Show More
@@ -1,454 +1,452 b'' | |||
|
1 | 1 | # hgweb/hgweb_mod.py - Web interface for a repository. |
|
2 | 2 | # |
|
3 | 3 | # Copyright 21 May 2005 - (c) 2005 Jake Edge <jake@edge2.net> |
|
4 | 4 | # Copyright 2005-2007 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 contextlib |
|
12 | 12 | import os |
|
13 | 13 | |
|
14 | 14 | from .common import ( |
|
15 | 15 | ErrorResponse, |
|
16 | 16 | HTTP_BAD_REQUEST, |
|
17 | 17 | HTTP_NOT_FOUND, |
|
18 | 18 | HTTP_NOT_MODIFIED, |
|
19 | 19 | HTTP_OK, |
|
20 | 20 | HTTP_SERVER_ERROR, |
|
21 | 21 | caching, |
|
22 | 22 | cspvalues, |
|
23 | 23 | permhooks, |
|
24 | 24 | ) |
|
25 | 25 | |
|
26 | 26 | from .. import ( |
|
27 | 27 | encoding, |
|
28 | 28 | error, |
|
29 | 29 | formatter, |
|
30 | 30 | hg, |
|
31 | 31 | hook, |
|
32 | 32 | profiling, |
|
33 | 33 | pycompat, |
|
34 | 34 | repoview, |
|
35 | 35 | templatefilters, |
|
36 | 36 | templater, |
|
37 | 37 | ui as uimod, |
|
38 | 38 | util, |
|
39 | 39 | wireprotoserver, |
|
40 | 40 | ) |
|
41 | 41 | |
|
42 | 42 | from . import ( |
|
43 | 43 | request as requestmod, |
|
44 | 44 | webcommands, |
|
45 | 45 | webutil, |
|
46 | 46 | wsgicgi, |
|
47 | 47 | ) |
|
48 | 48 | |
|
49 | 49 | archivespecs = util.sortdict(( |
|
50 | 50 | ('zip', ('application/zip', 'zip', '.zip', None)), |
|
51 | 51 | ('gz', ('application/x-gzip', 'tgz', '.tar.gz', None)), |
|
52 | 52 | ('bz2', ('application/x-bzip2', 'tbz2', '.tar.bz2', None)), |
|
53 | 53 | )) |
|
54 | 54 | |
|
55 | 55 | def getstyle(req, configfn, templatepath): |
|
56 | 56 | fromreq = req.form.get('style', [None])[0] |
|
57 | 57 | styles = ( |
|
58 | 58 | fromreq, |
|
59 | 59 | configfn('web', 'style'), |
|
60 | 60 | 'paper', |
|
61 | 61 | ) |
|
62 | 62 | return styles, templater.stylemap(styles, templatepath) |
|
63 | 63 | |
|
64 | 64 | def makebreadcrumb(url, prefix=''): |
|
65 | 65 | '''Return a 'URL breadcrumb' list |
|
66 | 66 | |
|
67 | 67 | A 'URL breadcrumb' is a list of URL-name pairs, |
|
68 | 68 | corresponding to each of the path items on a URL. |
|
69 | 69 | This can be used to create path navigation entries. |
|
70 | 70 | ''' |
|
71 | 71 | if url.endswith('/'): |
|
72 | 72 | url = url[:-1] |
|
73 | 73 | if prefix: |
|
74 | 74 | url = '/' + prefix + url |
|
75 | 75 | relpath = url |
|
76 | 76 | if relpath.startswith('/'): |
|
77 | 77 | relpath = relpath[1:] |
|
78 | 78 | |
|
79 | 79 | breadcrumb = [] |
|
80 | 80 | urlel = url |
|
81 | 81 | pathitems = [''] + relpath.split('/') |
|
82 | 82 | for pathel in reversed(pathitems): |
|
83 | 83 | if not pathel or not urlel: |
|
84 | 84 | break |
|
85 | 85 | breadcrumb.append({'url': urlel, 'name': pathel}) |
|
86 | 86 | urlel = os.path.dirname(urlel) |
|
87 | 87 | return reversed(breadcrumb) |
|
88 | 88 | |
|
89 | 89 | class requestcontext(object): |
|
90 | 90 | """Holds state/context for an individual request. |
|
91 | 91 | |
|
92 | 92 | Servers can be multi-threaded. Holding state on the WSGI application |
|
93 | 93 | is prone to race conditions. Instances of this class exist to hold |
|
94 | 94 | mutable and race-free state for requests. |
|
95 | 95 | """ |
|
96 | 96 | def __init__(self, app, repo): |
|
97 | 97 | self.repo = repo |
|
98 | 98 | self.reponame = app.reponame |
|
99 | 99 | |
|
100 | 100 | self.archivespecs = archivespecs |
|
101 | 101 | |
|
102 | 102 | self.maxchanges = self.configint('web', 'maxchanges') |
|
103 | 103 | self.stripecount = self.configint('web', 'stripes') |
|
104 | 104 | self.maxshortchanges = self.configint('web', 'maxshortchanges') |
|
105 | 105 | self.maxfiles = self.configint('web', 'maxfiles') |
|
106 | 106 | self.allowpull = self.configbool('web', 'allow-pull') |
|
107 | 107 | |
|
108 | 108 | # we use untrusted=False to prevent a repo owner from using |
|
109 | 109 | # web.templates in .hg/hgrc to get access to any file readable |
|
110 | 110 | # by the user running the CGI script |
|
111 | 111 | self.templatepath = self.config('web', 'templates', untrusted=False) |
|
112 | 112 | |
|
