VBox-CodingGuidelines.cpp@ 73358

Last change on this file since 73358 was 72792, checked in by vboxsync, 6 years ago
doc/VBox-CodingGuidelines.cpp: Rewrote intro and added a table of contents. Put sec_vbox_guideline_local after toc.
Property svn:eol-style set to `native` Property svn:keywords set to `Author Date Id Revision`
File size: 35.9 KB

Line
1	/* $Id: VBox-CodingGuidelines.cpp 72792 2018-07-02 10:36:19Z vboxsync $ */
2	/** @file
3	* VBox - Coding Guidelines.
4	*/
5
6	/*
7	* Copyright (C) 2006-2017 Oracle Corporation
8	*
9	* This file is part of VirtualBox Open Source Edition (OSE), as
10	* available from http://www.virtualbox.org. This file is free software;
11	* you can redistribute it and/or modify it under the terms of the GNU
12	* General Public License (GPL) as published by the Free Software
13	* Foundation, in version 2 as it comes in the "COPYING" file of the
14	* VirtualBox OSE distribution. VirtualBox OSE is distributed in the
15	* hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
16	*/
17
18	/** @page pg_vbox_guideline VBox Coding Guidelines
19	*
20	* The compulsory sections of these guidelines are to be followed in all of the
21	* VBox sources. Please note that local guidelines in parts of the VBox source
22	* tree may promote the optional ones to compulsory status. The VBox tree also
23	* contains some 3rd party sources where it is good to follow the local coding
24	* style while keeping these guidelines in mind.
25	*
26	* Contents:
27	* - @ref sec_vbox_guideline_compulsory
28	* - @ref sec_vbox_guideline_compulsory_sub64
29	* - @ref sec_vbox_guideline_compulsory_cppmain
30	* - @ref sec_vbox_guideline_compulsory_cppqtgui
31	* - @ref sec_vbox_guideline_compulsory_xslt
32	* - @ref sec_vbox_guideline_compulsory_doxygen
33	* - @ref sec_vbox_guideline_compulsory_guest
34	* - @ref sec_vbox_guideline_optional
35	* - @ref sec_vbox_guideline_optional_layout
36	* - @ref sec_vbox_guideline_optional_prefix
37	* - @ref sec_vbox_guideline_optional_misc
38	* - @ref sec_vbox_guideline_warnings
39	* - @ref sec_vbox_guideline_warnings_signed_unsigned_compare
40	* - @ref sec_vbox_guideline_svn
41	*
42	* Local guidelines overrides:
43	* - src/VBox/VMM/: @ref pg_vmm_guideline (src/VBox/VMM/Docs-CodingGuidelines.cpp)
44	* - src/VBox/ValidationKit/: @ref pg_validationkit_guideline (src/VBox/ValidationKit/ValidationKitCodingGuidelines.cpp)
45	* - src/VBox/Runtime/: All of @ref sec_vbox_guideline_optional is mandatory.
46	* - src/VBox/Main/: @ref sec_vbox_guideline_compulsory_cppmain
47	* - src/VBox/Frontends/VirtualBox/: @ref sec_vbox_guideline_compulsory_cppqtgui
48	*
49	*
50	* @section sec_vbox_guideline_compulsory Compulsory
51	*
52	* <ul>
53	*
54	* <li> The indentation size is 4 chars.
55	*
56	* <li> Tabs are only ever used in makefiles.
57	*
58	* <li> Use RT and VBOX types.
59	*
60	* <li> Use Runtime functions.
61	*
62	* <li> Use the standard bool, uintptr_t, intptr_t and [u]int[1-9+]_t types.
63	*
64	* <li> Avoid using plain unsigned and int.
65	*
66	* <li> Use static wherever possible. This makes the namespace less polluted
67	* and avoids nasty name clash problems which can occur, especially on
68	* Unix-like systems. (1) It also simplifies locating callers when
69	* changing it (single source file vs entire VBox tree).
70	*
71	* <li> Public names are of the form Domain[Subdomain[]]Method, using mixed
72	* casing to mark the words. The main domain is all uppercase.
73	* (Think like java, mapping domain and subdomain to packages/classes.)
74	*
75	* <li> Public names are always declared using the appropriate DECL macro. (2)
76	*
77	* <li> Internal names starts with a lowercased main domain.
78	*
79	* <li> Defines are all uppercase and separate words with underscore.
80	* This applies to enum values too.
81	*
82	* <li> Typedefs are all uppercase and contain no underscores to distinguish
83	* them from defines.
84	*
85	* <li> Pointer typedefs start with 'P'. If pointer to const then 'PC'.
86	*
87	* <li> Function typedefs start with 'FN'. If pointer to FN then 'PFN'.
88	*
89	* <li> All files are case sensitive.
90	*
91	* <li> Slashes are unix slashes ('/') runtime converts when necessary.
