1 $Id: TODO,v 1.115 2014/06/16 10:17:38 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 more 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, patches have been submitted upstream:
13 http://curl.haxx.se/mail/lib-2014-06/0070.html
15 3) Fix some more XXX: comments.
17 6) Remove actions that aren't needed anymore:
19 content-type-overwrite should probably stay as it's also
20 used by some of the CGI pages (XXX: name them).
22 crunch-client-header and crunch-server-header should probably
23 go, their only advantage is that their search strings can be
24 controlled through the CGI pages, other than that they only
27 crunch-if-none-match can be replaced with a header filter.
29 prevent-compression has a misleading name and could
30 be replaced with a header filter.
32 7) force-text-mode has a stupid name and should probably
33 be renamed to force-filter-mode.
35 8) handle-as-empty-document and handle-as-image should
36 be merged to something like handle-as{something} to
37 prevent them from being activated at the same time.
39 10) There's a bug in the CGI editor that turns the
40 first section's "Insert new section below" into
41 a "Insert new section above" button.
43 11) CGI templates should use semantically-correct HTML
46 12) Support pipelining for outgoing connections.
48 14) Allow to filter POST parameters.
50 15) If trusted CGI pages are requested without trusted referrer,
51 set the status code to 403 instead of 200.
53 16) Filter SSL encrypted content as well.
55 At the beginning we could use a unencrypted connection between
56 client and Privoxy, and use an encrypted connection between
57 Privoxy and the server.
59 This should be good enough for most of the content the
60 user would want to filter.
62 19) enable-forward-fallback. Syntax? Suggested by K.R.
64 21) User Manual delivery doesn't accept multiple slashes. Should it?
66 22) Verify action files properly (Including arguments) and
67 act accordingly (should probably intercept all requests
68 with a "Invalid option foo detected" CGI page).
70 23) Do the same in case of syntax errors in the configuration file,
71 instead of just exiting or ignoring the problem.
73 25) Handle multiple filters with the same name better. Reject them?
75 26) Let show-url-info detect clearly invalid URLs.
77 27) Make errno logging less thread-unsafe.
78 Verify that it's really an improvement.
80 28) Don't take default ports in case of invalid forwarding ports.
82 31) If a string action foo is disabled csp->action->string[ACTION_STRING_FOO]
83 doesn't necessarily contain NULL, but may contain the string of an
84 enabled foo action in an overruled section. Is it a bug? Does it matter?
86 32) In case of forwarding failures with socks port == 9050,
87 show extra info about Tor (the whole FAQ entry?).
89 36) Unload unused action files directly, even if they are
90 disabled without replacement.
92 38) In the final results, explicitly list disabled multi actions
93 with their parameters. Not as trivial as it sounds.
95 40) When running in daemon mode, Privoxy's working directory is '/'
96 which means it may not have permissions to dump core when necessary.
97 Figure out a way to solve this. Introduce a cwd config option?
99 41) Change documentation framework to one that works cross-platform.
102 42) Add a DTrace USDT provider. Now that FreeBSD has userland DTrace
103 support there's no longer any reason not to.
105 43) Write a tool to check URL patterns against URLs in the log file.
106 This could be included in Privoxy-Regression-Test.
108 44) Privoxy-Log-Parser: Consider highlighting "Connection" in:
109 23:13:03.506 283b6100 Header: Replaced: 'Connection: Keep-Alive' with 'Connection: close'
111 50) Investigate possible PCRS template speedup when searching
112 macros with strstr() before compiling pcrs commands.
113 Investigated, needs some restructuring but is probably worth it.
115 51) Make user-manual directive more generic to allow serving
116 the FAQ and other stuff, too. Consider changing the port
117 for "same origin policy" issues.
119 53) Find a more reliable hoster. Involves finding out what our
120 requirements are and which SF alternatives fulfil them.
121 It would probably also make sense to look into what other
122 projects did when migrating away from SF.
124 2014-05-13: Work in progress. Hosting wish list at the end
127 54) Move away from CVS to a more modern revision control system.
128 Find out if there are any objection against going with Git.
129 Using Git would also have the advantage that SF now pretends
130 to support it, so we could do it independently from 53).
132 55) Apply for Coverity scans: http://scan.coverity.com/
134 58) Move more template strings from the code into the actual templates.
136 59) Import the German template translation.
138 60) Ask the Russian translators for input on how to make their
141 61) Consider (optionally?) skipping the hostname comparison when
142 checking if a connections that goes to a HTTP proxy can be reused.
143 Do all HTTP proxy support that? Is it worth it?
145 63) Reject clearly too large requests earlier?
147 64) Use proper copyright attribution. "Privoxy Developers"
150 65) Polish Website. Probably involves ditching the Docbook
151 mess in favour of wml. There are already several threads
152 in the mailinglist archives about this. See also #41.
154 66) Stop hard-coding the number of action and filter files.
156 67) Clean up source code directory layout. Depends on 54 so
157 we don't lose the revision history.
