENV, extra={ |
|
157 | 158 | r'SCRIPT_NAME': r'/script', |
|
158 | 159 | r'PATH_INFO': r'/pathinfo', |
|
159 | 160 | }) |
|
160 | 161 | |
|
161 | 162 | self.assertEqual(r.url, b'http://testserver/script/pathinfo') |
|
162 | 163 | self.assertEqual(r.baseurl, b'http://testserver') |
|
163 | 164 | self.assertEqual(r.advertisedurl, b'http://testserver/script/pathinfo') |
|
164 | 165 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
165 | 166 | self.assertEqual(r.apppath, b'/script') |
|
166 | 167 | self.assertEqual(r.dispatchparts, [b'pathinfo']) |
|
167 | 168 | self.assertEqual(r.dispatchpath, b'pathinfo') |
|
168 | 169 | |
|
169 | 170 | r = parse(DEFAULT_ENV, extra={ |
|
170 | 171 | r'SCRIPT_NAME': r'/script1/script2', |
|
171 | 172 | r'PATH_INFO': r'/path1/path2', |
|
172 | 173 | }) |
|
173 | 174 | |
|
174 | 175 | self.assertEqual(r.url, |
|
175 | 176 | b'http://testserver/script1/script2/path1/path2') |
|
176 | 177 | self.assertEqual(r.baseurl, b'http://testserver') |
|
177 | 178 | self.assertEqual(r.advertisedurl, |
|
178 | 179 | b'http://testserver/script1/script2/path1/path2') |
|
179 | 180 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
180 | 181 | self.assertEqual(r.apppath, b'/script1/script2') |
|
181 | 182 | self.assertEqual(r.dispatchparts, [b'path1', b'path2']) |
|
182 | 183 | self.assertEqual(r.dispatchpath, b'path1/path2') |
|
183 | 184 | |
|
184 | 185 | r = parse(DEFAULT_ENV, extra={ |
|
185 | 186 | r'HTTP_HOST': r'hostserver', |
|
186 | 187 | r'SCRIPT_NAME': r'/script', |
|
187 | 188 | r'PATH_INFO': r'/pathinfo', |
|
188 | 189 | }) |
|
189 | 190 | |
|
190 | 191 | self.assertEqual(r.url, b'http://hostserver/script/pathinfo') |
|
191 | 192 | self.assertEqual(r.baseurl, b'http://hostserver') |
|
192 | 193 | self.assertEqual(r.advertisedurl, b'http://testserver/script/pathinfo') |
|
193 | 194 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
194 | 195 | self.assertEqual(r.apppath, b'/script') |
|
195 | 196 | self.assertEqual(r.dispatchparts, [b'pathinfo']) |
|
196 | 197 | self.assertEqual(r.dispatchpath, b'pathinfo') |
|
197 | 198 | |
|
198 | 199 | def testreponame(self): |
|
199 | 200 | """repository path components get stripped from URL.""" |
|
200 | 201 | |
|
201 | 202 | with self.assertRaisesRegexp(error.ProgrammingError, |
|
202 | 203 | b'reponame requires PATH_INFO'): |
|
203 | 204 | parse(DEFAULT_ENV, reponame=b'repo') |
|
204 | 205 | |
|
205 | 206 | with self.assertRaisesRegexp(error.ProgrammingError, |
|
206 | 207 | b'PATH_INFO does not begin with repo ' |
|
207 | 208 | b'name'): |
|
208 | 209 | parse(DEFAULT_ENV, reponame=b'repo', extra={ |
|
209 | 210 | r'PATH_INFO': r'/pathinfo', |
|
210 | 211 | }) |
|
211 | 212 | |
|
212 | 213 | with self.assertRaisesRegexp(error.ProgrammingError, |
|
213 | 214 | b'reponame prefix of PATH_INFO'): |
|
214 | 215 | parse(DEFAULT_ENV, reponame=b'repo', extra={ |
|
215 | 216 | r'PATH_INFO': r'/repoextra/path', |
|
216 | 217 | }) |
|
217 | 218 | |
|
218 | 219 | r = parse(DEFAULT_ENV, reponame=b'repo', extra={ |
|
219 | 220 | r'PATH_INFO': r'/repo/path1/path2', |
|
220 | 221 | }) |
|
221 | 222 | |
|
222 | 223 | self.assertEqual(r.url, b'http://testserver/repo/path1/path2') |
|
223 | 224 | self.assertEqual(r.baseurl, b'http://testserver') |
|
224 | 225 | self.assertEqual(r.advertisedurl, b'http://testserver/repo/path1/path2') |