113 | 113 | # This object is more expensive to build than simple config values. |
|
114 | 114 | # It is shared across requests. The app will replace the object |
|
115 | 115 | # if it is updated. Since this is a reference and nothing should |
|
116 | 116 | # modify the underlying object, it should be constant for the lifetime |
|
117 | 117 | # of the request. |
|
118 | 118 | self.websubtable = app.websubtable |
|
119 | 119 | |
|
120 | 120 | self.csp, self.nonce = cspvalues(self.repo.ui) |
|
121 | 121 | |
|
122 | 122 | # Trust the settings from the .hg/hgrc files by default. |
|
123 | 123 | def config(self, section, name, default=uimod._unset, untrusted=True): |
|
124 | 124 | return self.repo.ui.config(section, name, default, |
|
125 | 125 | untrusted=untrusted) |
|
126 | 126 | |
|
127 | 127 | def configbool(self, section, name, default=uimod._unset, untrusted=True): |
|
128 | 128 | return self.repo.ui.configbool(section, name, default, |
|
129 | 129 | untrusted=untrusted) |
|
130 | 130 | |
|
131 | 131 | def configint(self, section, name, default=uimod._unset, untrusted=True): |
|
132 | 132 | return self.repo.ui.configint(section, name, default, |
|
133 | 133 | untrusted=untrusted) |
|
134 | 134 | |
|
135 | 135 | def configlist(self, section, name, default=uimod._unset, untrusted=True): |
|
136 | 136 | return self.repo.ui.configlist(section, name, default, |
|
137 | 137 | untrusted=untrusted) |
|
138 | 138 | |
|
139 | 139 | def archivelist(self, nodeid): |
|
140 | 140 | allowed = self.configlist('web', 'allow_archive') |
|
141 | 141 | for typ, spec in self.archivespecs.iteritems(): |
|
142 | 142 | if typ in allowed or self.configbool('web', 'allow%s' % typ): |
|
143 | 143 | yield {'type': typ, 'extension': spec[2], 'node': nodeid} |
|
144 | 144 | |
|
145 | 145 | def templater(self, wsgireq, req): |
|
146 | 146 | # determine scheme, port and server name |
|
147 | 147 | # this is needed to create absolute urls |
|
148 | 148 | logourl = self.config('web', 'logourl') |
|
149 | 149 | logoimg = self.config('web', 'logoimg') |
|
150 | 150 | staticurl = (self.config('web', 'staticurl') |
|
151 |
or |
|
|
151 | or req.apppath + '/static/') | |
|
152 | 152 | if not staticurl.endswith('/'): |
|
153 | 153 | staticurl += '/' |
|
154 | 154 | |
|
155 | 155 | # some functions for the templater |
|
156 | 156 | |
|
157 | 157 | def motd(**map): |
|
158 | 158 | yield self.config('web', 'motd') |
|
159 | 159 | |
|
160 | 160 | # figure out which style to use |
|
161 | 161 | |
|
162 | 162 | vars = {} |
|
163 | 163 | styles, (style, mapfile) = getstyle(wsgireq, self.config, |
|
164 | 164 | self.templatepath) |
|
165 | 165 | if style == styles[0]: |
|
166 | 166 | vars['style'] = style |
|
167 | 167 | |
|
168 | 168 | sessionvars = webutil.sessionvars(vars, '?') |
|
169 | 169 | |
|
170 | 170 | if not self.reponame: |
|
171 | 171 | self.reponame = (self.config('web', 'name', '') |
|
172 | 172 | or wsgireq.env.get('REPO_NAME') |
|
173 |
or |
|
|
173 | or req.apppath or self.repo.root) | |
|
174 | 174 | |
|
175 | 175 | def websubfilter(text): |
|
176 | 176 | return templatefilters.websub(text, self.websubtable) |
|
177 | 177 | |
|
178 | 178 | # create the templater |
|
179 | 179 | # TODO: export all keywords: defaults = templatekw.keywords.copy() |
|
180 | 180 | defaults = { |
|
181 | 'url': pycompat.sysbytes(wsgireq.url), | |
|
181 | 'url': req.apppath + '/', | |
|
182 | 182 | 'logourl': logourl, |
|
183 | 183 | 'logoimg': logoimg, |
|
184 | 184 | 'staticurl': staticurl, |
|
185 | 185 | 'urlbase': req.advertisedbaseurl, |
|
186 | 186 | 'repo': self.reponame, |
|
187 | 187 | 'encoding': encoding.encoding, |
|
188 | 188 | 'motd': motd, |
|
189 | 189 | 'sessionvars': sessionvars, |
|
190 |
'pathdef': makebreadcrumb( |
|
|
190 | 'pathdef': makebreadcrumb(req.apppath), | |
|
191 | 191 | 'style': style, |
|
192 | 192 | 'nonce': self.nonce, |
|
193 | 193 | } |
|
194 | 194 | tres = formatter.templateresources(self.repo.ui, self.repo) |
|
195 | 195 | tmpl = templater.templater.frommapfile(mapfile, |
|
196 | 196 | filters={'websub': websubfilter}, |
