1 Privoxy Developer Manual
3 Copyright © 2001, 2002 by Privoxy Developers
5 $Id: developer-manual.sgml,v 2.3 2002/09/05 02:27:59 hal9 Exp $
7 The developer manual provides guidance on coding, testing, packaging,
8 documentation and other issues of importance to those involved with Privoxy
9 development. It is mandatory (and helpful!) reading for anyone who wants to
12 Please note that this document is constantly evolving. This copy represents the
13 state at the release of version 3.1.1. You can find the latest version of the
14 this manual at http://www.privoxy.org/developer-manual/. Please see the Contact
15 section on how to contact the developers.
17 -------------------------------------------------------------------------------
22 1.1. Quickstart to Privoxy Development
28 2.3. CVS Commit Guidelines
30 3. Documentation Guidelines
32 3.1. Quickstart to Docbook and SGML
33 3.2. Privoxy Documentation Style
34 3.3. Privoxy Custom Entities
41 4.2.1. Comment, Comment, Comment
42 4.2.2. Use blocks for comments
43 4.2.3. Keep Comments on their own line
44 4.2.4. Comment each logical step
45 4.2.5. Comment All Functions Thoroughly
46 4.2.6. Comment at the end of braces if the content is more than one
49 4.3. Naming Conventions
53 4.3.3. Header file prototypes
54 4.3.4. Enumerations, and #defines
59 4.4.1. Put braces on a line by themselves.
60 4.4.2. ALL control statements should have a block
61 4.4.3. Do not belabor/blow-up boolean expressions
62 4.4.4. Use white space freely because it is free
63 4.4.5. Don't use white space around structure operators
64 4.4.6. Make the last brace of a function stand out
65 4.4.7. Use 3 character indentions
69 4.5.1. Initialize all variables
73 4.6.1. Name functions that return a boolean as a question.
74 4.6.2. Always specify a return type for a function.
75 4.6.3. Minimize function calls when iterating by using variables
76 4.6.4. Pass and Return by Const Reference
77 4.6.5. Pass and Return by Value
78 4.6.6. Names of include files
79 4.6.7. Provide multiple inclusion protection
80 4.6.8. Use `extern "C"` when appropriate
81 4.6.9. Where Possible, Use Forward Struct Declaration Instead of
84 4.7. General Coding Practices
86 4.7.1. Turn on warnings
87 4.7.2. Provide a default case for all switch statements
88 4.7.3. Try to avoid falling through cases in a switch statement.
89 4.7.4. Use 'long' or 'short' Instead of 'int'
90 4.7.5. Don't mix size_t and other types
91 4.7.6. Declare each variable and struct on its own line.
92 4.7.7. Use malloc/zalloc sparingly
93 4.7.8. The Programmer Who Uses 'malloc' is Responsible for Ensuring
95 4.7.9. Add loaders to the `file_list' structure and in order
96 4.7.10. "Uncertain" new code and/or changes to existing code, use FIXME
98 4.8. Addendum: Template for files and function comment blocks:
100 5. Testing Guidelines
102 5.1. Testplan for releases
105 6. Releasing a New Version
108 6.2. Before the Release: Freeze
109 6.3. Building and Releasing the Packages
111 6.3.1. Note on Privoxy Packaging
112 6.3.2. Source Tarball
113 6.3.3. SuSE, Conectiva or Red Hat RPM
124 6.4. Uploading and Releasing Your Package
125 6.5. After the Release
127 7. Update the Webserver
128 8. Contacting the developers, Bug Reporting and Feature Requests
132 8.3. Request New Features
133 8.4. Report Ads or Other Actions-Related Problems
136 9. Privoxy Copyright, License and History
145 Privoxy, as an heir to Junkbuster, is an Open Source project and licensed under
146 the GPL. As such, Privoxy development is potentially open to anyone who has the
147 time, knowledge, and desire to contribute in any capacity. Our goals are simply
148 to continue the mission, to improve Privoxy, and to make it available to as
149 wide an audience as possible.
151 One does not have to be a programmer to contribute. Packaging, testing, and
152 porting, are all important jobs as well.
154 -------------------------------------------------------------------------------
156 1.1. Quickstart to Privoxy Development
158 You'll need an account on Sourceforge to support our development. Mail your ID
159 to the list and wait until a project manager has added you.
161 You will also need to have a cvs package installed, which will entail having
162 ssh installed as well (which seems to be a requirement of SourceForge), in
163 order to access the cvs repository. Having the GNU build tools is also going to
164 be important (particularly, autoconf and gmake).
166 For the time being (read, this section is under construction), please refer to
167 the extensive comments in the source code.
169 -------------------------------------------------------------------------------
171 2. The CVS Repository
173 If you intend to help us with programming, documentation or packaging you will
174 need write access to our holy grail, the CVS repository. Please read this
175 chapter completely before accessing via CVS.
177 -------------------------------------------------------------------------------
181 The project's CVS repository is hosted on SourceForge. Please refer to the
182 chapters 6 and 7 in SF's site documentation for the technical access details
183 for your operating system. For historical reasons, the CVS server is called
184 cvs.ijbswa.sourceforge.net, the repository is called ijbswa, and the source
185 tree module is called current.
187 -------------------------------------------------------------------------------
191 Within the CVS repository, there are modules and branches. As mentioned, the
192 sources are in the current "module". Other modules are present for platform
193 specific issues. There is a webview of the CVS hierarchy at http://
194 cvs.sourceforge.net/cgi-bin/viewcvs.cgi/ijbswa/, which might help with
195 visualizing how these pieces fit together.
197 Branches are used to fork a sub-development path from the main trunk. Within
198 the current module where the sources are, there is always at least one "branch"
199 from the main trunk devoted to a stable release series. The main trunk is where
200 active development takes place for the next stable series (e.g. 3.2.x). And for
201 testing bugfixes for the stable series. Just prior to each stable series (e.g.
202 3.0.x), a branch is created just for stable series releases (e.g. 3.0.0 ->
203 3.0.1 -> 3.0.2, etc). Once the initial stable release of any stable branch has
204 taken place, this branch is only used for bugfixes, which have had prior
205 testing before being committed to CVS. (See Version Numbers below for details
208 -------------------------------------------------------------------------------
210 2.3. CVS Commit Guidelines
212 The source tree is the heart of every software project. Every effort must be
213 made to ensure that it is readable, compilable and consistent at all times.
214 There are differing guidelines for the stable branch and the main development
215 trunk, and we ask anyone with CVS access to strictly adhere to the following
218 Basic Guidelines, for all branches:
220 * Never (read: never, ever) be tempted to commit that small change without
221 testing it thoroughly first. When we're close to a public release, ask a
222 fellow developer to review your changes.
224 * Your commit message should give a concise overview of what you changed (no
225 big details) and why you changed it Just check previous messages for good
228 * Don't use the same message on multiple files, unless it equally applies to
231 * If your changes span multiple files, and the code won't recompile unless
232 all changes are committed (e.g. when changing the signature of a function),
233 then commit all files one after another, without long delays in between. If
234 necessary, prepare the commit messages in advance.
236 * Before changing things on CVS, make sure that your changes are in line with
237 the team's general consensus on what should be done.
239 * Note that near a major public release, we get more cautious. There is
240 always the possibility to submit a patch to the patch tracker instead.
242 Stable branches are handled with decidedly more care, especially after the
243 initial *.*.0 release, and we are just in bugfix mode. In addition to the
244 above, the below applies only to the stable branch (currently the
245 v_3_0_branchpoint branch):
247 * Do not commit anything into the stable branch, unless immediately before a
248 new release! There needs to be testing done before it hits CVS, and to
249 ensure that all changes are appropriate just to fix whatever the problem
252 * Where possible, bugfixes and changes should be tested in the main
253 development trunk first. There may be occasions where this is not feasible,
256 * Alternately, proposed changes can be submitted as patches to the patch
257 tracker on Sourceforge first: http://sourceforge.net/tracker/?group_id=
258 11118&atid=311118. Then ask for peer review.
260 * Do not commit anything unless your proposed changes have been well tested
261 first, by other members of the project, and have prior approval of the
262 project leaders or consensus of the devel list.
264 * Do not even think about anything except bugfixes. No new features!
266 -------------------------------------------------------------------------------
268 3. Documentation Guidelines
270 All formal documents are maintained in Docbook SGML and located in the doc/
271 source/* directory. You will need Docbook, the Docbook DTD's and the Docbook
272 modular stylesheets (or comparable alternatives), and either jade or openjade
273 (recommended) installed in order to build docs from source. Currently there is
274 user-manual, FAQ, and, of course this, the developer-manual in this format. The
275 README, AUTHORS privoxy.1 (man page), and config files are also now maintained
276 as Docbook SGML. These files, when built, in the top-level source directory are
277 generated files! Also, the Privoxy index.html (and a variation on this file,
278 privoxy-index.html, meant for inclusion with doc packages), are maintained as
279 SGML as well. DO NOT edit these directly. Edit the SGML source, or contact
280 someone involved in the documentation (at present Hal).