|
225 | 226 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
226 | 227 | self.assertEqual(r.apppath, b'/repo') |
|
227 | 228 | self.assertEqual(r.dispatchparts, [b'path1', b'path2']) |
|
228 | 229 | self.assertEqual(r.dispatchpath, b'path1/path2') |
|
229 | 230 | self.assertEqual(r.reponame, b'repo') |
|
230 | 231 | |
|
231 | 232 | r = parse(DEFAULT_ENV, reponame=b'prefix/repo', extra={ |
|
232 | 233 | r'PATH_INFO': r'/prefix/repo/path1/path2', |
|
233 | 234 | }) |
|
234 | 235 | |
|
235 | 236 | self.assertEqual(r.url, b'http://testserver/prefix/repo/path1/path2') |
|
236 | 237 | self.assertEqual(r.baseurl, b'http://testserver') |
|
237 | 238 | self.assertEqual(r.advertisedurl, |
|
238 | 239 | b'http://testserver/prefix/repo/path1/path2') |
|
239 | 240 | self.assertEqual(r.advertisedbaseurl, b'http://testserver') |
|
240 | 241 | self.assertEqual(r.apppath, b'/prefix/repo') |
|
241 | 242 | self.assertEqual(r.dispatchparts, [b'path1', b'path2']) |
|
242 | 243 | self.assertEqual(r.dispatchpath, b'path1/path2') |
|
243 | 244 | self.assertEqual(r.reponame, b'prefix/repo') |
|
244 | 245 | |
|
246 | def testaltbaseurl(self): | |
|
247 | # Simple hostname remap. | |
|
248 | r = parse(DEFAULT_ENV, altbaseurl='http://altserver') | |
|
249 | ||
|
250 | self.assertEqual(r.url, b'http://testserver') | |
|
251 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
252 | self.assertEqual(r.advertisedurl, b'http://altserver') | |
|
253 | self.assertEqual(r.advertisedbaseurl, b'http://altserver') | |
|
254 | self.assertEqual(r.urlscheme, b'http') | |
|
255 | self.assertEqual(r.apppath, b'') | |
|
256 | self.assertEqual(r.dispatchparts, []) | |
|
257 | self.assertIsNone(r.dispatchpath) | |
|
258 | self.assertIsNone(r.reponame) | |
|
259 | ||
|
260 | # With a custom port. | |
|
261 | r = parse(DEFAULT_ENV, altbaseurl='http://altserver:8000') | |
|
262 | self.assertEqual(r.url, b'http://testserver') | |
|
263 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
264 | self.assertEqual(r.advertisedurl, b'http://altserver:8000') | |
|
265 | self.assertEqual(r.advertisedbaseurl, b'http://altserver:8000') | |
|
266 | self.assertEqual(r.urlscheme, b'http') | |
|
267 | self.assertEqual(r.apppath, b'') | |
|
268 | self.assertEqual(r.dispatchparts, []) | |
|
269 | self.assertIsNone(r.dispatchpath) | |
|
270 | self.assertIsNone(r.reponame) | |
|
271 | ||
|
272 | # With a changed protocol. | |
|
273 | r = parse(DEFAULT_ENV, altbaseurl='https://altserver') | |
|
274 | self.assertEqual(r.url, b'http://testserver') | |
|
275 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
276 | self.assertEqual(r.advertisedurl, b'https://altserver') | |
|
277 | self.assertEqual(r.advertisedbaseurl, b'https://altserver') | |
|
278 | # URL scheme is defined as the actual scheme, not advertised. | |
|
279 | self.assertEqual(r.urlscheme, b'http') | |
|
280 | self.assertEqual(r.apppath, b'') | |
|
281 | self.assertEqual(r.dispatchparts, []) | |
|
282 | self.assertIsNone(r.dispatchpath) | |
|
283 | self.assertIsNone(r.reponame) | |
|
284 | ||
|
285 | # Need to specify explicit port number for proper https:// alt URLs. | |
|
286 | r = parse(DEFAULT_ENV, altbaseurl='https://altserver:443') | |
|
287 | self.assertEqual(r.url, b'http://testserver') | |
|