|
197 | 197 | defaults=defaults, |
|
198 | 198 | resources=tres) |
|
199 | 199 | return tmpl |
|
200 | 200 | |
|
201 | 201 | |
|
202 | 202 | class hgweb(object): |
|
203 | 203 | """HTTP server for individual repositories. |
|
204 | 204 | |
|
205 | 205 | Instances of this class serve HTTP responses for a particular |
|
206 | 206 | repository. |
|
207 | 207 | |
|
208 | 208 | Instances are typically used as WSGI applications. |
|
209 | 209 | |
|
210 | 210 | Some servers are multi-threaded. On these servers, there may |
|
211 | 211 | be multiple active threads inside __call__. |
|
212 | 212 | """ |
|
213 | 213 | def __init__(self, repo, name=None, baseui=None): |
|
214 | 214 | if isinstance(repo, str): |
|
215 | 215 | if baseui: |
|
216 | 216 | u = baseui.copy() |
|
217 | 217 | else: |
|
218 | 218 | u = uimod.ui.load() |
|
219 | 219 | r = hg.repository(u, repo) |
|
220 | 220 | else: |
|
221 | 221 | # we trust caller to give us a private copy |
|
222 | 222 | r = repo |
|
223 | 223 | |
|
224 | 224 | r.ui.setconfig('ui', 'report_untrusted', 'off', 'hgweb') |
|
225 | 225 | r.baseui.setconfig('ui', 'report_untrusted', 'off', 'hgweb') |
|
226 | 226 | r.ui.setconfig('ui', 'nontty', 'true', 'hgweb') |
|
227 | 227 | r.baseui.setconfig('ui', 'nontty', 'true', 'hgweb') |
|
228 | 228 | # resolve file patterns relative to repo root |
|
229 | 229 | r.ui.setconfig('ui', 'forcecwd', r.root, 'hgweb') |
|
230 | 230 | r.baseui.setconfig('ui', 'forcecwd', r.root, 'hgweb') |
|
231 | 231 | # displaying bundling progress bar while serving feel wrong and may |
|
232 | 232 | # break some wsgi implementation. |
|
233 | 233 | r.ui.setconfig('progress', 'disable', 'true', 'hgweb') |
|
234 | 234 | r.baseui.setconfig('progress', 'disable', 'true', 'hgweb') |
|
235 | 235 | self._repos = [hg.cachedlocalrepo(self._webifyrepo(r))] |
|
236 | 236 | self._lastrepo = self._repos[0] |
|
237 | 237 | hook.redirect(True) |
|
238 | 238 | self.reponame = name |
|
239 | 239 | |
|
240 | 240 | def _webifyrepo(self, repo): |
|
241 | 241 | repo = getwebview(repo) |
|
242 | 242 | self.websubtable = webutil.getwebsubs(repo) |
|
243 | 243 | return repo |
|
244 | 244 | |
|
245 | 245 | @contextlib.contextmanager |
|
246 | 246 | def _obtainrepo(self): |
|
247 | 247 | """Obtain a repo unique to the caller. |
|
248 | 248 | |
|
249 | 249 | Internally we maintain a stack of cachedlocalrepo instances |
|
250 | 250 | to be handed out. If one is available, we pop it and return it, |
|
251 | 251 | ensuring it is up to date in the process. If one is not available, |
|
252 | 252 | we clone the most recently used repo instance and return it. |
|
253 | 253 | |
|
254 | 254 | It is currently possible for the stack to grow without bounds |
|
255 | 255 | if the server allows infinite threads. However, servers should |
|
256 | 256 | have a thread limit, thus establishing our limit. |
|
257 | 257 | """ |
|
258 | 258 | if self._repos: |
|
259 | 259 | cached = self._repos.pop() |
|
260 | 260 | r, created = cached.fetch() |
|
261 | 261 | else: |
|
262 | 262 | cached = self._lastrepo.copy() |
|
263 | 263 | r, created = cached.fetch() |
|
264 | 264 | if created: |
|
265 | 265 | r = self._webifyrepo(r) |
|
266 | 266 | |
|
267 | 267 | self._lastrepo = cached |
|
268 | 268 | self.mtime = cached.mtime |
|
269 | 269 | try: |
|
270 | 270 | yield r |
|
271 | 271 | finally: |
|
272 | 272 | self._repos.append(cached) |
|
273 | 273 | |
|
274 | 274 | def run(self): |
|
275 | 275 | """Start a server from CGI environment. |
|
276 | 276 | |
|
277 | 277 | Modern servers should be using WSGI and should avoid this |
|
278 | 278 | method, if possible. |
|
279 | 279 | """ |
|
280 | 280 | if not encoding.environ.get('GATEWAY_INTERFACE', |
|
281 | 281 | '').startswith("CGI/1."): |
|
282 | 282 | raise RuntimeError("This function is only intended to be " |
|
283 | 283 | "called while running as a CGI script.") |
|
284 | 284 | wsgicgi.launch(self) |
|
285 | 285 | |
|
286 | 286 | def __call__(self, env, respond): |
|
287 | 287 | """Run the WSGI application. |