282 config requires some special handling. The reason it is maintained this way is
283 so that the extensive comments in the file mirror those in user-manual. But the
284 conversion process requires going from SGML to HTML to text to special
285 formatting required for the embedded comments. Some of this does not survive so
286 well. Especially some of the examples that are longer than 80 characters. The
287 build process for this file outputs to config.new, which should be reviewed for
288 errors and mis-formatting. Once satisfied that it is correct, then it should be
289 hand copied to config.
291 Other, less formal documents (e.g. LICENSE, INSTALL) are maintained as plain
292 text files in the top-level source directory. At least for the time being.
294 Packagers are encouraged to include this documentation. For those without the
295 ability to build the docs locally, text versions of each are kept in CVS. HTML
296 versions are also now being kept in CVS under doc/webserver/*.
298 Formal documents are built with the Makefile targets of make dok, or
299 alternately make redhat-dok. If you have problems, try both. The build process
300 uses the document SGML sources in doc/source/*/* to update all text files in
301 doc/text/ and to update all HTML documents in doc/webserver/.
303 Documentation writers should please make sure documents build successfully
304 before committing to CVS, if possible.
306 How do you update the webserver (i.e. the pages on privoxy.org)?
308 1. First, build the docs by running make dok (or alternately make redhat-dok).
309 For PDF docs, do make dok-pdf.
311 2. Run make webserver which copies all files from doc/webserver to the
312 sourceforge webserver via scp.
314 Finished docs should be occasionally submitted to CVS (doc/webserver/*/*.html)
315 so that those without the ability to build them locally, have access to them if
316 needed. This is especially important just prior to a new release! Please do
317 this after the $VERSION and other release specific data in configure.in has
318 been updated (this is done just prior to a new release).
320 -------------------------------------------------------------------------------
322 3.1. Quickstart to Docbook and SGML
324 If you are not familiar with SGML, it is a markup language similar to HTML.
325 Actually, not a mark up language per se, but a language used to define markup
326 languages. In fact, HTML is an SGML application. Both will use "tags" to format
327 text and other content. SGML tags can be much more varied, and flexible, but do
328 much of the same kinds of things. The tags, or "elements", are definable in
329 SGML. There is no set "standards". Since we are using Docbook, our tags are
330 those that are defined by Docbook. Much of how the finish document is rendered
331 is determined by the "stylesheets". The stylesheets determine how each tag gets
332 translated to HTML, or other formats.
334 Tags in Docbook SGML need to be always "closed". If not, you will likely
335 generate errors. Example: <title>My Title</title>. They are also
336 case-insensitive, but we strongly suggest using all lower case. This keeps
337 compatibility with [Docbook] XML.
339 Our documents use "sections" for the most part. Sections will be processed into
340 HTML headers (e.g. h1 for sect1). The Docbook stylesheets will use these to
341 also generate the Table of Contents for each doc. Our TOC's are set to a depth
342 of three. Meaning sect1, sect2, and sect3 will have TOC entries, but sect4 will
343 not. Each section requires a <title> element, and at least one <para>. There is
344 a limit of five section levels in Docbook, but generally three should be
345 sufficient for our purposes.
347 Some common elements that you likely will use:
349 <para></para>, paragraph delimiter. Most text needs to be within paragraph
350 elements (there are some exceptions).
351 <emphasis></emphasis>, the stylesheets make this italics.
352 <filename></filename>, files and directories.
353 <command></command>, command examples.
354 <literallayout></literallayout>, like <pre>, more or less.
355 <itemizedlist></itemizedlist>, list with bullets.
356 <listitem></listitem>, member of the above.
357 <screen></screen>, screen output, implies <literallayout>.
358 <ulink url="example.com"></ulink>, like HTML <a> tag.
359 <quote></quote>, for, doh, quoting text.
361 Look at any of the existing docs for examples of all these and more.
363 You might also find "Writing Documentation Using DocBook - A Crash Course"
366 -------------------------------------------------------------------------------
368 3.2. Privoxy Documentation Style
370 It will be easier if everyone follows a similar writing style. This just makes
371 it easier to read what someone else has written if it is all done in a similar
376 * All tags should be lower case.
378 * Tags delimiting a block of text (even small blocks) should be on their own
386 Tags marking individual words, or few words, should be in-line:
388 Just to <emphasis>emphasize</emphasis>, some text goes here.
391 * Tags should be nested and step indented for block text like: (except
398 Some text goes here in our list example.
405 This makes it easier to find the text amongst the tags ;-)
407 * Use white space to separate logical divisions within a document, like
408 between sections. Running everything together consistently makes it harder
411 * Do not hesitate to make comments. Comments can either use the <comment>
412 element, or the <!-- --> style comment familiar from HTML. (Note in Docbook
413 v4.x <comment> is replaced by <remark>.)
415 * We have an international audience. Refrain from slang, or English
416 idiosyncrasies (too many to list :). Humor also does not translate well
419 * Try to keep overall line lengths in source files to 80 characters or less
420 for obvious reasons. This is not always possible, with lengthy URLs for
423 * Our documents are available in differing formats. Right now, they are just
424 plain text, TML, and PDF, but others are always a future possibility. Be
425 careful with URLs (<ulink>), and avoid this mistake:
427 My favorite site is <ulink url="http://example.com">here</ulink>.
429 This will render as "My favorite site is here", which is not real helpful
430 in a text doc. Better like this:
432 My favorite site is <ulink url="http://example.com">example.com</ulink>.
434 * All documents should be spell checked occasionally. aspell can check SGML
435 with the -H option. (ispell I think too.)
437 -------------------------------------------------------------------------------
439 3.3. Privoxy Custom Entities
441 Privoxy documentation is using a number of customized "entities" to facilitate
442 documentation maintenance.
444 We are using a set of "boilerplate" files with generic text, that is used by
445 multiple docs. This way we can write something once, and use it repeatedly
446 without having to re-write the same content over and over again. If editing
447 such a file, keep in mind that it should be generic. That is the purpose; so it
448 can be used in varying contexts without additional modifications.
450 We are also using what Docbook calls "internal entities". These are like
451 variables in programming. Well, sort of. For instance, we have the p-version
452 entity that contains the current Privoxy version string. You are strongly
453 encouraged to use these where possible. Some of these obviously require
454 re-setting with each release (done by the Makefile). A sampling of custom
455 entities are listed below. See any of the main docs for examples.
457 * Re- "boilerplate" text entities are defined like:
459 <!entity supported SYSTEM "supported.sgml">
461 In this example, the contents of the file, supported.sgml is available for
462 inclusion anywhere in the doc. To make this happen, just reference the now
463 defined entity: &supported; (starts with an ampersand and ends with a
464 semi-colon), and the contents will be dumped into the finished doc at that
467 * Commonly used "internal entities":
469 p-version: the Privoxy version string, e.g. "3.1.1".
470 p-status: the project status, either "alpha", "beta", or "stable".
471 p-not-stable: use to conditionally include text in "not stable" releases
473 p-stable: just the opposite.
474 p-text: this doc is only generated as text.
476 There are others in various places that are defined for a specific purpose.
479 -------------------------------------------------------------------------------
485 This set of standards is designed to make our lives easier. It is developed
486 with the simple goal of helping us keep the "new and improved Privoxy"
487 consistent and reliable. Thus making maintenance easier and increasing chances
488 of success of the project.
490 And that of course comes back to us as individuals. If we can increase our
491 development and product efficiencies then we can solve more of the request for
492 changes/improvements and in general feel good about ourselves. ;->
494 -------------------------------------------------------------------------------
498 4.2.1. Comment, Comment, Comment
502 Comment as much as possible without commenting the obvious. For example do not
503 comment "aVariable is equal to bVariable". Instead explain why aVariable should
504 be equal to the bVariable. Just because a person can read code does not mean
505 they will understand why or what is being done. A reader may spend a lot more
506 time figuring out what is going on when a simple comment or explanation would
507 have prevented the extra research. Please help your brother IJB'ers out!
509 The comments will also help justify the intent of the code. If the comment
510 describes something different than what the code is doing then maybe a
511 programming error is occurring.
515 /* if page size greater than 1k ... */
516 if ( PageLength() > 1024 )
518 ... "block" the page up ...
