1 $Id: TODO,v 1.105 2014/05/13 11:15:28 fabiankeil Exp $
3 Some Privoxy-related tasks, sorted by the time they
4 have been added, not by priority.
6 The latest version should be available at:
7 http://ijbswa.cvs.sourceforge.net/viewvc/ijbswa/current/TODO
9 1) Add some regression tests. Filters should be tested automatically
10 (variables too). Could probably reuse large parts of Privoxy-Filter-Test.
11 Note that there is currently work in progress to leverage curl's
12 test suite which could be used for this as well.
14 3) Fix some more XXX: comments.
16 6) Remove actions that aren't needed anymore:
18 content-type-overwrite should probably stay as it's also
19 used by some of the CGI pages (XXX: name them).
21 crunch-client-header and crunch-server-header should probably
22 go, their only advantage is that their search strings can be
23 controlled through the CGI pages, other than that they only
26 crunch-if-none-match can be replaced with a header filter.
28 prevent-compression has a misleading name and could
29 be replaced with a header filter.
31 7) force-text-mode has a stupid name and should probably
32 be renamed to force-filter-mode.
34 8) handle-as-empty-document and handle-as-image should
35 be merged to something like handle-as{something} to
36 prevent them from being activated at the same time.
38 10) There's a bug in the CGI editor that turns the
39 first section's "Insert new section below" into
40 a "Insert new section above" button.
42 11) CGI templates should use semantically-correct HTML
45 12) Support pipelining for outgoing connections.
47 14) Allow to filter POST parameters.
49 15) If trusted CGI pages are requested without trusted referrer,
50 set the status code to 403 instead of 200.
52 16) Filter SSL encrypted content as well.
54 At the beginning we could use a unencrypted connection between
55 client and Privoxy, and use an encrypted connection between
56 Privoxy and the server.
58 This should be good enough for most of the content the
59 user would want to filter.
61 19) enable-forward-fallback. Syntax? Suggested by K.R.
63 21) User Manual delivery doesn't accept multiple slashes. Should it?
65 22) Verify action files properly (Including arguments) and
66 act accordingly (should probably intercept all requests
67 with a "Invalid option foo detected" CGI page).
69 23) Do the same in case of syntax errors in the configuration file,
70 instead of just exiting or ignoring the problem.
72 25) Handle multiple filters with the same name better. Reject them?
74 26) Let show-url-info detect clearly invalid URLs.
76 27) Make errno logging less thread-unsafe.
77 Verify that it's really an improvement.
79 28) Don't take default ports in case of invalid forwarding ports.
81 31) If a string action foo is disabled csp->action->string[ACTION_STRING_FOO]
82 doesn't necessarily contain NULL, but may contain the string of an
83 enabled foo action in an overruled section. Is it a bug? Does it matter?
85 32) In case of forwarding failures with socks port == 9050,
86 show extra info about Tor (the whole FAQ entry?).
88 36) Unload unused action files directly, even if they are
89 disabled without replacement.
91 38) In the final results, explicitly list disabled multi actions
92 with their parameters. Not as trivial as it sounds.
94 40) When running in daemon mode, Privoxy's working directory is '/'
95 which means it may not have permissions to dump core when necessary.
96 Figure out a way to solve this. Introduce a cwd config option?
98 41) Change documentation framework to one that works cross-platform.
101 42) Add a DTrace USDT provider. Now that FreeBSD has userland DTrace
102 support there's no longer any reason not to.
104 43) Write a tool to check URL patterns against URLs in the log file.
105 This could be included in Privoxy-Regression-Test.
107 44) Privoxy-Log-Parser: Consider highlighting "Connection" in:
108 23:13:03.506 283b6100 Header: Replaced: 'Connection: Keep-Alive' with 'Connection: close'
110 50) Investigate possible PCRS template speedup when searching
111 macros with strstr() before compiling pcrs commands.
112 Investigated, needs some restructuring but is probably worth it.
114 51) Make user-manual directive more generic to allow serving
115 the FAQ and other stuff, too. Consider changing the port
116 for "same origin policy" issues.
118 53) Find a more reliable hoster. Involves finding out what our
119 requirements are and which SF alternatives fulfil them.
120 It would probably also make sense to look into what other
121 projects did when migrating away from SF.
123 2014-05-13: Work in progress. Hosting wish list at the end
126 54) Move away from CVS to a more modern revision control system.
127 Find out if there are any objection against going with Git.
128 Using Git would also have the advantage that SF now pretends
129 to support it, so we could do it independently from 53).
131 55) Apply for Coverity scans: http://scan.coverity.com/
133 56) Apply for the "free online access for qualified open-source
134 software projects" for the Co-Advisor HTTP compliance tests:
135 http://coad.measurement-factory.com/details.html#pricing
136 (Application sent at 2014-05-12)
138 57) Allow piping into external programs to allow more powerful
139 filters and policy decisions. Incomplete support available
140 in Fabian's popen branch.