|
288 | 288 | |
|
289 | 289 | This may be called by multiple threads. |
|
290 | 290 | """ |
|
291 | 291 | req = requestmod.wsgirequest(env, respond) |
|
292 | 292 | return self.run_wsgi(req) |
|
293 | 293 | |
|
294 | 294 | def run_wsgi(self, wsgireq): |
|
295 | 295 | """Internal method to run the WSGI application. |
|
296 | 296 | |
|
297 | 297 | This is typically only called by Mercurial. External consumers |
|
298 | 298 | should be using instances of this class as the WSGI application. |
|
299 | 299 | """ |
|
300 | 300 | with self._obtainrepo() as repo: |
|
301 | 301 | profile = repo.ui.configbool('profiling', 'enabled') |
|
302 | 302 | with profiling.profile(repo.ui, enabled=profile): |
|
303 | 303 | for r in self._runwsgi(wsgireq, repo): |
|
304 | 304 | yield r |
|
305 | 305 | |
|
306 | 306 | def _runwsgi(self, wsgireq, repo): |
|
307 | 307 | req = requestmod.parserequestfromenv(wsgireq.env) |
|
308 | 308 | rctx = requestcontext(self, repo) |
|
309 | 309 | |
|
310 | 310 | # This state is global across all threads. |
|
311 | 311 | encoding.encoding = rctx.config('web', 'encoding') |
|
312 | 312 | rctx.repo.ui.environ = wsgireq.env |
|
313 | 313 | |
|
314 | 314 | if rctx.csp: |
|
315 | 315 | # hgwebdir may have added CSP header. Since we generate our own, |
|
316 | 316 | # replace it. |
|
317 | 317 | wsgireq.headers = [h for h in wsgireq.headers |
|
318 | 318 | if h[0] != 'Content-Security-Policy'] |
|
319 | 319 | wsgireq.headers.append(('Content-Security-Policy', rctx.csp)) |
|
320 | 320 | |
|
321 | wsgireq.url = pycompat.sysstr(req.apppath) | |
|
322 | ||
|
323 | 321 | if r'PATH_INFO' in wsgireq.env: |
|
324 | 322 | parts = wsgireq.env[r'PATH_INFO'].strip(r'/').split(r'/') |
|
325 | 323 | repo_parts = wsgireq.env.get(r'REPO_NAME', r'').split(r'/') |
|
326 | 324 | if parts[:len(repo_parts)] == repo_parts: |
|
327 | 325 | parts = parts[len(repo_parts):] |
|
328 | 326 | query = r'/'.join(parts) |
|
329 | 327 | else: |
|
330 | 328 | query = wsgireq.env[r'QUERY_STRING'].partition(r'&')[0] |
|
331 | 329 | query = query.partition(r';')[0] |
|
332 | 330 | |
|
333 | 331 | # Route it to a wire protocol handler if it looks like a wire protocol |
|
334 | 332 | # request. |
|
335 | 333 | protohandler = wireprotoserver.parsehttprequest(rctx, wsgireq, query, |
|
336 | 334 | self.check_perm) |
|
337 | 335 | |
|
338 | 336 | if protohandler: |
|
339 | 337 | try: |
|
340 | 338 | if query: |
|
341 | 339 | raise ErrorResponse(HTTP_NOT_FOUND) |
|
342 | 340 | |
|
343 | 341 | return protohandler['dispatch']() |
|
344 | 342 | except ErrorResponse as inst: |
|
345 | 343 | return protohandler['handleerror'](inst) |
|
346 | 344 | |
|
347 | 345 | # translate user-visible url structure to internal structure |
|
348 | 346 | |
|
349 | 347 | args = query.split(r'/', 2) |
|
350 | 348 | if 'cmd' not in wsgireq.form and args and args[0]: |
|
351 | 349 | cmd = args.pop(0) |
|
352 | 350 | style = cmd.rfind('-') |
|
353 | 351 | if style != -1: |
|
354 | 352 | wsgireq.form['style'] = [cmd[:style]] |
|
355 | 353 | cmd = cmd[style + 1:] |
|
356 | 354 | |
|
357 | 355 | # avoid accepting e.g. style parameter as command |
|
358 | 356 | if util.safehasattr(webcommands, cmd): |
|
359 | 357 | wsgireq.form['cmd'] = [cmd] |
|
360 | 358 | |
|
361 | 359 | if cmd == 'static': |
|
362 | 360 | wsgireq.form['file'] = ['/'.join(args)] |
|
363 | 361 | else: |
|
364 | 362 | if args and args[0]: |
|
365 | 363 | node = args.pop(0).replace('%2F', '/') |
|
366 | 364 | wsgireq.form['node'] = [node] |
|
367 | 365 | if args: |
|
368 | 366 | wsgireq.form['file'] = args |
|
369 | 367 | |
|
370 | 368 | ua = wsgireq.env.get('HTTP_USER_AGENT', '') |
|
371 | 369 | if cmd == 'rev' and 'mercurial' in ua: |
|
372 | 370 | wsgireq.form['style'] = ['raw'] |
|
373 | 371 | |
|
374 | 372 | if cmd == 'archive': |
|
375 | 373 | fn = wsgireq.form['node'][0] |
|
376 | 374 | for type_, spec in rctx.archivespecs.iteritems(): |
|
377 | 375 | ext = spec[2] |
|
378 | 376 | if fn.endswith(ext): |
|