521 /* if page size is small, send it in blocks */
522 if ( PageLength() > 1024 )
524 ... "block" the page up ...
527 This demonstrates 2 cases of "what not to do". The first is a
528 "syntax comment". The second is a comment that does not fit what
529 is actually being done.
531 -------------------------------------------------------------------------------
533 4.2.2. Use blocks for comments
537 Comments can help or they can clutter. They help when they are differentiated
538 from the code they describe. One line comments do not offer effective
539 separation between the comment and the code. Block identifiers do, by
540 surrounding the code with a clear, definable pattern.
544 /*********************************************************************
545 * This will stand out clearly in your code!
546 *********************************************************************/
547 if ( thisVariable == thatVariable )
549 DoSomethingVeryImportant();
553 /* unfortunately, this may not */
554 if ( thisVariable == thatVariable )
556 DoSomethingVeryImportant();
560 if ( thisVariable == thatVariable ) /* this may not either */
562 DoSomethingVeryImportant();
567 If you are trying to add a small logic comment and do not wish to "disrupt" the
568 flow of the code, feel free to use a 1 line comment which is NOT on the same
571 -------------------------------------------------------------------------------
573 4.2.3. Keep Comments on their own line
577 It goes back to the question of readability. If the comment is on the same line
578 as the code it will be harder to read than the comment that is on its own line.
580 There are three exceptions to this rule, which should be violated freely and
581 often: during the definition of variables, at the end of closing braces, when
582 used to comment parameters.
586 /*********************************************************************
587 * This will stand out clearly in your code,
588 * But the second example won't.
589 *********************************************************************/
590 if ( thisVariable == thatVariable )
592 DoSomethingVeryImportant();
595 if ( thisVariable == thatVariable ) /*can you see me?*/
597 DoSomethingVeryImportant(); /*not easily*/
601 /*********************************************************************
602 * But, the encouraged exceptions:
603 *********************************************************************/
604 int urls_read = 0; /* # of urls read + rejected */
605 int urls_rejected = 0; /* # of urls rejected */
609 DoSomethingVeryImportant();
613 short DoSomethingVeryImportant(
614 short firstparam, /* represents something */
615 short nextparam /* represents something else */ )
619 } /* -END- DoSomethingVeryImportant */
621 -------------------------------------------------------------------------------
623 4.2.4. Comment each logical step
627 Logical steps should be commented to help others follow the intent of the
628 written code and comments will make the code more readable.
630 If you have 25 lines of code without a comment, you should probably go back
631 into it to see where you forgot to put one.
633 Most "for", "while", "do", etc... loops _probably_ need a comment. After all,
634 these are usually major logic containers.
636 -------------------------------------------------------------------------------
638 4.2.5. Comment All Functions Thoroughly
642 A reader of the code should be able to look at the comments just prior to the
643 beginning of a function and discern the reason for its existence and the
644 consequences of using it. The reader should not have to read through the code
645 to determine if a given function is safe for a desired use. The proper
646 information thoroughly presented at the introduction of a function not only
647 saves time for subsequent maintenance or debugging, it more importantly aids in
648 code reuse by allowing a user to determine the safety and applicability of any
649 function for the problem at hand. As a result of such benefits, all functions
650 should contain the information presented in the addendum section of this
653 -------------------------------------------------------------------------------
655 4.2.6. Comment at the end of braces if the content is more than one screen
660 Each closing brace should be followed on the same line by a comment that
661 describes the origination of the brace if the original brace is off of the
662 screen, or otherwise far away from the closing brace. This will simplify the
663 debugging, maintenance, and readability of the code.
665 As a suggestion , use the following flags to make the comment and its brace
668 use following a closing brace: } /* -END- if() or while () or etc... */
674 DoSomethingVeryImportant();
675 ...some long list of commands...
676 } /* -END- if x is 1 */
682 DoSomethingVeryImportant();
683 ...some long list of commands...
684 } /* -END- if ( 1 == X ) */
686 -------------------------------------------------------------------------------
688 4.3. Naming Conventions
690 4.3.1. Variable Names
694 Use all lowercase, and separate words via an underscore ('_'). Do not start an
695 identifier with an underscore. (ANSI C reserves these for use by the compiler
696 and system headers.) Do not use identifiers which are reserved in ANSI C++.
697 (E.g. template, class, true, false, ...). This is in case we ever decide to
702 int ms_iis5_hack = 0;
706 int msiis5hack = 0; int msIis5Hack = 0;
708 -------------------------------------------------------------------------------
710 4.3.2. Function Names
714 Use all lowercase, and separate words via an underscore ('_'). Do not start an
715 identifier with an underscore. (ANSI C reserves these for use by the compiler
716 and system headers.) Do not use identifiers which are reserved in ANSI C++.
717 (E.g. template, class, true, false, ...). This is in case we ever decide to
722 int load_some_file( struct client_state *csp )
726 int loadsomefile( struct client_state *csp )
727 int loadSomeFile( struct client_state *csp )
729 -------------------------------------------------------------------------------
731 4.3.3. Header file prototypes
735 Use a descriptive parameter name in the function prototype in header files. Use
736 the same parameter name in the header file that you use in the c file.
740 (.h) extern int load_aclfile( struct client_state *csp );
741 (.c) int load_aclfile( struct client_state *csp )
745 (.h) extern int load_aclfile( struct client_state * ); or
746 (.h) extern int load_aclfile();
747 (.c) int load_aclfile( struct client_state *csp )
749 -------------------------------------------------------------------------------
751 4.3.4. Enumerations, and #defines
755 Use all capital letters, with underscores between words. Do not start an
756 identifier with an underscore. (ANSI C reserves these for use by the compiler
761 (enumeration) : enum Boolean { FALSE, TRUE };
762 (#define) : #define DEFAULT_SIZE 100;
764 Note: We have a standard naming scheme for #defines that toggle a feature in
765 the preprocessor: FEATURE_>, where > is a short (preferably 1 or 2 word)
770 #define FEATURE_FORCE 1
773 #define FORCE_PREFIX blah
774 #endif /* def FEATURE_FORCE */
776 -------------------------------------------------------------------------------
782 Spell common words out entirely (do not remove vowels).
784 Use only widely-known domain acronyms and abbreviations. Capitalize all letters
787 Use underscore (_) to separate adjacent acronyms and abbreviations. Never
788 terminate a name with an underscore.
792 #define USE_IMAGE_LIST 1
796 #define USE_IMG_LST 1 or
797 #define _USE_IMAGE_LIST 1 or
798 #define USE_IMAGE_LIST_ 1 or
799 #define use_image_list 1 or
800 #define UseImageList 1
802 -------------------------------------------------------------------------------
806 4.4.1. Put braces on a line by themselves.
810 The brace needs to be on a line all by itself, not at the end of the statement.
811 Curly braces should line up with the construct that they're associated with.
812 This practice makes it easier to identify the opening and closing braces for a
824 if ( this == that ) { ... }
828 if ( this == that ) { ... }
830 Note: In the special case that the if-statement is inside a loop, and it is
831 trivial, i.e. it tests for a condition that is obvious from the purpose of the
832 block, one-liners as above may optically preserve the loop structure and make
835 Status: developer-discretion.
839 while ( more lines are read )
841 /* Please document what is/is not a comment line here */
842 if ( it's a comment ) continue;
844 do_something( line );
847 -------------------------------------------------------------------------------
849 4.4.2. ALL control statements should have a block
853 Using braces to make a block will make your code more readable and less prone
854 to error. All control statements should have a block defined.
866 if ( this == that ) DoSomething(); DoSomethingElse();
870 if ( this == that ) DoSomething();
872 Note: The first example in "Instead of" will execute in a manner other than
873 that which the developer desired (per indentation). Using code braces would
874 have prevented this "feature". The "explanation" and "exception" from the point
877 -------------------------------------------------------------------------------
879 4.4.3. Do not belabor/blow-up boolean expressions
883 structure->flag = ( condition );
887 if ( condition ) { structure->flag = 1; } else { structure->flag = 0; }
889 Note: The former is readable and concise. The later is wordy and inefficient.
890 Please assume that any developer new to the project has at least a "good"
891 knowledge of C/C++. (Hope I do not offend by that last comment ... 8-)
893 -------------------------------------------------------------------------------
895 4.4.4. Use white space freely because it is free
899 Make it readable. The notable exception to using white space freely is listed
900 in the next guideline.
906 int anotherValue = 0;
907 int thisVariable = 0;
909 if ( thisVariable == thatVariable )
911 firstValue = oldValue + ( ( someValue - anotherValue ) - whatever )
913 -------------------------------------------------------------------------------
915 4.4.5. Don't use white space around structure operators
919 - structure pointer operator ( "->" ) - member operator ( "." ) - functions and
922 It is a general coding practice to put pointers, references, and function
923 parentheses next to names. With spaces, the connection between the object and
924 variable/function name is not as clear.