92	*
93	* <li> char strings are UTF-8.
94	*
95	* <li> Strings from any external source must be treated with utmost care as
96	* they do not have to be valid UTF-8. Only trust internal strings.
97	*
98	* <li> All functions return VBox status codes. There are three general
99	* exceptions from this:
100	*
101	* <ol>
102	* <li>Predicate functions. These are function which are boolean in
103	* nature and usage. They return bool. The function name will
104	* include 'Has', 'Is' or similar.
105	* <li>Functions which by nature cannot possibly fail.
106	* These return void.
107	* <li>"Get"-functions which return what they ask for.
108	* A get function becomes a "Query" function if there is any
109	* doubt about getting what is ask for.
110	* </ol>
111	*
112	* <li> VBox status codes have three subdivisions:
113	* <ol>
114	* <li> Errors, which are VERR_ prefixed and negative.
115	* <li> Warnings, which are VWRN_ prefixed and positive.
116	* <li> Informational, which are VINF_ prefixed and positive.
117	* </ol>
118	*
119	* <li> Platform/OS operation are generalized and put in the IPRT.
120	*
121	* <li> Other useful constructs are also put in the IPRT.
122	*
123	* <li> The code shall not cause compiler warnings. Check this on ALL
124	* the platforms.
125	*
126	* <li> The use of symbols leading with single or double underscores is
127	* forbidden as that intrudes on reserved compiler/system namespace. (3)
128	*
129	* <li> All files have file headers with $Id and a file tag which describes
130	* the file in a sentence or two.
131	* Note: Use the svn-ps.cmd/svn-ps.sh utility with the -a option to add
132	* new sources with keyword expansion and exporting correctly
133	* configured.
134	*
135	* <li> All public functions are fully documented in Doxygen style using the
136	* javadoc dialect (using the 'at' instead of the 'slash' as
137	* commandprefix.)
138	*
139	* <li> All structures in header files are described, including all their
140	* members. (Doxygen style, of course.)
141	*
142	* <li> All modules have a documentation '\@page' in the main source file
143	* which describes the intent and actual implementation.
144	*
145	* <li> Code which is doing things that are not immediately comprehensible
146	* shall include explanatory comments.
147	*
148	* <li> Documentation and comments are kept up to date.
149	*
150	* <li> Headers in /include/VBox shall not contain any slash-slash C++
151	* comments, only ANSI C comments!
152	*
153	* <li> Comments on \#else indicates what begins while the comment on a
154	* \#endif indicates what ended. Only add these when there are more than
155	* a few lines (6-10) of \#ifdef'ed code, otherwise they're just clutter.
156	*
157	* <li> \#ifdefs around a single function shall be tight, i.e. no empty
158	* lines between it and the function documentation and body.
159	*
160	* <li> \#ifdefs around more than one function shall be relaxed, i.e. leave at
161	* least one line before the first function's documentation comment and
162	* one line after the end of the last function.
163	*
164	* <li> No 'else' after if block ending with 'return', 'break', or 'continue'.
165	*
166	* <li> The term 'last' is inclusive, whereas the term 'end' is exclusive.
167	*
168	* <li> Go through all of this: https://www.slideshare.net/olvemaudal/deep-c/
169	*
170	* </ul>
171	*
172	* (1) It is common practice on Unix to have a single symbol namespace for an
173	* entire process. If one is careless symbols might be resolved in a
174	* different way that one expects, leading to weird problems.
175	*
176	* (2) This is common practice among most projects dealing with modules in
177	* shared libraries. The Windows / PE __declspect(import) and
178	* __declspect(export) constructs are the main reason for this.
179	* OTOH, we do perhaps have a bit too detailed graining of this in VMM...
180	*
181	* (3) There are guys out there grepping public sources for symbols leading with
182	* single and double underscores as well as gotos and other things
183	* considered bad practice. They'll post statistics on how bad our sources
184	* are on some mailing list, forum or similar.
185	*
186	*
187	* @subsection sec_vbox_guideline_compulsory_sub64 64-bit and 32-bit
188	*
189	* Here are some amendments which address 64-bit vs. 32-bit portability issues.
190	*
191	* Some facts first:
192	*
193	* <ul>
194	*
195	* <li> On 64-bit Windows the type long remains 32-bit. On nearly all other
196	* 64-bit platforms long is 64-bit.
197	*
198	* <li> On all 64-bit platforms we care about, int is 32-bit, short is 16 bit
199	* and char is 8-bit.
200	* (I don't know about any platforms yet where this isn't true.)
201	*
202	* <li> size_t, ssize_t, uintptr_t, ptrdiff_t and similar are all 64-bit on
203	* 64-bit platforms. (These are 32-bit on 32-bit platforms.)
204	*
205	* <li> There is no inline assembly support in the 64-bit Microsoft compilers.