379 | 377 | wsgireq.form['node'] = [fn[:-len(ext)]] |
|
380 | 378 | wsgireq.form['type'] = [type_] |
|
381 | 379 | else: |
|
382 | 380 | cmd = wsgireq.form.get('cmd', [''])[0] |
|
383 | 381 | |
|
384 | 382 | # process the web interface request |
|
385 | 383 | |
|
386 | 384 | try: |
|
387 | 385 | tmpl = rctx.templater(wsgireq, req) |
|
388 | 386 | ctype = tmpl('mimetype', encoding=encoding.encoding) |
|
389 | 387 | ctype = templater.stringify(ctype) |
|
390 | 388 | |
|
391 | 389 | # check read permissions non-static content |
|
392 | 390 | if cmd != 'static': |
|
393 | 391 | self.check_perm(rctx, wsgireq, None) |
|
394 | 392 | |
|
395 | 393 | if cmd == '': |
|
396 | 394 | wsgireq.form['cmd'] = [tmpl.cache['default']] |
|
397 | 395 | cmd = wsgireq.form['cmd'][0] |
|
398 | 396 | |
|
399 | 397 | # Don't enable caching if using a CSP nonce because then it wouldn't |
|
400 | 398 | # be a nonce. |
|
401 | 399 | if rctx.configbool('web', 'cache') and not rctx.nonce: |
|
402 | 400 | caching(self, wsgireq) # sets ETag header or raises NOT_MODIFIED |
|
403 | 401 | if cmd not in webcommands.__all__: |
|
404 | 402 | msg = 'no such method: %s' % cmd |
|
405 | 403 | raise ErrorResponse(HTTP_BAD_REQUEST, msg) |
|
406 | 404 | elif cmd == 'file' and 'raw' in wsgireq.form.get('style', []): |
|
407 | 405 | rctx.ctype = ctype |
|
408 | 406 | content = webcommands.rawfile(rctx, wsgireq, tmpl) |
|
409 | 407 | else: |
|
410 | 408 | content = getattr(webcommands, cmd)(rctx, wsgireq, tmpl) |
|
411 | 409 | wsgireq.respond(HTTP_OK, ctype) |
|
412 | 410 | |
|
413 | 411 | return content |
|
414 | 412 | |
|
415 | 413 | except (error.LookupError, error.RepoLookupError) as err: |
|
416 | 414 | wsgireq.respond(HTTP_NOT_FOUND, ctype) |
|
417 | 415 | msg = pycompat.bytestr(err) |
|
418 | 416 | if (util.safehasattr(err, 'name') and |
|
419 | 417 | not isinstance(err, error.ManifestLookupError)): |
|
420 | 418 | msg = 'revision not found: %s' % err.name |
|
421 | 419 | return tmpl('error', error=msg) |
|
422 | 420 | except (error.RepoError, error.RevlogError) as inst: |
|
423 | 421 | wsgireq.respond(HTTP_SERVER_ERROR, ctype) |
|
424 | 422 | return tmpl('error', error=pycompat.bytestr(inst)) |
|
425 | 423 | except ErrorResponse as inst: |
|
426 | 424 | wsgireq.respond(inst, ctype) |
|
427 | 425 | if inst.code == HTTP_NOT_MODIFIED: |
|
428 | 426 | # Not allowed to return a body on a 304 |
|
429 | 427 | return [''] |
|
430 | 428 | return tmpl('error', error=pycompat.bytestr(inst)) |
|
431 | 429 | |
|
432 | 430 | def check_perm(self, rctx, req, op): |
|
433 | 431 | for permhook in permhooks: |
|
434 | 432 | permhook(rctx, req, op) |
|
435 | 433 | |
|
436 | 434 | def getwebview(repo): |
|
437 | 435 | """The 'web.view' config controls changeset filter to hgweb. Possible |
|
438 | 436 | values are ``served``, ``visible`` and ``all``. Default is ``served``. |
|
439 | 437 | The ``served`` filter only shows changesets that can be pulled from the |
|
440 | 438 | hgweb instance. The``visible`` filter includes secret changesets but |
|
441 | 439 | still excludes "hidden" one. |
|
442 | 440 | |
|
443 | 441 | See the repoview module for details. |
|
444 | 442 | |
|
445 | 443 | The option has been around undocumented since Mercurial 2.5, but no |
|
446 | 444 | user ever asked about it. So we better keep it undocumented for now.""" |
|
447 | 445 | # experimental config: web.view |
|
448 | 446 | viewconfig = repo.ui.config('web', 'view', untrusted=True) |
|
449 | 447 | if viewconfig == 'all': |
|
450 | 448 | return repo.unfiltered() |
|
451 | 449 | elif viewconfig in repoview.filtertable: |
|
452 | 450 | return repo.filtered(viewconfig) |
|
453 | 451 | else: |
|
454 | 452 | return repo.filtered('served') |
@@ -1,280 +1,279 b'' | |||
|
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 cgi |
|
12 | 12 | import errno |
|
13 | 13 | import socket |
|
14 | 14 | #import wsgiref.validate |
|
15 | 15 | |
|
16 | 16 | from .common import ( |
|
17 | 17 | ErrorResponse, |
|
18 | 18 | HTTP_NOT_MODIFIED, |
|
19 | 19 | statusmessage, |
|
20 | 20 | ) |