932 Instead of: aStruct -> aMember; aStruct . aMember; FunctionName ();
934 -------------------------------------------------------------------------------
936 4.4.6. Make the last brace of a function stand out
945 } /* -END- function1 */
950 } /* -END- function2 */
954 int function1( ... ) { ...code... return( retCode ); } int function2( ... ) { }
956 Note: Use 1 blank line before the closing brace and 2 lines afterward. This
957 makes the end of function standout to the most casual viewer. Although function
958 comments help separate functions, this is still a good coding practice. In
959 fact, I follow these rules when using blocks in "for", "while", "do" loops, and
960 long if {} statements too. After all whitespace is free!
962 Status: developer-discretion on the number of blank lines. Enforced is the end
963 of function comments.
965 -------------------------------------------------------------------------------
967 4.4.7. Use 3 character indentions
971 If some use 8 character TABs and some use 3 character TABs, the code can look *
972 very* ragged. So use 3 character indentions only. If you like to use TABs, pass
973 your code through a filter such as "expand -t3" before checking in your code.
977 static const char * const url_code_map[256] =
987 return( ALWAYS_TRUE );
991 return( HOW_DID_YOU_GET_HERE );
994 return( NEVER_GETS_HERE );
998 -------------------------------------------------------------------------------
1002 4.5.1. Initialize all variables
1006 Do not assume that the variables declared will not be used until after they
1007 have been assigned a value somewhere else in the code. Remove the chance of
1008 accidentally using an unassigned variable.
1016 Note: It is much easier to debug a SIGSEGV if the message says you are trying
1017 to access memory address 00000000 and not 129FA012; or arrayPtr[20] causes a
1018 SIGSEV vs. arrayPtr[0].
1020 Status: developer-discretion if and only if the variable is assigned a value
1021 "shortly after" declaration.
1023 -------------------------------------------------------------------------------
1027 4.6.1. Name functions that return a boolean as a question.
1031 Value should be phrased as a question that would logically be answered as a
1032 true or false statement
1036 ShouldWeBlockThis();
1040 -------------------------------------------------------------------------------
1042 4.6.2. Always specify a return type for a function.
1046 The default return for a function is an int. To avoid ambiguity, create a
1047 return for a function when the return has a purpose, and create a void return
1048 type if the function does not need to return anything.
1050 -------------------------------------------------------------------------------
1052 4.6.3. Minimize function calls when iterating by using variables
1056 It is easy to write the following code, and a clear argument can be made that
1057 the code is easy to understand:
1061 for ( size_t cnt = 0; cnt < blockListLength(); cnt ++ )
1066 Note: Unfortunately, this makes a function call for each and every iteration.
1067 This increases the overhead in the program, because the compiler has to look up
1068 the function each time, call it, and return a value. Depending on what occurs
1069 in the blockListLength() call, it might even be creating and destroying
1070 structures with each iteration, even though in each case it is comparing "cnt"
1071 to the same value, over and over. Remember too - even a call to blockListLength
1072 () is a function call, with the same overhead.
1074 Instead of using a function call during the iterations, assign the value to a
1075 variable, and evaluate using the variable.
1079 size_t len = blockListLength();
1081 for ( size_t cnt = 0; cnt < len; cnt ++ )
1086 Exceptions: if the value of blockListLength() *may* change or could *
1087 potentially* change, then you must code the function call in the for/while
1090 -------------------------------------------------------------------------------
1092 4.6.4. Pass and Return by Const Reference
1096 This allows a developer to define a const pointer and call your function. If
1097 your function does not have the const keyword, we may not be able to use your
1098 function. Consider strcmp, if it were defined as: extern int strcmp( char *s1,
1101 I could then not use it to compare argv's in main: int main( int argc, const
1102 char *argv[] ) { strcmp( argv[0], "privoxy" ); }
1104 Both these pointers are *const*! If the c runtime library maintainers do it, we
1107 -------------------------------------------------------------------------------
1109 4.6.5. Pass and Return by Value
1113 Most structures cannot fit onto a normal stack entry (i.e. they are not 4 bytes
1114 or less). Aka, a function declaration like: int load_aclfile( struct
1117 would not work. So, to be consistent, we should declare all prototypes with
1118 "pass by value": int load_aclfile( struct client_state *csp )
1120 -------------------------------------------------------------------------------
1122 4.6.6. Names of include files
1126 Your include statements should contain the file name without a path. The path
1127 should be listed in the Makefile, using -I as processor directive to search the
1128 indicated paths. An exception to this would be for some proprietary software
1129 that utilizes a partial path to distinguish their header files from system or
1134 #include <iostream.h> /* This is not a local include */
1135 #include "config.h" /* This IS a local include */
1139 /* This is not a local include, but requires a path element. */
1140 #include <sys/fileName.h>
1142 Note: Please! do not add "-I." to the Makefile without a _very_ good reason.
1143 This duplicates the #include "file.h" behavior.
1145 -------------------------------------------------------------------------------
1147 4.6.7. Provide multiple inclusion protection
1151 Prevents compiler and linker errors resulting from redefinition of items.
1153 Wrap each header file with the following syntax to prevent multiple inclusions
1154 of the file. Of course, replace PROJECT_H with your file name, with "." Changed
1155 to "_", and make it uppercase.
1159 #ifndef PROJECT_H_INCLUDED
1160 #define PROJECT_H_INCLUDED
1162 #endif /* ndef PROJECT_H_INCLUDED */
1164 -------------------------------------------------------------------------------
1166 4.6.8. Use `extern "C"` when appropriate
1170 If our headers are included from C++, they must declare our functions as
1171 `extern "C"`. This has no cost in C, but increases the potential re-usability
1179 #endif /* def __cplusplus */
1181 ... function definitions here ...
1185 #endif /* def __cplusplus */
1187 -------------------------------------------------------------------------------
1189 4.6.9. Where Possible, Use Forward Struct Declaration Instead of Includes
1193 Useful in headers that include pointers to other struct's. Modifications to
1194 excess header files may cause needless compiles.
1198 /*********************************************************************
1199 * We're avoiding an include statement here!
1200 *********************************************************************/
1202 extern file_list *xyz;
1204 Note: If you declare "file_list xyz;" (without the pointer), then including the
1205 proper header file is necessary. If you only want to prototype a pointer,
1206 however, the header file is unnecessary.
1208 Status: Use with discretion.
1210 -------------------------------------------------------------------------------
1212 4.7. General Coding Practices
1214 4.7.1. Turn on warnings
1218 Compiler warnings are meant to help you find bugs. You should turn on as many
1219 as possible. With GCC, the switch is "-Wall". Try and fix as many warnings as
1222 -------------------------------------------------------------------------------
1224 4.7.2. Provide a default case for all switch statements
1228 What you think is guaranteed is never really guaranteed. The value that you
1229 don't think you need to check is the one that someday will be passed. So, to
1230 protect yourself from the unknown, always have a default step in a switch
1235 switch( hash_string( cmd ) )
1237 case hash_actions_file :
1247 ... anomaly code goes here ...
1248 continue; / break; / exit( 1 ); / etc ...
1250 } /* end switch( hash_string( cmd ) ) */
1252 Note: If you already have a default condition, you are obviously exempt from
1253 this point. Of note, most of the WIN32 code calls `DefWindowProc' after the
1254 switch statement. This API call *should* be included in a default statement.
1256 Another Note: This is not so much a readability issue as a robust programming
1257 issue. The "anomaly code goes here" may be no more than a print to the STDERR
1258 stream (as in load_config). Or it may really be an ABEND condition.
1260 Status: Programmer discretion is advised.
1262 -------------------------------------------------------------------------------
1264 4.7.3. Try to avoid falling through cases in a switch statement.
1268 In general, you will want to have a 'break' statement within each 'case' of a
1269 switch statement. This allows for the code to be more readable and
1270 understandable, and furthermore can prevent unwanted surprises if someone else
1271 later gets creative and moves the code around.
1273 The language allows you to plan the fall through from one case statement to
1274 another simply by omitting the break statement within the case statement. This
1275 feature does have benefits, but should only be used in rare cases. In general,
1276 use a break statement for each case statement.
1278 If you choose to allow fall through, you should comment both the fact of the
1279 fall through and reason why you felt it was necessary.
1281 -------------------------------------------------------------------------------
1283 4.7.4. Use 'long' or 'short' Instead of 'int'
1287 On 32-bit platforms, int usually has the range of long. On 16-bit platforms,
1288 int has the range of short.