288 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
289 | self.assertEqual(r.advertisedurl, b'https://altserver') | |
|
290 | self.assertEqual(r.advertisedbaseurl, b'https://altserver') | |
|
291 | self.assertEqual(r.urlscheme, b'http') | |
|
292 | self.assertEqual(r.apppath, b'') | |
|
293 | self.assertEqual(r.dispatchparts, []) | |
|
294 | self.assertIsNone(r.dispatchpath) | |
|
295 | self.assertIsNone(r.reponame) | |
|
296 | ||
|
297 | # With only PATH_INFO defined. | |
|
298 | r = parse(DEFAULT_ENV, altbaseurl='http://altserver', extra={ | |
|
299 | r'PATH_INFO': r'/path1/path2', | |
|
300 | }) | |
|
301 | self.assertEqual(r.url, b'http://testserver/path1/path2') | |
|
302 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
303 | self.assertEqual(r.advertisedurl, b'http://altserver/path1/path2') | |
|
304 | self.assertEqual(r.advertisedbaseurl, b'http://altserver') | |
|
305 | self.assertEqual(r.urlscheme, b'http') | |
|
306 | self.assertEqual(r.apppath, b'') | |
|
307 | self.assertEqual(r.dispatchparts, [b'path1', b'path2']) | |
|
308 | self.assertEqual(r.dispatchpath, b'path1/path2') | |
|
309 | self.assertIsNone(r.reponame) | |
|
310 | ||
|
311 | # Path on alt URL. | |
|
312 | r = parse(DEFAULT_ENV, altbaseurl='http://altserver/altpath') | |
|
313 | self.assertEqual(r.url, b'http://testserver') | |
|
314 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
315 | self.assertEqual(r.advertisedurl, b'http://altserver/altpath') | |
|
316 | self.assertEqual(r.advertisedbaseurl, b'http://altserver') | |
|
317 | self.assertEqual(r.urlscheme, b'http') | |
|
318 | self.assertEqual(r.apppath, b'/altpath') | |
|
319 | self.assertEqual(r.dispatchparts, []) | |
|
320 | self.assertIsNone(r.dispatchpath) | |
|
321 | self.assertIsNone(r.reponame) | |
|
322 | ||
|
323 | # With a trailing slash. | |
|
324 | r = parse(DEFAULT_ENV, altbaseurl='http://altserver/altpath/') | |
|
325 | self.assertEqual(r.url, b'http://testserver') | |
|
326 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
327 | self.assertEqual(r.advertisedurl, b'http://altserver/altpath/') | |
|
328 | self.assertEqual(r.advertisedbaseurl, b'http://altserver') | |
|
329 | self.assertEqual(r.urlscheme, b'http') | |
|
330 | self.assertEqual(r.apppath, b'/altpath/') | |
|
331 | self.assertEqual(r.dispatchparts, []) | |
|
332 | self.assertIsNone(r.dispatchpath) | |
|
333 | self.assertIsNone(r.reponame) | |
|
334 | ||
|
335 | # PATH_INFO + path on alt URL. | |
|
336 | r = parse(DEFAULT_ENV, altbaseurl='http://altserver/altpath', extra={ | |
|
337 | r'PATH_INFO': r'/path1/path2', | |
|
338 | }) | |
|
339 | self.assertEqual(r.url, b'http://testserver/path1/path2') | |
|
340 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
341 | self.assertEqual(r.advertisedurl, | |
|
342 | b'http://altserver/altpath/path1/path2') | |
|
343 | self.assertEqual(r.advertisedbaseurl, b'http://altserver') | |
|
344 | self.assertEqual(r.urlscheme, b'http') | |
|
345 | self.assertEqual(r.apppath, b'/altpath') | |
|
346 | self.assertEqual(r.dispatchparts, [b'path1', b'path2']) | |
|
347 | self.assertEqual(r.dispatchpath, b'path1/path2') | |
|
348 | self.assertIsNone(r.reponame) | |
|
349 | ||
|
350 | # PATH_INFO + path on alt URL with trailing slash. | |
|
351 | r = parse(DEFAULT_ENV, altbaseurl='http://altserver/altpath/', extra={ | |
|
352 | r'PATH_INFO': r'/path1/path2', | |
|
353 | }) | |
|
354 | self.assertEqual(r.url, b'http://testserver/path1/path2') | |
|
355 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
356 | self.assertEqual(r.advertisedurl, | |
|
357 | b'http://altserver/altpath//path1/path2') | |
|
358 | self.assertEqual(r.advertisedbaseurl, b'http://altserver') | |
|
359 | self.assertEqual(r.urlscheme, b'http') | |
|
360 | self.assertEqual(r.apppath, b'/altpath/') | |
|
361 | self.assertEqual(r.dispatchparts, [b'path1', b'path2']) | |
|
362 | self.assertEqual(r.dispatchpath, b'path1/path2') | |
|
363 | self.assertIsNone(r.reponame) | |
|
364 | ||
|
365 | # Local SCRIPT_NAME is ignored. | |
|
366 | r = parse(DEFAULT_ENV, altbaseurl='http://altserver', extra={ | |
|
367 | r'SCRIPT_NAME': r'/script', | |
|
368 | r'PATH_INFO': r'/path1/path2', | |
|
369 | }) | |
|
370 | self.assertEqual(r.url, b'http://testserver/script/path1/path2') | |
|
371 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
372 | self.assertEqual(r.advertisedurl, b'http://altserver/path1/path2') | |
|
373 | self.assertEqual(r.advertisedbaseurl, b'http://altserver') | |
|
374 | self.assertEqual(r.urlscheme, b'http') | |
|
375 | self.assertEqual(r.apppath, b'') | |
|
376 | self.assertEqual(r.dispatchparts, [b'path1', b'path2']) | |
|
377 | self.assertEqual(r.dispatchpath, b'path1/path2') | |
|
378 | self.assertIsNone(r.reponame) | |
|
379 | ||
|
380 | # Use remote's path for script name, app path | |
|
381 | r = parse(DEFAULT_ENV, altbaseurl='http://altserver/altroot', extra={ | |
|
382 | r'SCRIPT_NAME': r'/script', | |
|
383 | r'PATH_INFO': r'/path1/path2', | |
|
384 | }) | |
|
385 | self.assertEqual(r.url, b'http://testserver/script/path1/path2') | |
|
386 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
387 | self.assertEqual(r.advertisedurl, | |
|
388 | b'http://altserver/altroot/path1/path2') | |
|
389 | self.assertEqual(r.advertisedbaseurl, b'http://altserver') | |
|
390 | self.assertEqual(r.urlscheme, b'http') | |
|
391 | self.assertEqual(r.apppath, b'/altroot') | |
|
392 | self.assertEqual(r.dispatchparts, [b'path1', b'path2']) | |
|
393 | self.assertEqual(r.dispatchpath, b'path1/path2') | |
|
394 | self.assertIsNone(r.reponame) | |
|
395 | ||
|
396 | # reponame is factored in properly. | |
|
397 | r = parse(DEFAULT_ENV, reponame=b'repo', | |
|
398 | altbaseurl='http://altserver/altroot', | |
|
399 | extra={ | |
|
400 | r'SCRIPT_NAME': r'/script', | |
|
401 | r'PATH_INFO': r'/repo/path1/path2', | |
|
402 | }) | |
|
403 | ||
|
404 | self.assertEqual(r.url, b'http://testserver/script/repo/path1/path2') | |
|
405 | self.assertEqual(r.baseurl, b'http://testserver') | |
|
406 | self.assertEqual(r.advertisedurl, | |
|
407 | b'http://altserver/altroot/repo/path1/path2') | |
|
408 | self.assertEqual(r.advertisedbaseurl, b'http://altserver') | |
|
409 | self.assertEqual(r.apppath, b'/altroot/repo') | |
|
410 | self.assertEqual(r.dispatchparts, [b'path1', b'path2']) | |
|
411 | self.assertEqual(r.dispatchpath, b'path1/path2') | |
|
412 | self.assertEqual(r.reponame, b'repo') | |
|
413 | ||
|
245 | 414 | if __name__ == '__main__': |
|
246 | 415 | import silenttestrunner |
|
247 | 416 | silenttestrunner.main(__name__) |
General Comments 0
You need to be logged in to leave comments.
Login now