142 58) Move more template strings from the code into the actual templates.
144 59) Import the German template translation.
146 60) Ask the Russian translators for input on how to make their
149 61) Consider (optionally?) skipping the hostname comparison when
150 checking if a connections that goes to a HTTP proxy can be reused.
151 Do all HTTP proxy support that? Is it worth it?
153 63) Reject clearly too large requests earlier?
155 64) Use proper copyright attribution. "Privoxy Developers"
158 65) Polish Website. Probably involves ditching the Docbook
159 mess in favour of wml. There are already several threads
160 in the mailinglist archives about this. See also #41.
162 66) Stop hard-coding the number of action and filter files.
164 67) Clean up source code directory layout. Depends on 54 so
165 we don't lose the revision history.
167 68) Use standard make syntax so we don't depend on GNU make.
169 69) Update autoconf setup (or move away from it).
171 70) If the server connection is reset but the headers are
172 received, consider passing the mess to the client instead
173 of showing the connect-failed template. Relates to #2698674.
175 74) Let Privoxy-Regression-Test optionally check that action
176 sections which disable actions actually are preceded by
177 sections that enable said actions.
179 75) Create a tool that creates Privoxy action (and filter?) files
180 out of adblock files. Could be implemented as option for
181 url-pattern-translator.pl.
183 76) Cache DNS responses. Note that this has been requested
184 several times by users, but is not a developer priority.
185 If you care about this, feel free to submit patches.
187 77) Allow to configure the IP address used in outgoing connections.
189 78) Allow to optionally use pcre's DFA algorithm.
191 79) Evaluate pcre alternatives.
193 80) Change FEATURE_EXTENDED_HOST_PATTERNS to support both
194 extended and vanilla host patterns at the same time.
196 Note that the requirement is to allow the user to decide
197 if the domain pattern should be interpreted as regex or
198 traditional host pattern and if it's not obvious that the
199 user made any decision, default to the latter.
201 Possible solutions would be:
203 1. An always-use-regex-domain-patterns config option
204 2. An enable-regex-domain-patterns-for-this-action-file option
205 3. An enable-regex-domain-patterns-for-this-action-file-until-the-user-says-otherwise option
206 4. A treat-the-domain-pattern-in-this-line-as-regex(-or-not) option
207 5. Combinations of the options above
209 With 2+4, 3+4 or 2+3+4 being the preferences until
212 82) Detect if the system time goes back in time let the user
213 know if it caused any connections to get closed.
215 84) Flesh out the user-manual delivery to serve pages from
216 other directories, too.
218 85) Once #84 is done, write a script that populates a directory with
219 various common third-party icons (stumbleupon.png, facebook.png ...)
220 and redirect requests for them to Privoxy.
222 86) Add a server-body-tagger action. This is trivial as as all the
223 functionality required to do it already exists.
225 87) Add a client-body-tagger action. This is less trivial as we currently
226 don't buffer client bodies. After 14) is implemented it would be
229 88) Investigate if there's a Perl module that Privoxy-Regression-Test
230 could optionally use to keep connections alive, preferably while
231 requiring less forks at the same time.
233 89) When multiple block actions apply, consider showing all the block
234 reasons on the blocked page that haven't been overruled, not just
237 91) Add an optional limit for internal redirects. It would probably
238 be reasonable to default to a limit of one and showing an error
239 message if the request for the redirect URL would be redirected
242 92) The statistics currently aren't calculated correctly by Privoxy
243 as each thread is only counted as one request which is no longer
244 correct. This should be fixed, or the statistic code removed.
245 Privoxy-Log-Parser's provides more detailed statistics, anyway.
247 93) Add a config directive to let Privoxy explicitly request either
248 IPv4 (or IPv6) addresses, even if the system supports both.
249 Could be useful as a workaround for misconfigured setups where the
250 libc returns IPv6 addresses even if there's no IPv6 connectivity.
252 94) Add a config directive to let Privoxy prefer either IPv4 (or IPv6)
253 addresses, instead of trusting the libc to return them in an order
254 that makes sense. Like #93, this could be useful as a workaround
255 for misconfigured setups.
257 95) Support a non-standard client header in CONNECT requests that
258 contains the URL of the requested resource, which is then treated
259 like the request URL.
261 This way the client could opt-in for path-based blocking of https
262 requests. Given that the headers from the CONNECT request aren't
263 forwarded to the destination server, an unencrypted URL should be
264 acceptable if the client and Privoxy are running on the same system
265 or in a trusted environment.
267 96) Filters should be easier to look up. Currently get_filter() has to
268 go through all filters and skip the filter types the caller isn't
271 98) When showing action section on the CGI pages, properly escape
272 line breaks so they can be copy&pasted into action files without
275 99) Figure out a mechanism through which a user can easily enable
276 site-specific action sections that are too aggressive to be
277 enabled by default. This could be similar to the presettings
278 in default.action, but could also be just another action file
279 that isn't used by default.