1290 Status: open-to-debate. In the case of most FSF projects (including X/
1291 GNU-Emacs), there are typedefs to int4, int8, int16, (or equivalence ... I
1292 forget the exact typedefs now). Should we add these to IJB now that we have a
1295 -------------------------------------------------------------------------------
1297 4.7.5. Don't mix size_t and other types
1301 The type of size_t varies across platforms. Do not make assumptions about
1302 whether it is signed or unsigned, or about how long it is. Do not compare a
1303 size_t against another variable of a different type (or even against a
1304 constant) without casting one of the values. Try to avoid using size_t if you
1307 -------------------------------------------------------------------------------
1309 4.7.6. Declare each variable and struct on its own line.
1313 It can be tempting to declare a series of variables all on one line. Don't.
1325 Explanation: - there is more room for comments on the individual variables -
1326 easier to add new variables without messing up the original ones - when
1327 searching on a variable to find its type, there is less clutter to "visually"
1330 Exceptions: when you want to declare a bunch of loop variables or other trivial
1331 variables; feel free to declare them on 1 line. You should, although, provide a
1332 good comment on their functions.
1334 Status: developer-discretion.
1336 -------------------------------------------------------------------------------
1338 4.7.7. Use malloc/zalloc sparingly
1342 Create a local struct (on the stack) if the variable will live and die within
1343 the context of one function call.
1345 Only "malloc" a struct (on the heap) if the variable's life will extend beyond
1346 the context of one function call.
1350 If a function creates a struct and stores a pointer to it in a
1351 list, then it should definitely be allocated via `malloc'.
1353 -------------------------------------------------------------------------------
1355 4.7.8. The Programmer Who Uses 'malloc' is Responsible for Ensuring 'free'
1359 If you have to "malloc" an instance, you are responsible for insuring that the
1360 instance is `free'd, even if the deallocation event falls within some other
1361 programmer's code. You are also responsible for ensuring that deletion is
1362 timely (i.e. not too soon, not too late). This is known as "low-coupling" and
1363 is a "good thing (tm)". You may need to offer a free/unload/destructor type
1364 function to accommodate this.
1368 int load_re_filterfile( struct client_state *csp ) { ... }
1369 static void unload_re_filterfile( void *f ) { ... }
1373 The developer cannot be expected to provide `free'ing functions for C run-time
1374 library functions ... such as `strdup'.
1376 Status: developer-discretion. The "main" use of this standard is for allocating
1377 and freeing data structures (complex or nested).
1379 -------------------------------------------------------------------------------
1381 4.7.9. Add loaders to the `file_list' structure and in order
1385 I have ordered all of the "blocker" file code to be in alpha order. It is
1386 easier to add/read new blockers when you expect a certain order.
1388 Note: It may appear that the alpha order is broken in places by POPUP tests
1389 coming before PCRS tests. But since POPUPs can also be referred to as
1390 KILLPOPUPs, it is clear that it should come first.
1392 -------------------------------------------------------------------------------
1394 4.7.10. "Uncertain" new code and/or changes to existing code, use FIXME
1398 If you have enough confidence in new code or confidence in your changes, but
1399 are not *quite* sure of the repercussions, add this:
1401 /* FIXME: this code has a logic error on platform XYZ, * attempting to fix */ #
1402 ifdef PLATFORM ...changed code here... #endif
1406 /* FIXME: I think the original author really meant this... */ ...changed code
1411 /* FIXME: new code that *may* break something else... */ ...new code here...
1413 Note: If you make it clear that this may or may not be a "good thing (tm)", it
1414 will be easier to identify and include in the project (or conversely exclude
1417 -------------------------------------------------------------------------------
1419 4.8. Addendum: Template for files and function comment blocks:
1421 Example for file comments:
1423 const char FILENAME_rcs[] = "$Id: developer-manual.sgml,v 2.3 2002/09/05 02:27:59 hal9 Exp $";
1424 /*********************************************************************
1428 * Purpose : (Fill me in with a good description!)
1430 * Copyright : Written by and Copyright (C) 2001 the SourceForge
1431 * Privoxy team. http://www.privoxy.org/
1433 * Based on the Internet Junkbuster originally written
1434 * by and Copyright (C) 1997 Anonymous Coders and
1435 * Junkbusters Corporation. http://www.junkbusters.com
1437 * This program is free software; you can redistribute it
1438 * and/or modify it under the terms of the GNU General
1439 * Public License as published by the Free Software
1440 * Foundation; either version 2 of the License, or (at
1441 * your option) any later version.
1443 * This program is distributed in the hope that it will
1444 * be useful, but WITHOUT ANY WARRANTY; without even the
1445 * implied warranty of MERCHANTABILITY or FITNESS FOR A
1446 * PARTICULAR PURPOSE. See the GNU General Public
1447 * License for more details.
1449 * The GNU General Public License should be included with
1450 * this file. If not, you can view it at
1451 * http://www.gnu.org/copyleft/gpl.html
1452 * or write to the Free Software Foundation, Inc., 59
1453 * Temple Place - Suite 330, Boston, MA 02111-1307, USA.
1458 *********************************************************************/
1463 ...necessary include files for us to do our work...
1465 const char FILENAME_h_rcs[] = FILENAME_H_VERSION;
1467 Note: This declares the rcs variables that should be added to the
1468 "show-proxy-args" page. If this is a brand new creation by you, you are free to
1469 change the "Copyright" section to represent the rights you wish to maintain.
1471 Note: The formfeed character that is present right after the comment flower box
1472 is handy for (X|GNU)Emacs users to skip the verbiage and get to the heart of
1473 the code (via `forward-page' and `backward-page'). Please include it if you
1476 Example for file header comments:
1480 #define FILENAME_H_VERSION "$Id: developer-manual.sgml,v 2.3 2002/09/05 02:27:59 hal9 Exp $"
1481 /*********************************************************************
1485 * Purpose : (Fill me in with a good description!)
1487 * Copyright : Written by and Copyright (C) 2001 the SourceForge
1488 * Privoxy team. http://www.privoxy.org/
1490 * Based on the Internet Junkbuster originally written
1491 * by and Copyright (C) 1997 Anonymous Coders and
1492 * Junkbusters Corporation. http://www.junkbusters.com
1494 * This program is free software; you can redistribute it
1495 * and/or modify it under the terms of the GNU General
1496 * Public License as published by the Free Software
1497 * Foundation; either version 2 of the License, or (at
1498 * your option) any later version.
1500 * This program is distributed in the hope that it will
1501 * be useful, but WITHOUT ANY WARRANTY; without even the
1502 * implied warranty of MERCHANTABILITY or FITNESS FOR A
1503 * PARTICULAR PURPOSE. See the GNU General Public
1504 * License for more details.
1506 * The GNU General Public License should be included with
1507 * this file. If not, you can view it at
1508 * http://www.gnu.org/copyleft/gpl.html
1509 * or write to the Free Software Foundation, Inc., 59
1510 * Temple Place - Suite 330, Boston, MA 02111-1307, USA.
1515 *********************************************************************/
1518 #include "project.h"
1524 ... function headers here ...
1527 /* Revision control strings from this header and associated .c file */
1528 extern const char FILENAME_rcs[];
1529 extern const char FILENAME_h_rcs[];
1536 #endif /* ndef _FILENAME_H */
1544 Example for function comments:
1546 /*********************************************************************
1548 * Function : FUNCTION_NAME
1550 * Description : (Fill me in with a good description!)
1553 * 1 : param1 = pointer to an important thing
1554 * 2 : x = pointer to something else
1556 * Returns : 0 => Ok, everything else is an error.
1558 *********************************************************************/
1559 int FUNCTION_NAME( void *param1, const char *x )
1566 Note: If we all follow this practice, we should be able to parse our code to
1567 create a "self-documenting" web page.
1569 -------------------------------------------------------------------------------
1571 5. Testing Guidelines
1575 -------------------------------------------------------------------------------
1577 5.1. Testplan for releases
1579 Explain release numbers. major, minor. developer releases. etc.
1581 1. Remove any existing rpm with rpm -e
1583 2. Remove any file that was left over. This includes (but is not limited to)
1591 + /etc/init.d/privoxy
1595 3. Install the rpm. Any error messages?
1597 4. start,stop,status Privoxy with the specific script (e.g. /etc/rc.d/init/
1598 privoxy stop). Reboot your machine. Does autostart work?
1600 5. Start browsing. Does Privoxy work? Logfile written?
1602 6. Remove the rpm. Any error messages? All files removed?
1604 -------------------------------------------------------------------------------
1608 Please submit test reports only with the test form at sourceforge. Three simple
1611 * Select category: the distribution you test on.