159 68) Use standard make syntax so we don't depend on GNU make.
161 69) Update autoconf setup (or move away from it).
163 70) If the server connection is reset but the headers are
164 received, consider passing the mess to the client instead
165 of showing the connect-failed template. Relates to #2698674.
167 74) Let Privoxy-Regression-Test optionally check that action
168 sections which disable actions actually are preceded by
169 sections that enable said actions.
171 75) Create a tool that creates Privoxy action (and filter?) files
172 out of adblock files. Could be implemented as option for
173 url-pattern-translator.pl.
175 76) Cache DNS responses. Note that this has been requested
176 several times by users, but is not a developer priority.
177 If you care about this, feel free to submit patches.
179 77) Allow to configure the IP address used in outgoing connections.
181 78) Allow to optionally use pcre's DFA algorithm.
183 79) Evaluate pcre alternatives.
185 80) Change FEATURE_EXTENDED_HOST_PATTERNS to support both
186 extended and vanilla host patterns at the same time.
188 Note that the requirement is to allow the user to decide
189 if the domain pattern should be interpreted as regex or
190 traditional host pattern and if it's not obvious that the
191 user made any decision, default to the latter.
193 Possible solutions would be:
195 1. An always-use-regex-domain-patterns config option
196 2. An enable-regex-domain-patterns-for-this-action-file option
197 3. An enable-regex-domain-patterns-for-this-action-file-until-the-user-says-otherwise option
198 4. A treat-the-domain-pattern-in-this-line-as-regex(-or-not) option
199 5. Combinations of the options above
201 With 2+4, 3+4 or 2+3+4 being the preferences until
204 82) Detect if the system time goes back in time let the user
205 know if it caused any connections to get closed.
207 84) Flesh out the user-manual delivery to serve pages from
208 other directories, too.
210 85) Once #84 is done, write a script that populates a directory with
211 various common third-party icons (stumbleupon.png, facebook.png ...)
212 and redirect requests for them to Privoxy.
214 86) Add a server-body-tagger action. This is trivial as as all the
215 functionality required to do it already exists.
217 87) Add a client-body-tagger action. This is less trivial as we currently
218 don't buffer client bodies. After 14) is implemented it would be
221 88) Investigate if there's a Perl module that Privoxy-Regression-Test
222 could optionally use to keep connections alive, preferably while
223 requiring less forks at the same time.
225 89) When multiple block actions apply, consider showing all the block
226 reasons on the blocked page that haven't been overruled, not just
229 91) Add an optional limit for internal redirects. It would probably
230 be reasonable to default to a limit of one and showing an error
231 message if the request for the redirect URL would be redirected
234 92) The statistics currently aren't calculated correctly by Privoxy
235 as each thread is only counted as one request which is no longer
236 correct. This should be fixed, or the statistic code removed.
237 Privoxy-Log-Parser's provides more detailed statistics, anyway.
239 93) Add a config directive to let Privoxy explicitly request either
240 IPv4 (or IPv6) addresses, even if the system supports both.
241 Could be useful as a workaround for misconfigured setups where the
242 libc returns IPv6 addresses even if there's no IPv6 connectivity.
244 94) Add a config directive to let Privoxy prefer either IPv4 (or IPv6)
245 addresses, instead of trusting the libc to return them in an order
246 that makes sense. Like #93, this could be useful as a workaround
247 for misconfigured setups.
249 95) Support a non-standard client header in CONNECT requests that
250 contains the URL of the requested resource, which is then treated
251 like the request URL.
253 This way the client could opt-in for path-based blocking of https
254 requests. Given that the headers from the CONNECT request aren't
255 forwarded to the destination server, an unencrypted URL should be
256 acceptable if the client and Privoxy are running on the same system
257 or in a trusted environment.
259 96) Filters should be easier to look up. Currently get_filter() has to
260 go through all filters and skip the filter types the caller isn't
263 98) When showing action section on the CGI pages, properly escape
264 line breaks so they can be copy&pasted into action files without
267 99) Figure out a mechanism through which a user can easily enable
268 site-specific action sections that are too aggressive to be
269 enabled by default. This could be similar to the presettings
270 in default.action, but could also be just another action file
271 that isn't used by default.
273 100) Create a cross-platform Privoxy control program and retire
274 the win32 GUI. Integrate support for Privoxy-Regression-Test,
275 Privoxy-Log-Parser, Privoxy-Filter-Test, uagen and similar tools.
277 102) Add an include directive to split the config file into several parts.
279 103) Potential performance improvement for large action files:
280 when figuring out which actions apply, check the action bit mask
281 before pattern matching and skip section that wouldn't modify the
282 actions already set. To increase the impact the sections would have
283 to be applied in reverse.
285 104) The code to modify global_toggle_state should be factored out into
286 a separate function. Currently we mess with it in three different
287 files, but only in w32log.c the tray icon is explicitly set.