206	*
207	* </ul>
208	*
209	* Now for the guidelines:
210	*
211	* <ul>
212	*
213	* <li> Never, ever, use int, long, ULONG, LONG, DWORD or similar to cast a
214	* pointer to integer. Use uintptr_t or intptr_t. If you have to use
215	* NT/Windows types, there is the choice of ULONG_PTR and DWORD_PTR.
216	*
217	* <li> Avoid where ever possible the use of the types 'long' and 'unsigned
218	* long' as these differs in size between windows and the other hosts
219	* (see above).
220	*
221	* <li> RT_OS_WINDOWS is defined to indicate Windows. Do not use __WIN32__,
222	* __WIN64__ and __WIN__ because they are all deprecated and scheduled
223	* for removal (if not removed already). Do not use the compiler
224	* defined _WIN32, _WIN64, or similar either. The bitness can be
225	* determined by testing ARCH_BITS.
226	* Example:
227	* @code
228	* #ifdef RT_OS_WINDOWS
229	* // call win32/64 api.
230	* #endif
231	* #ifdef RT_OS_WINDOWS
232	* # if ARCH_BITS == 64
233	* // call win64 api.
234	* # else // ARCH_BITS == 32
235	* // call win32 api.
236	* # endif // ARCH_BITS == 32
237	* #else // !RT_OS_WINDOWS
238	* // call posix api
239	* #endif // !RT_OS_WINDOWS
240	* @endcode
241	*
242	* <li> There are RT_OS_xxx defines for each OS, just like RT_OS_WINDOWS
243	* mentioned above. Use these defines instead of any predefined
244	* compiler stuff or defines from system headers.
245	*
246	* <li> RT_ARCH_X86 is defined when compiling for the x86 the architecture.
247	* Do not use __x86__, __X86__, __[Ii]386__, __[Ii]586__, or similar
248	* for this purpose.
249	*
250	* <li> RT_ARCH_AMD64 is defined when compiling for the AMD64 architecture.
251	* Do not use __AMD64__, __amd64__ or __x64_86__.
252	*
253	* <li> Take care and use size_t when you have to, esp. when passing a pointer
254	* to a size_t as a parameter.
255	*
256	* <li> Be wary of type promotion to (signed) integer. For example the
257	* following will cause u8 to be promoted to int in the shift, and then
258	* sign extended in the assignment 64-bit:
259	* @code
260	* uint8_t u8 = 0xfe;
261	* uint64_t u64 = u8 << 24;
262	* // u64 == 0xfffffffffe000000
263	* @endcode
264	*
265	* </ul>
266	*
267	* @subsection sec_vbox_guideline_compulsory_cppmain C++ guidelines for Main
268	*
269	* Main is currently (2009) full of hard-to-maintain code that uses complicated
270	* templates. The new mid-term goal for Main is to have less custom templates
271	* instead of more for the following reasons:
272	*
273	* <ul>
274	*
275	* <li> Template code is harder to read and understand. Custom templates create
276	* territories which only the code writer understands.
277	*
278	* <li> Errors in using templates create terrible C++ compiler messages.
279	*
280	* <li> Template code is really hard to look at in a debugger.
281	*
282	* <li> Templates slow down the compiler a lot.
283	*
284	* </ul>
285	*
286	* In particular, the following bits should be considered deprecated and should
287	* NOT be used in new code:
288	*
289	* <ul>
290	*
291	* <li> everything in include/iprt/cpputils.h (auto_ref_ptr, exception_trap_base,
292	* char_auto_ptr and friends)
293	*
294	* </ul>
295	*
296	* Generally, in many cases, a simple class with a proper destructor can achieve
297	* the same effect as a 1,000-line template include file, and the code is
298	* much more accessible that way.
299	*
300	* Using standard STL templates like std::list, std::vector and std::map is OK.
301	* Exceptions are:
302	*
303	* <ul>
304	*
305	* <li> Guest Additions because we don't want to link against libstdc++ there.
306	*
307	* <li> std::string should not be used because we have iprt::MiniString and
308	* com::Utf8Str which can convert efficiently with COM's UTF-16 strings.
309	*
310	* <li> std::auto_ptr<> in general; that part of the C++ standard is just broken.
311	* Write a destructor that calls delete.