1613 * Select group: the version of Privoxy that we are about to release.
1615 * Fill the Summary and Detailed Description with something intelligent (keep
1616 it short and precise).
1618 Do not mail to the mailing list (we cannot keep track on issues there).
1620 -------------------------------------------------------------------------------
1622 6. Releasing a New Version
1624 When we release versions of Privoxy, our work leaves our cozy secret lab and
1625 has to work in the cold RealWorld[tm]. Once it is released, there is no way to
1626 call it back, so it is very important that great care is taken to ensure that
1627 everything runs fine, and not to introduce problems in the very last minute.
1629 So when releasing a new version, please adhere exactly to the procedure
1630 outlined in this chapter.
1632 The following programs are required to follow this process: ncftpput (ncftp),
1633 scp, ssh (ssh), gmake (GNU's version of make), autoconf, cvs.
1635 -------------------------------------------------------------------------------
1637 6.1. Version numbers
1639 First you need to determine which version number the release will have. Privoxy
1640 version numbers consist of three numbers, separated by dots, like in X.Y.Z
1641 (e.g. 3.0.0), where:
1643 * X, the version major, is rarely ever changed. It is increased by one if
1644 turning a development branch into stable substantially changes the
1645 functionality, user interface or configuration syntax. Majors 1 and 2 were
1646 Junkbuster, and 3 will be the first stable Privoxy release.
1648 * Y, the version minor, represents the branch within the major version. At
1649 any point in time, there are two branches being maintained: The stable
1650 branch, with an even minor, say, 2N, in which no functionality is being
1651 added and only bug-fixes are made, and 2N+1, the development branch, in
1652 which the further development of Privoxy takes place. This enables us to
1653 turn the code upside down and inside out, while at the same time providing
1654 and maintaining a stable version. The minor is reset to zero (and one) when
1655 the major is incremented. When a development branch has matured to the
1656 point where it can be turned into stable, the old stable branch 2N is given
1657 up (i.e. no longer maintained), the former development branch 2N+1 becomes
1658 the new stable branch 2N+2, and a new development branch 2N+3 is opened.
1660 * Z, the point or sub version, represents a release of the software within a
1661 branch. It is therefore incremented immediately before each code freeze. In
1662 development branches, only the even point versions correspond to actual
1663 releases, while the odd ones denote the evolving state of the sources on
1664 CVS in between. It follows that Z is odd on CVS in development branches
1665 most of the time. There, it gets increased to an even number immediately
1666 before a code freeze, and is increased to an odd number again immediately
1667 thereafter. This ensures that builds from CVS snapshots are easily
1668 distinguished from released versions. The point version is reset to zero
1669 when the minor changes.
1671 In summary, the main CVS trunk is the development branch where new features are
1672 being worked on for the next stable series. This should almost always be where
1673 the most activity takes place. There is always at least one stable branch from
1674 the trunk, e.g now it is 3.0, which is only used to release stable versions.
1675 Once the initial .0 release of the stable branch has been done, then as a rule,
1676 only bugfixes that have had prior testing should be committed to the stable
1677 branch. At that point, it is mostly "hands off". Once there are enough bugfixes
1678 to justify a new release, the version of this branch is again incremented
1679 Example: 3.0.0 -> 3.0.1 -> 3.0.2, etc are all stable releases from within the
1680 stable branch. 3.1.x is currently the main trunk, and where work on 3.2.x is
1681 taking place. If any questions, please post to the devel list before committing
1684 -------------------------------------------------------------------------------
1686 6.2. Before the Release: Freeze
1688 The following must be done by one of the developers prior to each new release.
1690 * Make sure that everybody who has worked on the code in the last couple of
1691 days has had a chance to yell "no!" in case they have pending changes/fixes
1692 in their pipelines. Announce the freeze so that nobody will interfere with
1693 last minute changes.
1695 * Increment the version number (point from odd to even in development
1696 branches!) in configure.in. (RPM spec files will need to be incremented as
1699 * If default.action has changed since last release (i.e. software release or
1700 standalone actions file release), bump up its version info to A.B in this
1703 {+add-header{X-Actions-File-Version: A.B} -filter -no-popups}
1705 Then change the version info in doc/webserver/actions/index.php, line:
1706 '$required_actions_file_version = "A.B";'
1708 * All documentation should be rebuild after the version bump. Finished docs
1709 should be then be committed to CVS (for those without the ability to build
1710 these). Some docs may require rather obscure processing tools. config, the
1711 man page (and the html version of the man page), and the PDF docs fall in
1712 this category. REAMDE, the man page, AUTHORS, and config should all also be
1713 committed to CVS for other packagers. The formal docs should be uploaded to
1714 the webserver. See the Section "Updating the webserver" in this manual for
1717 * The User Manual is also used for context sensitive help for the CGI editor.
1718 This is version sensitive, so that the user will get appropriate help for
1719 his/her release. So with each release a fresh version should be uploaded to
1720 the webserver (this is in addition to the main User Manual link from the
1721 main page since we need to keep manuals for various versions available).
1722 The CGI pages will link to something like http://privoxy.org/$(VERSION)/
1723 user-manual/. This will need to be updated for each new release. There is
1724 no Makefile target for this at this time!!! It needs to be done manually.
1726 * All developers should look at the ChangeLog and make sure noteworthy
1727 changes are referenced.
1729 * Commit all files that were changed in the above steps!
1731 * Tag all files in CVS with the version number with "cvs tag v_X_Y_Z". Don't
1732 use vX_Y_Z, ver_X_Y_Z, v_X.Y.Z (won't work) etc.
1734 * If the release was in a development branch, increase the point version from
1735 even to odd (X.Y.(Z+1)) again in configure.in and commit your change.
1737 * On the webserver, copy the user manual to a new top-level directory called
1738 X.Y.Z. This ensures that help links from the CGI pages, which have the
1739 version as a prefix, will go into the right version of the manual. If this
1740 is a development branch release, also symlink X.Y.(Z-1) to X.Y.Z and X.Y.
1741 (Z+1) to . (i.e. dot).
1743 -------------------------------------------------------------------------------
1745 6.3. Building and Releasing the Packages
1747 Now the individual packages can be built and released. Note that for GPL
1748 reasons the first package to be released is always the source tarball.
1750 For all types of packages, including the source tarball, you must make sure
1751 that you build from clean sources by exporting the right version from CVS into
1752 an empty directory (just press return when asked for a password):
1754 mkdir dist # delete or choose different name if it already exists
1756 cvs -d:pserver:anonymous@cvs.ijbswa.sourceforge.net:/cvsroot/ijbswa login
1757 cvs -z3 -d:pserver:anonymous@cvs.ijbswa.sourceforge.net:/cvsroot/ijbswa export -r v_X_Y_Z current
1759 Do NOT change a single bit, including, but not limited to version information
1760 after export from CVS. This is to make sure that all release packages, and with
1761 them, all future bug reports, are based on exactly the same code.
1763 Please find additional instructions for the source tarball and the individual
1764 platform dependent binary packages below. And details on the Sourceforge
1765 release process below that.
1767 -------------------------------------------------------------------------------
1769 6.3.1. Note on Privoxy Packaging
1771 Please keep these general guidelines in mind when putting together your
1772 package. These apply to all platforms!
1774 * Privoxy requires write access to: all *.action files, all logfiles, and the
1775 trust file. You will need to determine the best way to do this for your
1778 * Please include up to date documentation. At a bare minimum:
1780 LICENSE (top-level directory)
1782 README (top-level directory)
1784 AUTHORS (top-level directory)
1786 man page (top-level directory, Unix-like platforms only)
1788 The User Manual (doc/webserver/user-manual/)
1790 FAQ (doc/webserver/faq/)
1792 Also suggested: Developer Manual (doc/webserver/developer-manual) and
1793 ChangeLog (top-level directory). FAQ and the manuals are HTML docs. There
1794 are also text versions in doc/text/ which could conceivably also be
1797 The documentation has been designed such that the manuals are linked to
1798 each other from parallel directories, and should be packaged that way.
1799 privoxy-index.html can also be included and can serve as a focal point for
1800 docs and other links of interest (and possibly renamed to index.html). This
1801 should be one level up from the manuals. There is a link also on this page
1802 to an HTMLized version of the man page. To avoid 404 for this, it is in CVS
1803 as doc/webserver/man-page/privoxy-man-page.html, and should be included
1804 along with the manuals. There is also a css stylesheets that can be
1805 included for better presentation: p_doc.css. This should be in the same
1806 directory with privoxy-index.html, (i.e. one level up from the manual
1809 * user.action is designed for local preferences. Make sure this does not get
1812 * Other configuration files should be installed as the new defaults, but all
1813 previously installed configuration files should be preserved as backups.