288 The logging is inconsistent as well. For details see #3525694.
290 105) Add support for socks authentication.
292 106) actionlist.h should be embedded in a way that causes less text
295 107) Support more pcrs variables, for example $destination-ip-address
296 and $source-ip-address.
298 108) Allow to use a somewhat random string instead of PRIVOXY-FORCE.
300 109) Let log_error() support the format specifier %S which should
301 work like %s but escape new lines like %N. This would be useful
302 to log the result of header filters which may inject new lines.
304 110) Add a global-buffer-limit directive that roughly limits how
305 much malloc'ed memory Privoxy will use and can potentially
306 be smaller than (buffer-limit * max-client-connections).
308 111) Reject requests if hosts and ports in request line and Host
309 header don't match (before filters have been applied).
311 112) If a header filter is used to inject another header by inserting
312 a \r\n (undocumented feature), detect it and split the headers so
313 following header actions do not treat them as a single string.
314 Alternatively add another header injection mechanism.
316 113) Log statistics upon receiving a certain signal (SIGINFO or SIGUSR1).
318 114) Properly deal with status code 100. The current "Continue hack"
319 can cause problems for gpg when uploading keys through Privoxy.
321 115) Add ICAP (RFC 3507) support. FR #3615158.
323 116) Due to the use of sscanf(), Privoxy currently will fail to properly
324 parse chunks whose size can't be represented with 32 bit. This is
325 unlikely to cause problems in the real world, but should eventually
326 be fixed anyway. See also:
327 https://bugzilla.mozilla.org/show_bug.cgi?id=959100
329 118) There should be "escaped" dynamic variables that are guaranteed
330 not to break filters.
332 119) Evaluate using pcre's jit mode.
334 120) Add an option to limit pcre's recursion limit below the default.
335 On some platforms the recursion limit doesn't prevent pcre from
336 running out of stack space, causing the kernel to kill Privoxy
339 121) Add HTTP/2 support. As a first step, incomming HTTP/1.x requests
340 should be translated to outgoing HTTP/2 requests where possible
341 (and if desired by the user).
343 122) Allow customized log messages.
345 123) Evaluate if the voluntarily-disclose-session-keys option in Firefox
346 (and other browsers) can be leveraged. Probably depends on #16.
348 124) Add support for the "lightweight OS capability and sandbox framework"
349 Capsicum. http://www.cl.cam.ac.uk/research/security/capsicum/
351 125) Allow clients to HTTPS-encrypt the proxy connection.
353 126) Run the Co-Advisor HTTP compliance tests, evaluate the results,
354 fix the compliance issues that aren't by design and document
356 Note that Privoxy developers qualified for free account upgrades:
357 http://coad.measurement-factory.com/details.html#pricing
359 127) Add "real" CGI support (serve program output instead of forwarding
360 the request). The work is mostly done due to +external-filter{}.
362 128) Add a config directive to control the stack limit.
364 129) Completely implement RFC 7230 4.1 (Chunked Transfer Coding).
365 Currently Privoxy doesn't properly deal with trailers which
366 are rarely used in the real world but should be supported anyway.
368 130) Move header_tagger() out of the parser structs and let it execute
369 taggers one-by-one against all headers so the header order has less
370 influence on the tagging result. As a bonus, dynamic taggers would
371 have to be compiled less often.
373 131) The handle-as-empty-doc-returns-ok directive should be replaced with
374 an action so the behaviour can be enabled on a per-request basis.
376 ##########################################################################
378 Hosting wish list (relevant for #53)
383 - Mailinglists (Mailman with public archives preferred)
384 - Webspace (on a Unix-like OS that works with the webserver targets
386 - Source code repositories (currently CVS, but migrating away
387 from it is TODO #54 anyway and shouldn't be too much trouble)
388 - Commit mails (preferably with unified diffs)
390 (Unsorted) details to look at when evaluating hosters:
392 1. Preferably no third-party ads and trackers.
393 External images, CSS and JavaScript may count as trackers
394 but texts like "supported by company XYZ" may be acceptable.
396 2. JavaScript should be optional or not used at all.
398 3. Services we don't need shouldn't be enabled anyway.
399 (We currently don't use Web forums, wikis, surveys etc.)
401 4. It would be preferable if the hoster didn't have a bad track
402 record as far as user experience, security and privacy are
403 concerned and if the terms of service are "reasonable" and
404 haven't changed too often in the past. Updates in the past
405 should have been improvements and not regressions.
407 5. It would be preferable if most of the server administration
408 is done by a trusted third-party (or at least not a lot of work
411 6. The server(s) should be located in a country with laws we can
412 understand and follow (or at least not unintentionally violate).
414 7. A server location in a country with some kind of due process
415 and strong data protection laws (at least on paper) would be
418 8. Given that Privoxy is a free software project it would be
419 preferable if the hoster would use free software where possible.
421 9. Migrating away from the hoster in the future without losing
422 any important data should be possible without writing web