312	*
313	* </ul>
314	*
315	* @subsection sec_vbox_guideline_compulsory_cppqtgui C++ guidelines for the Qt GUI
316	*
317	* The Qt GUI is currently (2010) on its way to become more compatible to the
318	* rest of VirtualBox coding style wise. From now on, all the coding style
319	* rules described in this file are also mandatory for the Qt GUI. Additionally
320	* the following rules should be respected:
321	*
322	* <ul>
323	*
324	* <li> GUI classes which correspond to GUI tasks should be prefixed by UI (no VBox anymore)
325	*
326	* <li> Classes which extents some of the Qt classes should be prefix by QI
327	*
328	* <li> General task classes should be prefixed by C
329	*
330	* <li> Slots are prefixed by slt -> sltName
331	*
332	* <li> Signals are prefixed by sig -> sigName
333	*
334	* <li> Use Qt classes for lists, strings and so on, the use of STL classes should
335	* be avoided
336	*
337	* <li> All files like .cpp, .h, .ui, which belong together are located in the
338	* same directory and named the same
339	*
340	* </ul>
341	*
342	*
343	* @subsection sec_vbox_guideline_compulsory_xslt XSLT
344	*
345	* XSLT (eXtensible Stylesheet Language Transformations) is used quite a bit in
346	* the Main API area of VirtualBox to generate sources and bindings to that API.
347	* There are a couple of common pitfalls worth mentioning:
348	*
349	* <ul>
350	*
351	* <li> Never do repeated //interface[\@name=...] and //enum[\@name=...] lookups
352	* because they are expensive. Instead delcare xsl:key elements for these
353	* searches and do the lookup using the key() function. xsltproc uses
354	* (per current document) hash tables for each xsl:key, i.e. very fast.
355	*
356	* <li> When output type is 'text' make sure to call xsltprocNewlineOutputHack
357	* from typemap-shared.inc.xsl every few KB of output, or xsltproc will
358	* end up wasting all the time reallocating the output buffer.
359	*
360	* </ul>
361	*
362	*
363	* @subsection sec_vbox_guideline_compulsory_doxygen Doxygen Comments
364	*
365	* As mentioned above, we shall use doxygen/javadoc style commenting of public
366	* functions, typedefs, classes and such. It is mandatory to use this style
367	* everywhere!
368	*
369	* A couple of hints on how to best write doxygen comments:
370	*
371	* <ul>
372	*
373	* <li> A good class, method, function, structure or enum doxygen comment
374	* starts with a one line sentence giving a brief description of the
375	* item. Details comes in a new paragraph (after blank line).
376	*
377	* <li> Except for list generators like \@todo, \@cfgm, \@gcfgm and others,
378	* all doxygen comments are related to things in the code. So, for
379	* instance you DO NOT add a doxygen \@note comment in the middle of a
380	* because you've got something important to note, you add a normal
381	* comment like 'Note! blah, very importan blah!'
382	*
383	* <li> We do NOT use TODO/XXX/BUGBUG or similar markers in the code to flag
384	* things needing fixing later, we always use \@todo doxygen comments.
385	*
386	* <li> There is no colon after the \@todo. And it is ALWAYS in a doxygen
387	* comment.
388	*
389	* <li> The \@retval tag is used to explain status codes a method/function may
390	* returns. It is not used to describe output parameters, that is done
391	* using the \@param or \@param[out] tag.
392	*
393	* </ul>
394	*
395	* See https://www.stack.nl/~dimitri/doxygen/manual/index.html for the official
396	* doxygen documention.
397	*
398	*
399	*
400	* @subsection sec_vbox_guideline_compulsory_guest Handling of guest input
401	*
402	* First, guest input should ALWAYS be consider to be TOXIC and constructed with
403	* MALICIOUS intent! Max paranoia level!
404	*
405	* Second, when getting inputs from memory shared with the guest, be EXTREMELY
406	* careful to not re-read input from shared memory after validating it, because
407	* that will create TOCTOU problems. So, after reading input from shared memory
408	* always use the RT_UNTRUSTED_NONVOLATILE_COPY_FENCE() macro. For more details
409	* on TOCTOU: https://en.wikipedia.org/wiki/Time_of_check_to_time_of_use
410	*
411	* Thirdly, considering the recent speculation side channel issues, spectre v1
412	* in particular, we would like to be ready for future screwups. This means
413	* having input validation in a separate block of code that ends with one (or
414	* more) RT_UNTRUSTED_VALIDATED_FENCE().
415	*
416	* So the rules:
417	*
418	* <ul>
419	*
420	* <li> Mark all pointers to shared memory with RT_UNTRUSTED_VOLATILE_GUEST.
421	*
422	* <li> Copy volatile data into local variables or heap before validating
423	* them (see RT_COPY_VOLATILE() and RT_BCOPY_VOLATILE().
424	*
425	* <li> Place RT_UNTRUSTED_NONVOLATILE_COPY_FENCE() after a block copying
426	* volatile data.
427	*
428	* <li> Always validate untrusted inputs in a block ending with a
429	* RT_UNTRUSTED_VALIDATED_FENCE().
430	*
431	* <li> Use the ASSERT_GUEST_XXXX macros from VBox/AssertGuest.h to validate
432	* guest input. (Do NOT use iprt/assert.h macros.)
433	*
434	* <li> Validation of an input B may require using another input A to look up
435	* some data, in which case its necessary to insert an
436	* RT_UNTRUSTED_VALIDATED_FENCE() after validating A and before A is used
437	* for the lookup.