1814 This is just good manners :-)
1816 * Please check platform specific notes in this doc, if you haven't done
1817 "Privoxy" packaging before for other platform specific issues. Conversely,
1818 please add any notes that you know are important for your platform (or
1819 contact one of the doc maintainers to do this if you can't).
1821 * Packagers should do a "clean" install of their package after building it.
1822 So any previous installs should be removed first to ensure the integrity of
1823 the newly built package. Then run the package for a while to make sure
1824 there are no obvious problems, before uploading.
1826 -------------------------------------------------------------------------------
1828 6.3.2. Source Tarball
1830 First, make sure that you have freshly exported the right version into an empty
1831 directory. (See "Building and releasing packages" above). Then run:
1834 autoheader && autoconf && ./configure
1840 To upload the package to Sourceforge, simply issue
1844 Go to the displayed URL and release the file publicly on Sourceforge. For the
1845 change log field, use the relevant section of the ChangeLog file.
1847 -------------------------------------------------------------------------------
1849 6.3.3. SuSE, Conectiva or Red Hat RPM
1851 In following text, replace dist with either "rh" for Red Hat or "suse" for
1854 First, make sure that you have freshly exported the right version into an empty
1855 directory. (See "Building and releasing packages" above).
1857 As the only exception to not changing anything after export from CVS, now
1858 examine the file privoxy-dist.spec and make sure that the version information
1859 and the RPM release number are correct. The RPM release numbers for each
1860 version start at one. Hence it must be reset to one if this is the first RPM
1861 for dist which is built from version X.Y.Z. Check the file list if unsure.
1862 Else, it must be set to the highest already available RPM release number for
1863 that version plus one.
1868 autoheader && autoconf && ./configure
1874 To upload the package to Sourceforge, simply issue
1876 make dist-upload rpm_packagerev
1878 where rpm_packagerev is the RPM release number as determined above. Go to the
1879 displayed URL and release the file publicly on Sourceforge. Use the release
1880 notes and change log from the source tarball package.
1882 -------------------------------------------------------------------------------
1886 First, make sure that you have freshly exported the right version into an empty
1887 directory. (See "Building and releasing packages" above). Then get the OS/2
1890 cvs -z3 -d:pserver:anonymous@cvs.ijbswa.sourceforge.net:/cvsroot/ijbswa co os2setup
1892 You will need a mix of development tools. The main compilation takes place with
1893 IBM Visual Age C++. Some ancillary work takes place with GNU tools, available
1894 from various sources like hobbes.nmsu.edu. Specificially, you will need
1895 autoheader, autoconf and sh tools. The packaging takes place with WarpIN,
1896 available from various sources, including its home page: xworkplace.
1898 Change directory to the os2setup directory. Edit the os2build.cmd file to set
1899 the final executable filename. For example,
1901 installExeName='privoxyos2_setup_X.Y.Z.exe'
1903 Next, edit the IJB.wis file so the release number matches in the PACKAGEID
1906 PACKAGEID="Privoxy Team\Privoxy\Privoxy Package\X\Y\Z"
1908 You're now ready to build. Run:
1912 You will find the WarpIN-installable executable in the ./files directory.
1913 Upload this anonymously to uploads.sourceforge.net/incoming, create a release
1914 for it, and you're done. Use the release notes and Change Log from the source
1917 -------------------------------------------------------------------------------
1921 Login to Sourceforge's compilefarm via ssh:
1923 ssh cf.sourceforge.net
1925 Choose the right operating system (not the Debian one). When logged in, make
1926 sure that you have freshly exported the right version into an empty directory.
1927 (See "Building and releasing packages" above). Then run:
1930 autoheader && autoconf && ./configure
1936 which creates a gzip'ed tar archive. Sadly, you cannot use make solaris-upload
1937 on the Sourceforge machine (no ncftpput). You now have to manually upload the
1938 archive to Sourceforge's ftp server and release the file publicly. Use the
1939 release notes and Change Log from the source tarball package.
1941 -------------------------------------------------------------------------------
1945 You should ensure you have the latest version of Cygwin (from http://
1946 www.cygwin.com/). Run the following commands from within a Cygwin bash shell.
1948 First, make sure that you have freshly exported the right version into an empty
1949 directory. (See "Building and releasing packages" above). Then get the Windows
1952 cvs -z3 -d:pserver:anonymous@cvs.ijbswa.sourceforge.net:/cvsroot/ijbswa co winsetup
1954 Then you can build the package. This is fully automated, and is controlled by
1955 winsetup/GNUmakefile. All you need to do is:
1960 Now you can manually rename privoxy_setup.exe to privoxy_setup_X_Y_Z.exe, and
1961 upload it to SourceForge. When releasing the package on SourceForge, use the
1962 release notes and Change Log from the source tarball package.
1964 -------------------------------------------------------------------------------
1968 First, make sure that you have freshly exported the right version into an empty
1969 directory. (See "Building and releasing packages" above). Then add a log entry
1970 to debian/changelog, if it is not already there, for example by running:
1972 debchange -v 3.1.1-alpha-1 "New upstream version"
1976 dpkg-buildpackage -rfakeroot -us -uc -b
1978 This will create ../privoxy_3.1.1-alpha-1_i386.deb which can be uploaded. To
1979 upload the package to Sourceforge, simply issue
1983 -------------------------------------------------------------------------------
1987 First, make sure that you have freshly exported the right version into an empty
1988 directory. (See "Building and releasing packages" above). Then get the Mac OSX
1991 cvs -z3 -d:pserver:anonymous@cvs.ijbswa.sourceforge.net:/cvsroot/ijbswa co osxsetup
1998 This will run autoheader, autoconf and configure as well as make. Finally, it
1999 will copy over the necessary files to the ./osxsetup/files directory for
2000 further processing by PackageMaker.
2002 Bring up PackageMaker with the PrivoxyPackage.pmsp definition file, modify the
2003 package name to match the release, and hit the "Create package" button. If you
2004 specify ./Privoxy.pkg as the output package name, you can then create the
2005 distributable zip file with the command:
2007 zip -r privoxyosx_setup_x.y.z.zip Privoxy.pkg
2009 You can then upload privoxyosx_setup_x.y.z.zip anonymously to
2010 uploads.sourceforge.net/incoming, create a release for it, and you're done. Use
2011 the release notes and Change Log from the source tarball package.
2013 -------------------------------------------------------------------------------
2017 Login to Sourceforge's compile-farm via ssh:
2019 ssh cf.sourceforge.net
2021 Choose the right operating system. When logged in, make sure that you have
2022 freshly exported the right version into an empty directory. (See "Building and
2023 releasing packages" above). Then run:
2026 autoheader && autoconf && ./configure
2032 which creates a gzip'ed tar archive. Sadly, you cannot use make freebsd-upload
2033 on the Sourceforge machine (no ncftpput). You now have to manually upload the
2034 archive to Sourceforge's ftp server and release the file publicly. Use the
2035 release notes and Change Log from the source tarball package.
2037 -------------------------------------------------------------------------------
2041 First, make sure that you have freshly exported the right version into an empty
2042 directory. (See "Building and releasing packages" above). Then run:
2045 autoheader && autoconf && ./configure
2049 -------------------------------------------------------------------------------
2053 First, make sure that you have freshly exported the right version into an empty
2054 directory. (See "Building and releasing packages" above). Then run:
2057 autoheader && autoconf && ./configure
2061 -------------------------------------------------------------------------------
2065 Login to Sourceforge's compilefarm via ssh:
2067 ssh cf.sourceforge.net
2069 Choose the right operating system. When logged in, make sure that you have
2070 freshly exported the right version into an empty directory. (See "Building and
2071 releasing packages" above). Then run:
2074 autoheader && autoconf && ./configure
2080 which creates a gzip'ed tar archive. Sadly, you cannot use make aix-upload on
2081 the Sourceforge machine (no ncftpput). You now have to manually upload the
2082 archive to Sourceforge's ftp server and release the file publicly. Use the
2083 release notes and Change Log from the source tarball package.
2085 -------------------------------------------------------------------------------
2087 6.4. Uploading and Releasing Your Package
2089 After the package is ready, it is time to upload it to SourceForge, and go
2090 through the release steps. The upload is done via FTP:
2092 * Upload to: ftp://upload.sourceforge.net/incoming
2096 * password: ijbswa-developers@lists.sourceforge.net
2098 Or use the make targets as described above.
2100 Once this done go to http://sourceforge.net/project/admin/editpackages.php?
2101 group_id=11118, making sure you are logged in. Find your target platform in the
2102 second column, and click Add Release. You will then need to create a new
2103 release for your package, using the format of $VERSION ($CODE_STATUS), e.g.