|
21 | 21 | |
|
22 | 22 | from ..thirdparty import ( |
|
23 | 23 | attr, |
|
24 | 24 | ) |
|
25 | 25 | from .. import ( |
|
26 | 26 | pycompat, |
|
27 | 27 | util, |
|
28 | 28 | ) |
|
29 | 29 | |
|
30 | 30 | shortcuts = { |
|
31 | 31 | 'cl': [('cmd', ['changelog']), ('rev', None)], |
|
32 | 32 | 'sl': [('cmd', ['shortlog']), ('rev', None)], |
|
33 | 33 | 'cs': [('cmd', ['changeset']), ('node', None)], |
|
34 | 34 | 'f': [('cmd', ['file']), ('filenode', None)], |
|
35 | 35 | 'fl': [('cmd', ['filelog']), ('filenode', None)], |
|
36 | 36 | 'fd': [('cmd', ['filediff']), ('node', None)], |
|
37 | 37 | 'fa': [('cmd', ['annotate']), ('filenode', None)], |
|
38 | 38 | 'mf': [('cmd', ['manifest']), ('manifest', None)], |
|
39 | 39 | 'ca': [('cmd', ['archive']), ('node', None)], |
|
40 | 40 | 'tags': [('cmd', ['tags'])], |
|
41 | 41 | 'tip': [('cmd', ['changeset']), ('node', ['tip'])], |
|
42 | 42 | 'static': [('cmd', ['static']), ('file', None)] |
|
43 | 43 | } |
|
44 | 44 | |
|
45 | 45 | def normalize(form): |
|
46 | 46 | # first expand the shortcuts |
|
47 | 47 | for k in shortcuts: |
|
48 | 48 | if k in form: |
|
49 | 49 | for name, value in shortcuts[k]: |
|
50 | 50 | if value is None: |
|
51 | 51 | value = form[k] |
|
52 | 52 | form[name] = value |
|
53 | 53 | del form[k] |
|
54 | 54 | # And strip the values |
|
55 | 55 | bytesform = {} |
|
56 | 56 | for k, v in form.iteritems(): |
|
57 | 57 | bytesform[pycompat.bytesurl(k)] = [ |
|
58 | 58 | pycompat.bytesurl(i.strip()) for i in v] |
|
59 | 59 | return bytesform |
|
60 | 60 | |
|
61 | 61 | @attr.s(frozen=True) |
|
62 | 62 | class parsedrequest(object): |
|
63 | 63 | """Represents a parsed WSGI request / static HTTP request parameters.""" |
|
64 | 64 | |
|
65 | 65 | # Full URL for this request. |
|
66 | 66 | url = attr.ib() |
|
67 | 67 | # URL without any path components. Just <proto>://<host><port>. |
|
68 | 68 | baseurl = attr.ib() |
|
69 | 69 | # Advertised URL. Like ``url`` and ``baseurl`` but uses SERVER_NAME instead |
|
70 | 70 | # of HTTP: Host header for hostname. This is likely what clients used. |
|
71 | 71 | advertisedurl = attr.ib() |
|
72 | 72 | advertisedbaseurl = attr.ib() |
|
73 | 73 | # WSGI application path. |
|
74 | 74 | apppath = attr.ib() |
|
75 | 75 | # List of path parts to be used for dispatch. |
|
76 | 76 | dispatchparts = attr.ib() |
|
77 | 77 | # URL path component (no query string) used for dispatch. |
|
78 | 78 | dispatchpath = attr.ib() |
|
79 | 79 | # Raw query string (part after "?" in URL). |
|
80 | 80 | querystring = attr.ib() |
|
81 | 81 | |
|
82 | 82 | def parserequestfromenv(env): |
|
83 | 83 | """Parse URL components from environment variables. |
|
84 | 84 | |
|
85 | 85 | WSGI defines request attributes via environment variables. This function |
|
86 | 86 | parses the environment variables into a data structure. |
|
87 | 87 | """ |
|
88 | 88 | # PEP-0333 defines the WSGI spec and is a useful reference for this code. |
|
89 | 89 | |
|
90 | 90 | # We first validate that the incoming object conforms with the WSGI spec. |
|
91 | 91 | # We only want to be dealing with spec-conforming WSGI implementations. |
|
92 | 92 | # TODO enable this once we fix internal violations. |
|
93 | 93 | #wsgiref.validate.check_environ(env) |
|
94 | 94 | |
|
95 | 95 | # PEP-0333 states that environment keys and values are native strings |
|
96 | 96 | # (bytes on Python 2 and str on Python 3). The code points for the Unicode |
|
97 | 97 | # strings on Python 3 must be between \00000-\000FF. We deal with bytes |
|
98 | 98 | # in Mercurial, so mass convert string keys and values to bytes. |
|
99 | 99 | if pycompat.ispy3: |
|
100 | 100 | env = {k.encode('latin-1'): v for k, v in env.iteritems()} |
|
101 | 101 | env = {k: v.encode('latin-1') if isinstance(v, str) else v |
|
102 | 102 | for k, v in env.iteritems()} |
|
103 | 103 | |
|
104 | 104 | # https://www.python.org/dev/peps/pep-0333/#environ-variables defines |
|
105 | 105 | # the environment variables. |
|