281 100) Create a cross-platform Privoxy control program and retire
282 the win32 GUI. Integrate support for Privoxy-Regression-Test,
283 Privoxy-Log-Parser, Privoxy-Filter-Test, uagen and similar tools.
285 102) Add an include directive to split the config file into several parts.
287 103) Potential performance improvement for large action files:
288 when figuring out which actions apply, check the action bit mask
289 before pattern matching and skip section that wouldn't modify the
290 actions already set. To increase the impact the sections would have
291 to be applied in reverse.
293 104) The code to modify global_toggle_state should be factored out into
294 a separate function. Currently we mess with it in three different
295 files, but only in w32log.c the tray icon is explicitly set.
296 The logging is inconsistent as well. For details see #3525694.
298 105) Add support for socks authentication.
300 106) actionlist.h should be embedded in a way that causes less text
303 107) Support more pcrs variables, for example $destination-ip-address
304 and $source-ip-address.
306 108) Allow to use a somewhat random string instead of PRIVOXY-FORCE.
308 109) Let log_error() support the format specifier %S which should
309 work like %s but escape new lines like %N. This would be useful
310 to log the result of header filters which may inject new lines.
312 110) Add a global-buffer-limit directive that roughly limits how
313 much malloc'ed memory Privoxy will use and can potentially
314 be smaller than (buffer-limit * max-client-connections).
316 111) Reject requests if hosts and ports in request line and Host
317 header don't match (before filters have been applied).
319 112) If a header filter is used to inject another header by inserting
320 a \r\n (undocumented feature), detect it and split the headers so
321 following header actions do not treat them as a single string.
322 Alternatively add another header injection mechanism.
324 113) Log statistics upon receiving a certain signal (SIGINFO or SIGUSR1).
326 114) Properly deal with status code 100. The current "Continue hack"
327 can cause problems for gpg when uploading keys through Privoxy.
329 115) Add ICAP (RFC 3507) support. FR #3615158.
331 116) Due to the use of sscanf(), Privoxy currently will fail to properly
332 parse chunks whose size can't be represented with 32 bit. This is
333 unlikely to cause problems in the real world, but should eventually
334 be fixed anyway. See also:
335 https://bugzilla.mozilla.org/show_bug.cgi?id=959100
337 118) There should be "escaped" dynamic variables that are guaranteed
338 not to break filters.
340 119) Evaluate using pcre's jit mode.
342 120) Add an option to limit pcre's recursion limit below the default.
343 On some platforms the recursion limit doesn't prevent pcre from
344 running out of stack space, causing the kernel to kill Privoxy
347 121) Add HTTP/2 support. As a first step, incomming HTTP/1.x requests
348 should be translated to outgoing HTTP/2 requests where possible
349 (and if desired by the user).
351 122) Allow customized log messages.
353 123) Evaluate if the voluntarily-disclose-session-keys option in Firefox
354 (and other browsers) can be leveraged. Probably depends on #16.
356 124) Add support for the "lightweight OS capability and sandbox framework"
357 Capsicum. http://www.cl.cam.ac.uk/research/security/capsicum/
359 125) Allow clients to HTTPS-encrypt the proxy connection.
361 ##########################################################################
363 Hosting wish list (relevant for #53)
368 - Mailinglists (Mailman with public archives preferred)
369 - Webspace (on a Unix-like OS that works with the webserver targets
371 - Source code repositories (currently CVS, but migrating away
372 from it is TODO #54 anyway and shouldn't be too much trouble)
373 - Commit mails (preferably with unified diffs)
375 (Unsorted) details to look at when evaluating hosters:
377 1. Preferably no third-party ads and trackers.
378 External images, CSS and JavaScript may count as trackers
379 but texts like "supported by company XYZ" may be acceptable.
381 2. JavaScript should be optional or not used at all.
383 3. Services we don't need shouldn't be enabled anyway.
384 (We currently don't use Web forums, wikis, surveys etc.)
386 4. It would be preferable if the hoster didn't have a bad track
387 record as far as user experience, security and privacy are
388 concerned and if the terms of services are "reasonable" and
389 haven't changed too often in the past. Updates in the past
390 should have been improvements and not regressions.
392 5. It would be preferable if most of the server administration
393 is done by a trusted third-party (or at least not a lot of work
396 6. The server(s) should be located in a country with laws we can
397 understand and follow (or at least not unintentionally violate).
399 7. A server location in a country with some kind of due process
400 and strong data protection laws (at least on paper) would be
403 8. Given that Privoxy is a free software project it would be
404 preferable if the hoster would use free software where possible.
406 9. Migrating away from the hoster in the future without losing
407 any important data should be possible without writing web