2106 Now just follow the prompts. Be sure to add any appropriate Release notes. You
2107 should see your freshly uploaded packages in "Step 2. Add Files To This
2108 Release". Check the appropriate box(es). Remember at each step to hit the
2109 "Refresh/Submit" buttons! You should now see your file(s) listed in Step 3.
2110 Fill out the forms with the appropriate information for your platform, being
2111 sure to hit "Update" for each file. If anyone is monitoring your platform,
2112 check the "email" box at the very bottom to notify them of the new package.
2115 If you have made errors, or need to make changes, you can go through
2116 essentially the same steps, but select Edit Release, instead of Add Release.
2118 -------------------------------------------------------------------------------
2120 6.5. After the Release
2122 When all (or: most of the) packages have been uploaded and made available, send
2123 an email to the announce mailing list, Subject: "Version X.Y.Z available for
2124 download". Be sure to include the download location, the release notes and the
2125 Changelog. Also, post an updated News item on the project page Sourceforge, and
2126 update the Home page and docs linked from the Home page (see below).
2128 -------------------------------------------------------------------------------
2130 7. Update the Webserver
2132 The webserver should be updated at least with each stable release. When
2133 updating, please follow these steps to make sure that no broken links,
2134 inconsistent contents or permission problems will occur (as it has many times
2137 If you have changed anything in the stable-branch documentation source SGML
2140 make dok dok-pdf # (or 'make redhat-dok dok-pdf' if 'make dok' doesn't work for you)
2142 That will generate doc/webserver/user-manual, doc/webserver/developer-manual,
2143 doc/webserver/faq, doc/pdf/*.pdf and doc/webserver/index.html automatically.
2145 If you changed the manual page sources, generate doc/webserver/man-page/
2146 privoxy-man-page.html by running "make man". (This is a separate target due to
2147 dependencies on some obscure perl scripts [now in CVS, but not well tested].
2148 See comments in GNUmakefile.)
2150 If you want to add new files to the webserver, create them locally in the doc/
2151 webserver/* directory (or create new directories under doc/webserver).
2153 Next, commit any changes from the above steps to CVS. All set? If these are
2154 docs in the stable branch, then do:
2158 This will do the upload to the webserver (www.privoxy.org) and ensure all files
2159 and directories there are group writable.
2161 Please do NOT use any other means of transferring files to the webserver to
2162 avoid permission problems. Also, please do not upload docs from development
2163 branches or versions. The publicly posted docs should be in sync with the last
2166 -------------------------------------------------------------------------------
2168 8. Contacting the developers, Bug Reporting and Feature Requests
2170 We value your feedback. In fact, we rely on it to improve Privoxy and its
2171 configuration. However, please note the following hints, so we can provide you
2172 with the best support:
2174 -------------------------------------------------------------------------------
2178 For casual users, our support forum at SourceForge is probably best suited:
2179 http://sourceforge.net/tracker/?group_id=11118&atid=211118
2181 All users are of course welcome to discuss their issues on the users mailing
2182 list, where the developers also hang around.
2184 -------------------------------------------------------------------------------
2188 Please report all bugs only through our bug tracker: http://sourceforge.net/
2189 tracker/?group_id=11118&atid=111118.
2191 Before doing so, please make sure that the bug has not already been submitted
2192 and observe the additional hints at the top of the submit form.
2194 Please try to verify that it is a Privoxy bug, and not a browser or site bug
2195 first. If unsure, try toggling off Privoxy, and see if the problem persists.
2196 The appendix of the user manual also has helpful information on action
2197 debugging. If you are using your own custom configuration, please try the stock
2198 configs to see if the problem is configuration related.
2200 If not using the latest version, chances are that the bug has been found and
2201 fixed in the meantime. We would appreciate if you could take the time to
2202 upgrade to the latest version (or even the latest CVS snapshot) and verify your
2203 bug, but this is not required for reporting.
2205 -------------------------------------------------------------------------------
2207 8.3. Request New Features
2209 You are welcome to submit ideas on new features or other proposals for
2210 improvement through our feature request tracker at http://sourceforge.net/
2211 tracker/?atid=361118&group_id=11118.
2213 -------------------------------------------------------------------------------
2215 8.4. Report Ads or Other Actions-Related Problems
2217 Please send feedback on ads that slipped through, innocent images that were
2218 blocked, and any other problems relating to the default.action file through our
2219 actions feedback mechanism located at http://www.privoxy.org/actions/. On this
2220 page, you will also find a bookmark which will take you back there from any
2221 troubled site and even pre-fill the form!
2223 New, improved default.action files will occasionally be made available based on
2224 your feedback. These will be announced on the ijbswa-announce list and
2225 available from our the files section of our project page.
2227 -------------------------------------------------------------------------------
2231 For any other issues, feel free to use the mailing lists. Technically
2232 interested users and people who wish to contribute to the project are also
2233 welcome on the developers list! You can find an overview of all Privoxy-related
2234 mailing lists, including list archives, at: http://sourceforge.net/mail/?
2237 -------------------------------------------------------------------------------
2239 9. Privoxy Copyright, License and History
2241 Copyright © 2001, 2002 by Privoxy Developers <developers@privoxy.org>
2243 Some source code is based on code Copyright © 1997 by Anonymous Coders and
2244 Junkbusters, Inc. and licensed under the GNU General Public License.
2246 -------------------------------------------------------------------------------
2250 Privoxy is free software; you can redistribute it and/or modify it under the
2251 terms of the GNU General Public License, version 2, as published by the Free
2252 Software Foundation.
2254 This program is distributed in the hope that it will be useful, but WITHOUT ANY
2255 WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
2256 PARTICULAR PURPOSE. See the GNU General Public License for more details, which
2257 is available from the Free Software Foundation, Inc, 59 Temple Place - Suite
2258 330, Boston, MA 02111-1307, USA.
2260 You should have received a copy of the GNU General Public License along with
2261 this program; if not, write to the
2264 Foundation, Inc. 59 Temple Place - Suite 330
2265 Boston, MA 02111-1307
2268 -------------------------------------------------------------------------------
2272 In the beginning, there was the Internet Junkbuster, by Anonymous Coders and
2273 Junkbusters Corporation. It saved many users a lot of pain in the early days of
2274 web advertising and user tracking.
2276 But the web, its protocols and standards, and with it, the techniques for
2277 forcing users to consume ads, give up autonomy over their browsing, and for
2278 spying on them, kept evolving. Unfortunately, the Internet Junkbuster did not.
2279 Version 2.0.2, published in 1998, was (and is) the last official release
2280 available from Junkbusters Corporation. Fortunately, it had been released under
2281 the GNU GPL, which allowed further development by others.
2283 So Stefan Waldherr started maintaining an improved version of the software, to
2284 which eventually a number of people contributed patches. It could already
2285 replace banners with a transparent image, and had a first version of pop-up
2286 killing, but it was still very closely based on the original, with all its
2287 limitations, such as the lack of HTTP/1.1 support, flexible per-site
2288 configuration, or content modification. The last release from this effort was
2289 version 2.0.2-10, published in 2000.
2291 Then, some developers picked up the thread, and started turning the software
2292 inside out, upside down, and then reassembled it, adding many new features
2295 The result of this is Privoxy, whose first stable release, 3.0, was released
2298 -------------------------------------------------------------------------------
2302 Other references and sites of interest to Privoxy users:
2304 http://www.privoxy.org/, the Privoxy Home page.
2306 http://www.privoxy.org/faq/, the Privoxy FAQ.
2308 http://sourceforge.net/projects/ijbswa/, the Project Page for Privoxy on
2311 http://config.privoxy.org/, the web-based user interface. Privoxy must be
2312 running for this to work. Shortcut: http://p.p/
2314 http://www.privoxy.org/actions/, to submit "misses" to the developers.
2316 http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/ijbswa/contrib/, cool and fun
2317 ideas from Privoxy users.
2319 http://www.junkbusters.com/ht/en/cookies.html, an explanation how cookies are
2320 used to track web users.
2322 http://www.junkbusters.com/ijb.html, the original Internet Junkbuster.
2324 http://www.waldherr.org/junkbuster/, Stefan Waldherr's version of Junkbuster,
2325 from which Privoxy was derived.
2327 http://privacy.net/analyze/, a useful site to check what information about you
2328 is leaked while you browse the web.
2330 http://www.squid-cache.org/, a very popular caching proxy, which is often used
2331 together with Privoxy.
2333 http://www.privoxy.org/developer-manual/, the Privoxy developer manual.