Commit Graph

295 Commits

Author SHA1 Message Date
Daniel Marjamäki af7bf5c805 removed 'explicit constructors' check. see comments in pull request #51 for the reasons. 2011-11-02 17:12:46 +01:00
Edoardo Prezioso 9dd3360cd6 Fix some complainted choices made in my older commits.
Note: probably you should do 'make clean' before using 'make' to rebuild it again. Maybe it's me but 'make' reports various errors when linking.
2011-11-01 13:23:47 +01:00
Edoardo Prezioso 433f4640a9 Fix some GCC warnings regarding the sign conversion. 2011-10-30 18:34:49 +01:00
Edoardo Prezioso 4fc7e86d7d Fixed: with checkConst and checkVariableScope wanted still required "information" settings instead of "style". 2011-10-30 10:58:15 +01:00
Richard Quirk 2ca932a3ae Add checks for explicit constructors
Single-argument constructors should be explicit. Constructors with
multiple arguments should not be marked explicit.
2011-10-29 18:24:16 +02:00
Thomas Jarosch 434fb933a8 Implement Token::findsimplematch und use it for simple patterns 2011-10-27 15:59:22 +02:00
Thomas Jarosch 54adb910ec Use Token::simpleMatch instead of Token::Match for simple patterns 2011-10-27 15:59:22 +02:00
Daniel Marjamäki 6f8e42a5af changed the astyle formatting flags 2011-10-13 20:53:06 +02:00
Kimmo Varis 3cfe7ca1a7 Move "information" errors to "style" errors.
"information" severity is documented in lib/errorlogger.h as:

  Checking information.
  Information message about the checking (process) itself. These
  messages inform about header files not found etc issues that are
  not errors in the code but something user needs to know.

It IS NOT for errors in the code. All the current "information"-
severity errors fit nicely into description of the "style"-
severity.

We definitely need to separate processing information and actual
errors in the code. It is highly confusing for users to mix these
two different things. Hence all current "information" code error
messages are moved to "style" category.

Ticket: #3165 (Stop misusing the 'information' error severity!)
2011-10-05 20:44:00 +03:00
Robert Reif 272783347b another variation of false negative from #3149 2011-09-30 16:26:08 -04:00
Robert Reif 1dcb8b2382 really fix #3149 (false negative: Technically the member function 'A::f' can be const.) 2011-09-30 11:16:34 -04:00
Robert Reif 556d523e4f fix #3149 (false negative: Technically the member function 'A::f' can be const.) 2011-09-29 20:16:52 -04:00
Robert Reif 540207533b fix #3008 (New check: Order of initialisation list) 2011-09-27 21:07:37 -04:00
Robert Reif 1717bda382 fix wrong information about constness of function 2011-09-16 18:07:25 -04:00
Robert Reif 2670525b4f better fix for #3114 (infinite recursion when operator= is overloaded) 2011-09-14 19:58:11 -04:00
Robert Reif 2bc7da2c64 quick fix for #3114 (infinite recursion when operator= is overloaded) 2011-09-14 19:04:06 -04:00
Robert Reif 2d952c65e4 fix #3051 (False positive: operator= should return reference (when function takes pointer argument)) 2011-08-28 11:14:15 -04:00
Robert Reif 9a0d076295 fix #3052 (False Positive - Technically the member function 'Example::Clear' can be const.) 2011-08-28 09:21:00 -04:00
Robert Reif 1d7ab77251 fix #3043 (False Positive - Member variable 'ProgramRecPriorityInfo::profile' is not assigned a value in 'ProgramRecPriorityInfo::operator=') 2011-08-25 23:27:10 -04:00
Robert Reif c7cb38b0b5 fix #3040 (False positive - Technically the member function 'PSIPTable::SetSection' can be const.) 2011-08-22 20:34:00 -04:00
Kimmo Varis cfcfa3f000 Use "enabled" list for the style checking.
Settings-class currently enables style checking via dedicated
boolean attribute. All other CLI's enable-options are handled
through the enable-list. This commit moves style-check enabling
to use the enable-list.