438	*
439	* For example A is a view identifier, idView, and B is an offset into
440	* the view's framebuffer area, offView. To validate offView (B) it is
441	* necessary to get the size of the views framebuffer region:
442	* @code
443	* uint32_t const idView = pReq->idView; // A
444	* uint32_t const offView = pReq->offView; // B
445	* RT_UNTRUSTED_NONVOLATILE_COPY_FENCE();
446	*
447	* ASSERT_GUEST_RETURN(idView < pThis->cView,
448	* VERR_INVALID_PARAMETER);
449	* RT_UNTRUSTED_VALIDATED_FENCE();
450	* const MYVIEW *pView = &pThis->aViews[idView];
451	* ASSERT_GUEST_RETURN(offView < pView->cbFramebufferArea,
452	* VERR_OUT_OF_RANGE);
453	* RT_UNTRUSTED_VALIDATED_FENCE();
454	* @endcode
455	*
456	* <li> Take care to make sure input check are not subject to integer overflow problems.
457	*
458	* For instance when validating an area, you must not just add cbDst + offDst
459	* and check against pThis->offEnd or something like that. Rather do:
460	* @code
461	* uint32_t const offDst = pReq->offDst;
462	* uint32_t const cbDst = pReq->cbDst;
463	* RT_UNTRUSTED_NONVOLATILE_COPY_FENCE();
464	*
465	* ASSERT_GUEST_RETURN( cbDst <= pThis->cbSrc
466	* && offDst < pThis->cbSrc - cbDst,
467	* VERR_OUT_OF_RANGE);
468	* RT_UNTRUSTED_VALIDATED_FENCE();
469	* @endcode
470	*
471	* <li> Input validation does not only apply to shared data cases, but also to
472	* I/O port and MMIO handlers.
473	*
474	* <li> Ditto for kernel drivers working with usermode inputs.
475	*
476	* </ul>
477	*
478	*
479	* Problem patterns:
480	* - https://en.wikipedia.org/wiki/Time_of_check_to_time_of_use
481	* - https://googleprojectzero.blogspot.de/2018/01/reading-privileged-memory-with-side.html
482	* (Variant 1 only).
483	* - https://en.wikipedia.org/wiki/Integer_overflow
484	*
485	*
486	*
487	* @section sec_vbox_guideline_optional Optional
488	*
489	* First part is the actual coding style and all the prefixes. The second part
490	* is a bunch of good advice.
491	*
492	*
493	* @subsection sec_vbox_guideline_optional_layout The code layout
494	*
495	* <ul>
496	*
497	* <li> Max line length is 130 chars. Exceptions are table-like
498	* code/initializers and Log*() statements (don't waste unnecessary
499	* vertical space on debug logging).
500	*
501	* <li> Comments should try stay within the usual 80 columns as these are
502	* denser and too long lines may be harder to read.
503	*
504	* <li> Curly brackets are not indented. Example:
505	* @code
506	* if (true)
507	* {
508	* Something1();
509	* Something2();
510	* }
511	* else
512	* {
513	* SomethingElse1().
514	* SomethingElse2().
515	* }
516	* @endcode
517	*
518	* <li> Space before the parentheses when it comes after a C keyword.
519	*
520	* <li> No space between argument and parentheses. Exception for complex
521	* expression. Example:
522	* @code
523	* if (PATMR3IsPatchGCAddr(pVM, GCPtr))
524	* @endcode
525	*
526	* <li> The else of an if is always the first statement on a line. (No curly
527	* stuff before it!)
528	*
529	* <li> else and if go on the same line if no { compound statement }
530	* follows the if. Example:
531	* @code
532	* if (fFlags & MYFLAGS_1)
533	* fFlags &= ~MYFLAGS_10;
534	* else if (fFlags & MYFLAGS_2)
535	* {
536	* fFlags &= ~MYFLAGS_MASK;
537	* fFlags \|= MYFLAGS_5;
538	* }
539	* else if (fFlags & MYFLAGS_3)
540	* @endcode
541	*
542	* <li> Slightly complex boolean expressions are split into multiple lines,
543	* putting the operators first on the line and indenting it all according
544	* to the nesting of the expression. The purpose is to make it as easy as
545	* possible to read. Example:
546	* @code
547	* if ( RT_SUCCESS(rc)
548	* \|\| (fFlags & SOME_FLAG))
549	* @endcode
550	*
551	* <li> When 'if' or 'while' statements gets long, the closing parentheses
552	* goes right below the opening parentheses. This may be applied to
553	* sub-expression. Example:
554	* @code
555	* if ( RT_SUCCESS(rc)
556	* \|\| ( fSomeStuff
557	* && fSomeOtherStuff
558	* && fEvenMoreStuff
559	* )
560	* \|\| SomePredicateFunction()
561	* )
562	* {
563	* ...