106 | 106 | # https://www.python.org/dev/peps/pep-0333/#url-reconstruction defines |
|
107 | 107 | # how URLs are reconstructed. |
|
108 | 108 | fullurl = env['wsgi.url_scheme'] + '://' |
|
109 | 109 | advertisedfullurl = fullurl |
|
110 | 110 | |
|
111 | 111 | def addport(s): |
|
112 | 112 | if env['wsgi.url_scheme'] == 'https': |
|
113 | 113 | if env['SERVER_PORT'] != '443': |
|
114 | 114 | s += ':' + env['SERVER_PORT'] |
|
115 | 115 | else: |
|
116 | 116 | if env['SERVER_PORT'] != '80': |
|
117 | 117 | s += ':' + env['SERVER_PORT'] |
|
118 | 118 | |
|
119 | 119 | return s |
|
120 | 120 | |
|
121 | 121 | if env.get('HTTP_HOST'): |
|
122 | 122 | fullurl += env['HTTP_HOST'] |
|
123 | 123 | else: |
|
124 | 124 | fullurl += env['SERVER_NAME'] |
|
125 | 125 | fullurl = addport(fullurl) |
|
126 | 126 | |
|
127 | 127 | advertisedfullurl += env['SERVER_NAME'] |
|
128 | 128 | advertisedfullurl = addport(advertisedfullurl) |
|
129 | 129 | |
|
130 | 130 | baseurl = fullurl |
|
131 | 131 | advertisedbaseurl = advertisedfullurl |
|
132 | 132 | |
|
133 | 133 | fullurl += util.urlreq.quote(env.get('SCRIPT_NAME', '')) |
|
134 | 134 | advertisedfullurl += util.urlreq.quote(env.get('SCRIPT_NAME', '')) |
|
135 | 135 | fullurl += util.urlreq.quote(env.get('PATH_INFO', '')) |
|
136 | 136 | advertisedfullurl += util.urlreq.quote(env.get('PATH_INFO', '')) |
|
137 | 137 | |
|
138 | 138 | if env.get('QUERY_STRING'): |
|
139 | 139 | fullurl += '?' + env['QUERY_STRING'] |
|
140 | 140 | advertisedfullurl += '?' + env['QUERY_STRING'] |
|
141 | 141 | |
|
142 | 142 | # When dispatching requests, we look at the URL components (PATH_INFO |
|
143 | 143 | # and QUERY_STRING) after the application root (SCRIPT_NAME). But hgwebdir |
|
144 | 144 | # has the concept of "virtual" repositories. This is defined via REPO_NAME. |
|
145 | 145 | # If REPO_NAME is defined, we append it to SCRIPT_NAME to form a new app |
|
146 | 146 | # root. We also exclude its path components from PATH_INFO when resolving |
|
147 | 147 | # the dispatch path. |
|
148 | 148 | |
|
149 | # TODO the use of trailing slashes in apppath is arguably wrong. We need it | |
|
150 | # to appease low-level parts of hgweb_mod for now. | |
|
151 | 149 | apppath = env['SCRIPT_NAME'] |
|
152 | if not apppath.endswith('/'): | |
|
153 | apppath += '/' | |
|
154 | 150 | |
|
155 | 151 | if env.get('REPO_NAME'): |
|
156 | apppath += env.get('REPO_NAME') + '/' | |
|
152 | if not apppath.endswith('/'): | |
|
153 | apppath += '/' | |
|
154 | ||
|
155 | apppath += env.get('REPO_NAME') | |
|
157 | 156 | |
|
158 | 157 | if 'PATH_INFO' in env: |
|
159 | 158 | dispatchparts = env['PATH_INFO'].strip('/').split('/') |
|
160 | 159 | |
|
161 | 160 | # Strip out repo parts. |
|
162 | 161 | repoparts = env.get('REPO_NAME', '').split('/') |
|
163 | 162 | if dispatchparts[:len(repoparts)] == repoparts: |
|
164 | 163 | dispatchparts = dispatchparts[len(repoparts):] |
|
165 | 164 | else: |
|
166 | 165 | dispatchparts = [] |
|
167 | 166 | |
|
168 | 167 | dispatchpath = '/'.join(dispatchparts) |
|
169 | 168 | |
|
170 | 169 | querystring = env.get('QUERY_STRING', '') |
|
171 | 170 | |
|
172 | 171 | return parsedrequest(url=fullurl, baseurl=baseurl, |
|
173 | 172 | advertisedurl=advertisedfullurl, |
|
174 | 173 | advertisedbaseurl=advertisedbaseurl, |
|
175 | 174 | apppath=apppath, |
|
176 | 175 | dispatchparts=dispatchparts, dispatchpath=dispatchpath, |
|
177 | 176 | querystring=querystring) |
|
178 | 177 | |
|
179 | 178 | class wsgirequest(object): |
|
180 | 179 | """Higher-level API for a WSGI request. |
|
181 | 180 | |
|
182 | 181 | WSGI applications are invoked with 2 arguments. They are used to |
|
183 | 182 | instantiate instances of this class, which provides higher-level APIs |
|
184 | 183 | for obtaining request parameters, writing HTTP output, etc. |
|
185 | 184 | """ |
|
186 | 185 | def __init__(self, wsgienv, start_response): |
|
187 | 186 | version = wsgienv[r'wsgi.version'] |
|
188 | 187 | if (version < (1, 0)) or (version >= (2, 0)): |
|
189 | 188 | raise RuntimeError("Unknown and unsupported WSGI version %d.%d" |