Main advantage is the consistency how options are handled/stored
in the Settings class. Which also unifies using them for the other
code. You need to enable certain type of checks? Use the
addEnabled()-method. You want to check if certain type of checks
are enabled? Use the isEnabled()-method.
2011-08-07 10:28:52 +03:00
Robert Reif ac6d67dc4d add support for using global qualified :: memset on class and class member initialization 2011-08-04 20:37:27 -04:00
Robert Reif e687ee9a9a assume all std::*::size() and std::*::empty() are const in CheckClass::checkConstFunc() 2011-07-30 10:30:31 -04:00
Robert Reif c6a804d306 add std::*.empty() to possible const functions in CheckClass::checkConstFunc 2011-07-30 09:44:20 -04:00
Robert Reif 74c1bdde77 fix #1593 (false negative: the function can be declared as const) 2011-07-30 08:48:11 -04:00
Robert Reif ca2e8b057b fix #2947 (False positive: member variable is not initialized (a[x::y] = 0;)) 2011-07-27 12:03:44 -04:00
Robert Reif 6eecab5d73 fix #2807 (Wrong handling of function references when prefixed with &) 2011-06-29 07:19:34 -04:00
Robert Reif 0f6da27b9f fix #2795 (hash operator[] is not 'const') (original patch by Sebastien Debrard) (modified to use the symbol database) 2011-06-03 22:00:27 -04:00
Robert Reif 42316f3e0b fix #2792 (false positive: Member variable 'class::m_val' is not assigned a value in 'class::operator=') 2011-05-24 20:25:34 -04:00
Robert Reif 6abf29bd22 fix spelling of initialized in checkclass messages 2011-04-26 20:37:16 -04:00
Greg Hewgill 3fc1db51d1 Change symbol database such that the typestart token skips over type modifiers (const/static/mutable).
This fixes checking for the case of a memset() on a static variable.
2011-04-23 01:13:23 +12:00
Robert Reif 79e5d2db15 run astyle 2011-04-20 19:07:28 -04:00
Daniel Marjamäki 061eab4d22 Virtual destructors: Enabled the check again. I think it is conclusive now. Ticket: #2728 2011-04-20 18:03:16 +02:00
Daniel Marjamäki e5d43d4ed2 Renamed Settings::stupid to Settings::experimental 2011-04-10 15:55:08 +02:00
Daniel Marjamäki 30ee9ba6e4 Added Settings::stupid flag that can be used to hide checking that generates false positives. 2011-04-10 13:23:45 +02:00
Robert Reif 77aebd357e Refactoring: add Token::isAssignmentOp and use it in CheckClass::checkConstFunc 2011-04-09 14:36:05 -04:00
Robert Reif a39444a40f change CheckClass::operatorEqToSelf to make better use of the symbol database 2011-03-31 19:53:35 -04:00
Robert Reif cf65a73c47 fix CheckClass::operatorEq to check return type more closely (to match check description) 2011-03-31 19:40:28 -04:00
Robert Reif dd5e15b40c output a debug message when a member variable is used that has a varid of 0 2011-03-27 14:27:14 -04:00
Robert Reif 1e1434e191 make member variable change detection simpler and more accurate for CheckClass::checkConst 2011-03-27 13:59:12 -04:00
Robert Reif c27ecf5aeb only break when we know we are inc/dec a member variable in CheckClass::checkConstFunc 2011-03-25 23:22:28 -04:00
Robert Reif 168db82fd6 better checking of assignment to array element in CheckClass::checkConstFunc 2011-03-25 23:02:13 -04:00
Robert Reif f8e1735b0f add support to CheckClass::checkConstFunc for ++/-- array elements 2011-03-25 22:37:32 -04:00
Robert Reif ca50bc7850 move 2 CheckClass helper functions to follow externally called functions so code matches comments 2011-03-25 22:21:40 -04:00
Robert Reif 61e720c82b fix #1288 (Use of memset on struct - nested structs not handled) 2011-03-23 21:58:58 -04:00
Robert Reif 14c07e988e fix pointer/reference check in CheckClass::noMemset symbol database conversion 2011-03-22 22:29:39 -04:00
Robert Reif 5314cc02b2 convert remainder of CheckClass::noMemset to use symbol database 2011-03-22 22:23:57 -04:00
Robert Reif 468a983db1 use the symbol database to look up base classes in CheckClass::noMemset check 2011-03-22 21:24:28 -04:00
Robert Reif d36ed9aff1 fix #2670 (False positive: function can be const, overloaded functions) 2011-03-22 19:23:36 -04:00
Robert Reif 2277cb6965 fix #2663 (False negative: function can be const (changing unknown or uninitialised variable)) 2011-03-20 13:29:52 -04:00