564	* }
565	* @endcode
566	*
567	* <li> The case is indented from the switch (to avoid having the braces for
568	* the 'case' at the same level as the 'switch' statement).
569	*
570	* <li> If a case needs curly brackets they contain the entire case, are not
571	* indented from the case, and the break or return is placed inside them.
572	* Example:
573	* @code
574	* switch (pCur->eType)
575	* {
576	* case PGMMAPPINGTYPE_PAGETABLES:
577	* {
578	* unsigned iPDE = pCur->GCPtr >> PGDIR_SHIFT;
579	* unsigned iPT = (pCur->GCPtrEnd - pCur->GCPtr) >> PGDIR_SHIFT;
580	* while (iPT-- > 0)
581	* if (pPD->a[iPDE + iPT].n.u1Present)
582	* return VERR_HYPERVISOR_CONFLICT;
583	* break;
584	* }
585	* }
586	* @endcode
587	*
588	* <li> In a do while construction, the while is on the same line as the
589	* closing "}" if any are used.
590	* Example:
591	* @code
592	* do
593	* {
594	* stuff;
595	* i--;
596	* } while (i > 0);
597	* @endcode
598	*
599	* <li> Comments are in C style. C++ style comments are used for temporary
600	* disabling a few lines of code.
601	*
602	* <li> No unnecessary parentheses in expressions (just don't over do this
603	* so that gcc / msc starts bitching). Find a correct C/C++ operator
604	* precedence table if needed.
605	*
606	* <li> 'for (;;)' is preferred over 'while (true)' and 'while (1)'.
607	*
608	* <li> Parameters are indented to the start parentheses when breaking up
609	* function calls, declarations or prototypes. (This is in line with
610	* how 'if', 'for' and 'while' statements are done as well.) Example:
611	* @code
612	* RTPROCESS hProcess;
613	* int rc = RTProcCreateEx(papszArgs[0],
614	* papszArgs,
615	* RTENV_DEFAULT,
616	* fFlags,
617	* NULL, // phStdIn
618	* NULL, // phStdOut
619	* NULL, // phStdErr
620	* NULL, // pszAsUser
621	* NULL, // pszPassword
622	* &hProcess);
623	* @endcode
624	*
625	* <li> That Dijkstra is dead is no excuse for using gotos.
626	*
627	* <li> Using do-while-false loops to avoid gotos is considered very bad form.
628	* They create hard to read code. They tend to be either too short (i.e.
629	* pointless) or way to long (split up the function already), making
630	* tracking the state is difficult and prone to bugs. Also, they cause
631	* the compiler to generate suboptimal code, because the break branches
632	* are by preferred over the main code flow (MSC has no branch hinting!).
633	* Instead, do make use the 130 columns (i.e. nested ifs) and split
634	* the code up into more functions!
635	*
636	* <li> Avoid code like
637	* @code
638	* int foo;
639	* int rc;
640	* ...
641	* rc = FooBar();
642	* if (RT_SUCCESS(rc))
643	* {
644	* foo = getFoo();
645	* ...
646	* pvBar = RTMemAlloc(sizeof(*pvBar));
647	* if (!pvBar)
648	* rc = VERR_NO_MEMORY;
649	* }
650	* if (RT_SUCCESS(rc))
651	* {
652	* buzz = foo;
653	* ...
654	* }
655	* @endcode
656	* The intention of such code is probably to save some horizontal space
657	* but unfortunately it's hard to read and the scope of certain varables
658	* (e.g. foo in this example) is not optimal. Better use the following
659	* style:
660	* @code
661	* int rc;
662	* ...
663	* rc = FooBar();
664	* if (RT_SUCCESS(rc))
665	* {
666	* int foo = getFoo();
667	* ...
668	* pvBar = RTMemAlloc(sizeof(*pvBar));
669	* if (pvBar)
670	* {
671	* buzz = foo;
672	* ...
673	* }
674	* else
675	* rc = VERR_NO_MEMORY;
676	* }
677	* @endcode
678	*
679	* </ul>
680	*
681	* @subsection sec_vbox_guideline_optional_prefix Variable / Member Prefixes
682	*
683	* Prefixes are meant to provide extra context clues to a variable/member, we
684	* therefore avoid using prefixes that just indicating the type if a better
685	* choice is available.
686	*
687	*
688	* The prefixes:
689	*
690	* <ul>
691	*
692	* <li> The 'g_' (or 'g') prefix means a global variable, either on file or module level.
693	*
694	* <li> The 's_' (or 's') prefix means a static variable inside a function or
695	* class. This is not used for static variables on file level, use 'g_'
696	* for those (logical, right).
697	*
698	* <li> The 'm_' (or 'm') prefix means a class data member.