|
190 | 189 | % version) |
|
191 | 190 | self.inp = wsgienv[r'wsgi.input'] |
|
192 | 191 | self.err = wsgienv[r'wsgi.errors'] |
|
193 | 192 | self.threaded = wsgienv[r'wsgi.multithread'] |
|
194 | 193 | self.multiprocess = wsgienv[r'wsgi.multiprocess'] |
|
195 | 194 | self.run_once = wsgienv[r'wsgi.run_once'] |
|
196 | 195 | self.env = wsgienv |
|
197 | 196 | self.form = normalize(cgi.parse(self.inp, |
|
198 | 197 | self.env, |
|
199 | 198 | keep_blank_values=1)) |
|
200 | 199 | self._start_response = start_response |
|
201 | 200 | self.server_write = None |
|
202 | 201 | self.headers = [] |
|
203 | 202 | |
|
204 | 203 | def __iter__(self): |
|
205 | 204 | return iter([]) |
|
206 | 205 | |
|
207 | 206 | def read(self, count=-1): |
|
208 | 207 | return self.inp.read(count) |
|
209 | 208 | |
|
210 | 209 | def drain(self): |
|
211 | 210 | '''need to read all data from request, httplib is half-duplex''' |
|
212 | 211 | length = int(self.env.get('CONTENT_LENGTH') or 0) |
|
213 | 212 | for s in util.filechunkiter(self.inp, limit=length): |
|
214 | 213 | pass |
|
215 | 214 | |
|
216 | 215 | def respond(self, status, type, filename=None, body=None): |
|
217 | 216 | if not isinstance(type, str): |
|
218 | 217 | type = pycompat.sysstr(type) |
|
219 | 218 | if self._start_response is not None: |
|
220 | 219 | self.headers.append((r'Content-Type', type)) |
|
221 | 220 | if filename: |
|
222 | 221 | filename = (filename.rpartition('/')[-1] |
|
223 | 222 | .replace('\\', '\\\\').replace('"', '\\"')) |
|
224 | 223 | self.headers.append(('Content-Disposition', |
|
225 | 224 | 'inline; filename="%s"' % filename)) |
|
226 | 225 | if body is not None: |
|
227 | 226 | self.headers.append((r'Content-Length', str(len(body)))) |
|
228 | 227 | |
|
229 | 228 | for k, v in self.headers: |
|
230 | 229 | if not isinstance(v, str): |
|
231 | 230 | raise TypeError('header value must be string: %r' % (v,)) |
|
232 | 231 | |
|
233 | 232 | if isinstance(status, ErrorResponse): |
|
234 | 233 | self.headers.extend(status.headers) |
|
235 | 234 | if status.code == HTTP_NOT_MODIFIED: |
|
236 | 235 | # RFC 2616 Section 10.3.5: 304 Not Modified has cases where |
|
237 | 236 | # it MUST NOT include any headers other than these and no |
|
238 | 237 | # body |
|
239 | 238 | self.headers = [(k, v) for (k, v) in self.headers if |
|
240 | 239 | k in ('Date', 'ETag', 'Expires', |
|
241 | 240 | 'Cache-Control', 'Vary')] |
|
242 | 241 | status = statusmessage(status.code, pycompat.bytestr(status)) |
|
243 | 242 | elif status == 200: |
|
244 | 243 | status = '200 Script output follows' |
|
245 | 244 | elif isinstance(status, int): |
|
246 | 245 | status = statusmessage(status) |
|
247 | 246 | |
|
248 | 247 | self.server_write = self._start_response( |
|
249 | 248 | pycompat.sysstr(status), self.headers) |
|
250 | 249 | self._start_response = None |
|
251 | 250 | self.headers = [] |
|
252 | 251 | if body is not None: |
|
253 | 252 | self.write(body) |
|
254 | 253 | self.server_write = None |
|
255 | 254 | |
|
256 | 255 | def write(self, thing): |
|
257 | 256 | if thing: |
|
258 | 257 | try: |
|
259 | 258 | self.server_write(thing) |
|
260 | 259 | except socket.error as inst: |
|
261 | 260 | if inst[0] != errno.ECONNRESET: |
|
262 | 261 | raise |
|
263 | 262 | |
|
264 | 263 | def writelines(self, lines): |
|
265 | 264 | for line in lines: |
|
266 | 265 | self.write(line) |
|
267 | 266 | |
|
268 | 267 | def flush(self): |
|
269 | 268 | return None |
|
270 | 269 | |
|
271 | 270 | def close(self): |
|
272 | 271 | return None |
|
273 | 272 | |
|
274 | 273 | def wsgiapplication(app_maker): |
|
275 | 274 | '''For compatibility with old CGI scripts. A plain hgweb() or hgwebdir() |
|
276 | 275 | can and should now be used as a WSGI application.''' |
|
277 | 276 | application = app_maker() |
|
278 | 277 | def run_wsgi(env, respond): |
|
279 | 278 | return application(env, respond) |
|
280 | 279 | return run_wsgi |
General Comments 0
You need to be logged in to leave comments.
Login now