699	*
700	* In new code in Main, use "m_" (and common sense). As an exception,
701	* in Main, if a class encapsulates its member variables in an anonymous
702	* structure which is declared in the class, but defined only in the
703	* implementation (like this: 'class X { struct Data; Data *m; }'), then
704	* the pointer to that struct is called 'm' itself and its members then
705	* need no prefix, because the members are accessed with 'm->member'
706	* already which is clear enough.
707	*
708	* <li> The 'a_' prefix means a parameter (argument) variable. This is
709	* sometimes written 'a' in parts of the source code that does not use
710	* the array prefix.
711	*
712	* <li> The 'p' prefix means pointer. For instance 'pVM' is pointer to VM.
713	*
714	* <li> The 'r' prefix means that something is passed by reference.
715	*
716	* <li> The 'k' prefix means that something is a constant. For instance
717	* 'enum { kStuff };'. This is usually not used in combination with
718	* 'p', 'r' or any such thing, it's main main use is to make enums
719	* easily identifiable.
720	*
721	* <li> The 'a' prefix means array. For instance 'aPages' could be read as
722	* array of pages.
723	*
724	* <li> The 'c' prefix means count. For instance 'cbBlock' could be read,
725	* count of bytes in block. (1)
726	*
727	* <li> The 'cx' prefix means width (count of 'x' units).
728	*
729	* <li> The 'cy' prefix means height (count of 'y' units).
730	*
731	* <li> The 'x', 'y' and 'z' prefix refers to the x-, y- , and z-axis
732	* respectively.
733	*
734	* <li> The 'off' prefix means offset.
735	*
736	* <li> The 'i' or 'idx' prefixes usually means index. Although the 'i' one
737	* can sometimes just mean signed integer.
738	*
739	* <li> The 'i[1-9]+' prefix means a fixed bit size variable. Frequently
740	* used with the int[1-9]+_t types where the width is really important.
741	* In most cases 'i' is more appropriate. [type]
742	*
743	* <li> The 'e' (or 'enm') prefix means enum.
744	*
745	* <li> The 'u' prefix usually means unsigned integer. Exceptions follows.
746	*
747	* <li> The 'u[1-9]+' prefix means a fixed bit size variable. Frequently
748	* used with the uint[1-9]+_t types and with bitfields where the width is
749	* really important. In most cases 'u' or 'b' (byte) would be more
750	* appropriate. [type]
751	*
752	* <li> The 'b' prefix means byte or bytes. [type]
753	*
754	* <li> The 'f' prefix means flags. Flags are unsigned integers of some kind
755	* or booleans.
756	*
757	* <li> TODO: need prefix for real float. [type]
758	*
759	* <li> The 'rd' prefix means real double and is used for 'double' variables.
760	* [type]
761	*
762	* <li> The 'lrd' prefix means long real double and is used for 'long double'
763	* variables. [type]
764	*
765	* <li> The 'ch' prefix means a char, the (signed) char type. [type]
766	*
767	* <li> The 'wc' prefix means a wide/windows char, the RTUTF16 type. [type]
768	*
769	* <li> The 'uc' prefix means a Unicode Code point, the RTUNICP type. [type]
770	*
771	* <li> The 'uch' prefix means unsigned char. It's rarely used. [type]
772	*
773	* <li> The 'sz' prefix means zero terminated character string (array of
774	* chars). (UTF-8)
775	*
776	* <li> The 'wsz' prefix means zero terminated wide/windows character string
777	* (array of RTUTF16).
778	*
779	* <li> The 'usz' prefix means zero terminated Unicode string (array of
780	* RTUNICP).
781	*
782	* <li> The 'str' prefix means C++ string; either a std::string or, in Main,
783	* a Utf8Str or, in Qt, a QString. When used with 'p', 'r', 'a' or 'c'
784	* the first letter should be capitalized.
785	*
786	* <li> The 'bstr' prefix, in Main, means a UTF-16 Bstr. When used with 'p',
787	* 'r', 'a' or 'c' the first letter should be capitalized.
788	*
789	* <li> The 'pfn' prefix means pointer to function. Common usage is 'pfnCallback'
790	* and such like.
791	*
792	* <li> The 'psz' prefix is a combination of 'p' and 'sz' and thus means
793	* pointer to a zero terminated character string. (UTF-8)
794	*
795	* <li> The 'pcsz' prefix is used to indicate constant string pointers in
796	* parts of the code. Most code uses 'psz' for const and non-const
797	* string pointers, so please ignore this one.
798	*
799	* <li> The 'l' prefix means (signed) long. We try avoid using this,
800	* expecially with the 'LONG' types in Main as these are not 'long' on
801	* 64-bit non-Windows platforms and can cause confusion. Alternatives:
802	* 'i' or 'i32'. [type]
803	*
804	* <li> The 'ul' prefix means unsigned long. We try avoid using this,
805	* expecially with the 'ULONG' types in Main as these are not 'unsigned
806	* long' on 64-bit non-Windows platforms and can cause confusion.
807	* Alternatives: 'u' or 'u32'. [type]
808	*
809	* </ul>
810	*
811	* (1) Except in the occasional 'pcsz' prefix, the 'c' prefix is never ever
812	* used in the meaning 'const'.
813	*
814	*
815	* @subsection sec_vbox_guideline_optional_misc Misc / Advice / Stuff
816	*
817	* <ul>
818	*
819	* <li> When writing code think as the reader.
820	*
821	* <li> When writing code think as the compiler. (2)
822	*
823	* <li> When reading code think as if it's full of bugs - find them and fix them.
824	*
825	* <li> Pointer within range tests like:
826	* @code
827	* if ((uintptr_t)pv >= (uintptr_t)pvBase && (uintptr_t)pv < (uintptr_t)pvBase + cbRange)
828	* @endcode
829	* Can also be written as (assuming cbRange unsigned):
830	* @code
831	* if ((uintptr_t)pv - (uintptr_t)pvBase < cbRange)
832	* @endcode
833	* Which is shorter and potentially faster. (1)
834	*
835	* <li> Avoid unnecessary casting. All pointers automatically cast down to
836	* void *, at least for non class instance pointers.
837	*
838	* <li> It's very very bad practise to write a function larger than a
839	* screen full (1024x768) without any comprehensibility and explaining
840	* comments.
841	*
842	* <li> More to come....
843	*
844	* </ul>
845	*
846	* (1) Important, be very careful with the casting. In particular, note that
847	* a compiler might treat pointers as signed (IIRC).
848	*
849	* (2) "A really advanced hacker comes to understand the true inner workings of
850	* the machine - he sees through the language he's working in and glimpses
851	* the secret functioning of the binary code - becomes a Ba'al Shem of
852	* sorts." (Neal Stephenson "Snow Crash")
853	*
854	*
855	*
856	* @section sec_vbox_guideline_warnings Compiler Warnings
857	*
858	* The code should when possible compile on all platforms and compilers without any
859	* warnings. That's a nice idea, however, if it means making the code harder to read,
860	* less portable, unreliable or similar, the warning should not be fixed.
861	*
862	* Some of the warnings can seem kind of innocent at first glance. So, let's take the
863	* most common ones and explain them.
864	*
865	*
866	* @subsection sec_vbox_guideline_warnings_signed_unsigned_compare Signed / Unsigned Compare
867	*
868	* GCC says: "warning: comparison between signed and unsigned integer expressions"
869	* MSC says: "warning C4018: '<\|<=\|==\|>=\|>' : signed/unsigned mismatch"
870	*
871	* The following example will not output what you expect:
872	@code
873	#include <stdio.h>
874	int main()
875	{
876	signed long a = -1;
877	unsigned long b = 2294967295;
878	if (a < b)
879	printf("%ld < %lu: true\n", a, b);
880	else
881	printf("%ld < %lu: false\n", a, b);
882	return 0;
883	}
884	@endcode
885	* If I understood it correctly, the compiler will convert a to an
886	* unsigned long before doing the compare.
887	*
888	*
889	*
890	* @section sec_vbox_guideline_svn Subversion Commit Rules
891	*
892	*
893	* Before checking in:
894	*
895	* <ul>
896	*
897	* <li> Check Tinderbox and make sure the tree is green across all platforms. If it's
898	* red on a platform, don't check in. If you want, warn in the \#vbox channel and
899	* help make the responsible person fix it.
900	* NEVER CHECK IN TO A BROKEN BUILD.
901	*
902	* <li> When checking in keep in mind that a commit is atomic and that the Tinderbox and
903	* developers are constantly checking out the tree. Therefore do not split up the
904	* commit unless it's into 100% independent parts. If you need to split it up in order
905	* to have sensible commit comments, make the sub-commits as rapid as possible.
906	*
907	* <li> If you make a user visible change, such as fixing a reported bug,
908	* make sure you add an entry to doc/manual/user_ChangeLogImpl.xml.
909	*
910	* <li> If you are adding files make sure set the right attributes.
911	* svn-ps.sh/cmd was created for this purpose, please make use of it.
912	*
913	* </ul>
914	*
915	* After checking in:
916	*
917	* <ul>
918	*
919	* <li> After checking-in, you watch Tinderbox until your check-ins clear. You do not
920	* go home. You do not sleep. You do not log out or experiment with drugs. You do
921	* not become unavailable. If you break the tree, add a comment saying that you're
922	* fixing it. If you can't fix it and need help, ask in the \#innotek channel or back
923	* out the change.
924	*
925	* </ul>
926	*
927	* (Inspired by mozilla tree rules.)
928	*
929	*
930	*/
931

Note: See TracBrowser for help on using the repository browser.

source: vbox/trunk/doc/VBox-CodingGuidelines.cpp@ 73358

Download in other formats: