Penguin

Differences between current version and revision by previous author of gcc(1).

Other diffs: Previous Major Revision, Previous Revision, or view the Annotated Edit History

Newer page: version 3 Last edited on Saturday, August 23, 2003 6:49:12 pm by StuartYeates
Older page: version 2 Last edited on Tuesday, June 4, 2002 12:22:03 am by perry Revert
@@ -1,6 +1,6 @@
-GCC  
 !!!GCC 
+!Table of Contents  
 NAME 
 SYNOPSIS 
 WARNING 
 DESCRIPTION 
@@ -27,9 +27,9 @@
 ---- 
 !!NAME 
  
  
-gcc, g++ - GNU project C and C++ Compiler (gcc-2.95) 
+gcc, g++ - [ GNU] project [ C] and [ C++] [ Compiler] (gcc-2.95) 
 !!SYNOPSIS 
  
  
 __gcc__ [[ ''option'' | ''filename'' ]...__ 
@@ -37,9 +37,9 @@
 !!WARNING 
  
  
 The information in this man page is an extract from the full 
-documentation of the GNU C compiler , and is limited to the 
+documentation of the [ GNU] C [Compiler] , and is limited to the 
 meaning of the options. 
  
  
 This man page is not kept up to date except when volunteers 
@@ -59,9 +59,9 @@
  
  
 For complete and current documentation, refer to the Info 
 file `__gcc__' or the manual ''Using and Porting GNU CC 
-(for version 2.)''. Both are made from the Texinfo source 
+(for version 2.)''. Both are made from the [ Texinfo|texinfo(5)] source 
 file __gcc.texinfo__. 
 !!DESCRIPTION 
  
  
@@ -89,18 +89,18 @@
 Suffixes of source file names indicate the language and kind 
 of processing to be done: 
  
  
-__.c__ C source; preprocess, compile, assemble  
-__.C__ C++ source; preprocess, compile, assemble  
-__.cc__ C++ source; preprocess, compile, assemble  
-__.cxx__ C++ source; preprocess, compile, assemble  
-__.m__ Objective-C source; preprocess, compile, assemble  
-__.i__ preprocessed C; compile, assemble  
-__.ii__ preprocessed C++; compile, assemble  
-__.s__ Assembler source; assemble  
-__.S__ Assembler source; preprocess, assemble  
-__.h__ Preprocessor file; not usually named on command line 
+ * __.c__ [ C] source; preprocess, compile, assemble  
+ * __.C__ [ C++] source; preprocess, compile, assemble  
+ * __.cc__ C++ source; preprocess, compile, assemble  
+ * __.cxx__ C++ source; preprocess, compile, assemble  
+ * __.m__ [ Objective-C] source; preprocess, compile, assemble  
+ * __.i__ preprocessed C; compile, assemble  
+ * __.ii__ preprocessed C++; compile, assemble  
+ * __.s__ [ Assembler|AssemblyLanguage] source; assemble  
+ * __.S__ Assembler source; preprocess, assemble  
+ * __.h__ Preprocessor file; not usually named on command line 
 Files with other suffixes are passed to the linker. Common cases include: 
  
  
 __.o__ Object file 
@@ -144,9 +144,9 @@
  
 __Warning Options__ 
  
  
--fsyntax-only -pedantic -pedantic-errors -w -W -Wall 
+-fsyntax-only -pedantic -pedantic-errors -w -W [ -Wall]  
 -Waggregate-return -Wcast-align -Wcast-qual -Wchar-subscript 
 -Wcomment -Wconversion -Wenum-clash -Werror -Wformat 
 -Wid-clash-''len'' -Wimplicit -Wimplicit-int 
 -Wimplicit-function-declaration -Winline -Wlong-long -Wmain 
@@ -322,12 +322,11 @@
  
 Specify explicitly the ''language'' for the following 
 input files (rather than choosing a default based on the 
 file name suffix) . This option applies to all following 
-input files until the next `__-x__' op- tion . Possible  
-values of ''language'' are `__c__', `__objec-  
-tive -c__', `__c-header__', `__c++__',  
-`__cpp-output__', `__assem- bler __', and 
+input files until the next `__-x__' option . Possible  
+values of ''language'' are `__c__', `__objective -c__', `__c-header__', `__c++__',  
+`__cpp-output__', `__assembler __', and 
 `__assembler-with-cpp__'. 
  
  
 __-x none__ 
@@ -340,9 +339,9 @@
  
  
 If you want only some of the four stages (preprocess, com- 
 pile, assemble, link), you can use `__-x__' (or filename 
-suf- fixes ) to tell __gcc__ where to start, and one of 
+suffixes ) to tell __gcc__ where to start, and one of 
 the options `__-c__', `__-S__', or `__-E__' to say 
 where __gcc__ is to stop. Note that some combinations 
 (for example, `__-x cpp-output -E__') instruct __gcc__ 
 to do nothing at all. 
@@ -351,9 +350,9 @@
 __-c__ 
  
  
 Compile or assemble the source files, but do not link. The 
-compiler output is an object file corre- sponding to each 
+compiler output is an object file corresponding to each 
 source file. 
  
  
 By default, GCC makes the object file name for a source file 
@@ -422,20 +421,20 @@
  
 __-v__ 
  
  
-Print (on standard error output) the commands exe- cuted to 
+Print (on standard error output) the commands executed to 
 run the stages of compilation. Also print the version number 
 of the compiler driver program and of the preprocessor and 
 the compiler proper. 
  
  
 __-pipe__ 
  
  
-Use pipes rather than temporary files for communi- cation  
+Use pipes rather than temporary files for communication  
 between the various stages of compilation. This fails to 
-work on some systems where the assem- bler cannot read from 
+work on some systems where the assembler cannot read from 
 a pipe; but the GNU assembler has no trouble. 
 !!LANGUAGE OPTIONS 
  
  
@@ -449,9 +448,9 @@
 Support all ANSI standard C programs. 
  
  
 This turns off certain features of GNU C that are 
-incompatible with ANSI C, such as the __asm__, 
+incompatible with [ ANSI] [ C] , such as the __asm__, 
 __inline__ and __typeof__ keywords, and predefined 
 macros such as __unix__ and __vax__ that identify the 
 type of system you are using. It also enables the 
 undesirable and rarely used ANSI trigraph feature, and 
@@ -461,11 +460,11 @@
 The alternate keywords ____asm____, ____extension____, 
 ____inline____ and ____typeof____ continue to work 
 despite `__-ansi__'. You would not want to use them in an 
 ANSI C program, of course, but it is useful to put them in 
-header files that might be included in compila- tions done 
+header files that might be included in compilations done 
 with `__-ansi__'. Alternate predefined macros such as 
-____unix____ and ____vax____ are also avail- able
+____unix____ and ____vax____ are also available
 with or without `__-ansi__'. 
  
  
 The `__-ansi__' option does not cause non-ANSI programs 
@@ -485,18 +484,18 @@
 __-fno-asm__ 
  
  
 Do not recognize __asm__, __inline__ or __typeof__ 
-as a key- word . These words may then be used as identifiers. 
+as a keyword . These words may then be used as identifiers. 
 You can use ____asm____, ____inline____ and 
-____typeof____ in- stead . `__-ansi__' implies 
+____typeof____ instead . `__-ansi__' implies 
 `__-fno-asm__'. 
  
  
 __-fno-builtin__ 
  
  
-Don't recognize built-in functions that do not be- gin with 
+Don't recognize built-in functions that do not begin with 
 two leading underscores. Currently, the functions affected 
 include ___exit__, __abort__, __abs__, __allo- 
 ca__, __cos__, __exit__, __fabs__, __labs__, 
 __memcmp__, __memcpy__, __sin__, __sqrt__, 
@@ -519,18 +518,18 @@
  
 __-ffreestanding__ 
  
  
-Compile for a freestanding environment; this im- plies the 
+Compile for a freestanding environment; this implies the 
 `__-fno-builtin__' option, and implies that __main__ 
 has no special requirements. 
  
  
 __-fno-strict-prototype__ 
  
  
 Treat a function declaration with no arguments, such as 
-`__int foo__ ();', as C would treat it--as say- ing  
+`__int foo__ ();', as C would treat it--as saying  
 nothing about the number of arguments or their types (C++ 
 only). Normally, such a declaration in C++ means that the 
 function __foo__ takes no arguments. 
  
@@ -579,9 +578,9 @@
 __-fenum-int-equiv__ 
  
  
 Permit implicit conversion of __int__ to enumeration 
-types (C++ only). Normally GNU C++ allows conver- sion of 
+types (C++ only). Normally GNU C++ allows conversion of 
 __enum__ to __int__, but not the other way 
 around. 
  
  
@@ -591,18 +590,18 @@
 Produce smaller code for template declarations, by 
 generating only a single copy of each template function 
 where it is defined (C++ only). To use this option 
 successfully, you must also mark all files that use 
-templates with either `__#pragma im- plementation __' (the  
-definition) or `__#pragma inter- face __' 
+templates with either `__#pragma implementation __' (the  
+definition) or `__#pragma interface __' 
 (declarations). 
  
  
 When your code is compiled with `__-fexternal-tem- 
 plates__', all template instantiations are external. You 
 must arrange for all necessary instantiations to appear in 
 the implementation file; you can do this with a 
-__typedef__ that references each instantia- tion needed. 
+__typedef__ that references each instantiation needed. 
 Conversely, when you compile using the default option 
 `__-fno-external-templates__', all template 
 instantiations are explicitly internal. 
  
@@ -610,9 +609,9 @@
 __-fall-virtual__ 
  
  
 Treat all possible member functions as virtual, im- 
-plicitly. All member functions (except for con- structor  
+plicitly. All member functions (except for constructor  
 functions and __new__ or __delete__ member opera- 
 tors) are treated as virtual functions of the class where 
 they appear. 
  
@@ -712,9 +711,9 @@
  
 Store string constants in the writable data segment and 
 don't uniquize them. This is for compatibility with old 
 programs which assume they can write into string constants. 
-`__-traditional__' also has this ef- fect
+`__-traditional__' also has this effect
  
  
 Writing into string constants is a very bad idea; 
 ``constants'' should be constant. 
@@ -725,10 +724,10 @@
 each C source file before actual compilation. 
  
  
 If you use the `__-E__' option, GCC does nothing except 
-pre- processing . Some of these options make sense only  
-togeth- er with `__-E__' because they cause the 
+preprocessing . Some of these options make sense only  
+together with `__-E__' because they cause the 
 preprocessor output to be unsuitable for actual 
 compilation. 
  
  
@@ -741,21 +740,21 @@
 the command line are always processed before 
 `__-include__ ''file''', regardless of the order in 
 which they are written. All the `__-include__' and 
 `__-imacros__' options are processed in the order in 
-which they are writ- ten
+which they are written
  
  
 __-imacros__ ''file'' 
  
  
 Process ''file'' as input, discarding the resulting 
 output, before processing the regular input file. Because 
-the output generated from ''file'' is discard- ed , the 
+the output generated from ''file'' is discarded , the 
 only effect of `__-imacros__ ''file''' is to make the 
 macros defined in ''file'' available for use in the main 
 input. The preprocessor evaluates any `__-D__' and 
-`__-U__' options on the command line before pro- cessing  
+`__-U__' options on the command line before processing  
 `__-imacros__''file''', regardless of the order in 
 which they are written. All the `__-include__' and 
 `__-imacros__' options are processed in the order in 
 which they are written. 
@@ -790,9 +789,9 @@
 __-nostdinc__ 
  
  
 Do not search the standard system directories for header 
-files. Only the directories you have speci- fied with 
+files. Only the directories you have specified with 
 `__-I__' options (and the current directory, if 
 appropriate) are searched. 
  
  
@@ -812,9 +811,9 @@
  
 __-undef__ 
  
  
-Do not predefine any nonstandard macros. (Includ- ing  
+Do not predefine any nonstandard macros. (Including  
 architecture flags). 
  
  
 __-E__ 
@@ -843,22 +842,22 @@
  
  
 Tell the preprocessor to output a rule suitable for 
 __make__ describing the dependencies of each object file. 
-For each source file, the preprocessor out- puts one 
+For each source file, the preprocessor outputs one 
 __make__-rule whose target is the object file name for 
 that source file and whose dependencies are all the files 
 `__#include__'d in it. This rule may be a single line or 
-may be continued with `__\__'-new- line if it is long. 
+may be continued with `__\__'-newline if it is long. 
 The list of rules is printed on standard output instead of 
 the preprocessed C program. 
  
  
 `__-M__' implies `__-E__'. 
  
  
 `__-MG__' says to treat missing header files as gener- 
-ated files and assume they live in the same direc- tory as 
+ated files and assume they live in the same directory as 
 the source file. It must be specified in addition to 
 `__-M__'. 
  
  
@@ -877,9 +876,9 @@
  
 Like `__-M__' but the dependency information is written 
 to files with names made by replacing `__.o__' with 
 `__.d__' at the end of the output file names. This is in 
-addition to compiling the file as speci- fied --`__-MD__' 
+addition to compiling the file as specified --`__-MD__' 
 does not inhibit ordinary compilation the way `__-M__' 
 does. 
  
  
@@ -897,9 +896,9 @@
  
 __-H__ 
  
  
-Print the name of each header file used, in addi- tion to 
+Print the name of each header file used, in addition to 
 other normal activities. 
  
  
 __-A__''question''__(__''answer''__)__ 
@@ -907,9 +906,9 @@
  
 Assert the answer ''answer'' for ''question'', in case 
 it is tested with a preprocessor conditional such as `__#if 
 #__''question''__(__''answer''__)__'. 
-`__-A-__' disables the stan- dard assertions that 
+`__-A-__' disables the standard assertions that 
 normally describe the target machine. 
  
  
 __-A__''question'' 
@@ -926,9 +925,9 @@
 __-D__''macro'' 
  
  
 Define macro ''macro'' with the string `__1__' as its 
-defi- nition
+definition
  
  
 __-D__''macro''__=__''defn'' 
  
@@ -989,12 +988,12 @@
  
 ''object-file-name'' 
  
  
-A file name that does not end in a special recog- nized  
+A file name that does not end in a special recognized  
 suffix is considered to name an object file or library. 
 (Object files are distinguished from libraries by the linker 
-according to the file con- tents .) If GCC does a link step, 
+according to the file contents .) If GCC does a link step, 
 these object files are used as input to the 
 linker. 
  
  
@@ -1017,9 +1016,9 @@
  
  
 Normally the files found this way are library files--archive 
 files whose members are object files. The linker handles an 
-archive file by scan- ning through it for members which 
+archive file by scanning through it for members which 
 define symbols that have so far been referenced but not 
 defined. However, if the linker finds an ordinary object 
 file rather than a library, the object file is linked in the 
 usual fashion. The only difference between using an 
@@ -1052,9 +1051,9 @@
  
 __-static__ 
  
  
-On systems that support dynamic linking, this pre- vents  
+On systems that support dynamic linking, this prevents  
 linking with the shared libraries. On other systems, this 
 option has no effect. 
  
  
@@ -1069,9 +1068,9 @@
 __-symbolic__ 
  
  
 Bind references to global symbols when building a shared 
-object. Warn about any unresolved refer- ences (unless 
+object. Warn about any unresolved references (unless 
 overridden by the link editor option `__-Xlinker -z 
 -Xlinker defs__'). Only a few systems support this 
 option. 
  
@@ -1083,12 +1082,12 @@
 this to supply system-specific linker options which GNU CC 
 does not know how to recognize. 
  
  
-If you want to pass an option that takes an argu- ment , you 
+If you want to pass an option that takes an argument , you 
 must use `__-Xlinker__' twice, once for the option and 
 once for the argument. For example, to pass `__-assert 
-definitions__', you must write `__-Xlink- er -assert 
+definitions__', you must write `__-Xlinker -assert 
 -Xlinker definitions__'. It does not work to write 
 `__-Xlinker __', because 
 this passes the entire string as a single argument, which is 
 not what the linker expects. 
@@ -1142,9 +1141,9 @@
  
  
 In addition, the `__-I-__' option inhibits the use of the 
 current directory (where the current input file came from) 
-as the first search directory for `__#in- clude  
+as the first search directory for `__#include  
 __''file''____'. There is no way to 
 override this effect of `__-I-__'. With `__-I.__' you 
 can specify searching the directory which was current when 
 the compiler was invoked. That is not exactly the same as 
@@ -1166,9 +1165,9 @@
  
 __-B__''prefix'' 
  
  
-This option specifies where to find the executa- bles
+This option specifies where to find the executables
 libraries and data files of the compiler it- 
 self. 
  
  
@@ -1185,9 +1184,9 @@
 found, or if `__-B__' was not specified, the driver tries 
 two standard prefixes, which are `__/usr/lib/gcc/__' and 
 `__/usr/local/lib/gcc-lib/__'. If neither of those 
 results in a file name that is found, the compiler driver 
-searches for the unmodi- fied program name, using the 
+searches for the unmodified program name, using the 
 directories specified in your `__PATH__' environment 
 variable. 
  
  
@@ -1244,22 +1243,22 @@
 reject all programs that use forbidden 
 extensions. 
  
  
-Valid ANSI standard C programs should compile prop- erly  
+Valid ANSI standard C programs should compile properly  
 with or without this option (though a rare few will require 
-`__-ansi__'). However, without this op- tion , certain GNU  
-extensions and traditional C fea- tures are supported as 
+`__-ansi__'). However, without this option , certain GNU  
+extensions and traditional C features are supported as 
 well. With this option, they are rejected. There is no 
 reason to ''use'' this option; it exists only to satisfy 
 pedants. 
  
  
 `__-pedantic__' does not cause warning messages for use 
 of the alternate keywords whose names begin and end with 
 `______'. Pedantic warnings are also disabled in the 
-expression that follows ____extension____. Howev- er ,  
-only system header files should use these es- cape routes; 
+expression that follows ____extension____. However ,  
+only system header files should use these escape routes; 
 application programs should avoid them. 
  
  
 __-pedantic-errors__ 
@@ -1287,11 +1286,11 @@
 because __longjmp__ cannot in fact be called at the place 
 which would cause a problem. 
  
  
-A function can return either with or without a val- ue
+A function can return either with or without a value
 (Falling off the end of the function body is considered 
-returning without a value.) For exam- ple , this function 
+returning without a value.) For example , this function 
 would evoke such a warning: 
  
  
 foo (a) 
@@ -1300,9 +1299,9 @@
 Spurious warnings can occur because GNU CC does not realize that certain functions (including __abort__ and __longjmp__) will never return. 
  
  
 An expression-statement or the left-hand side of a comma 
-expression contains no side effects. To sup- press the 
+expression contains no side effects. To suppress the 
 warning, cast the unused expression to void. For example, an 
 expression such as `__x[[i,j]__' will cause a warning, but 
 `__x[[(void)i,j]__' will not. 
  
@@ -1335,17 +1334,17 @@
 __-Wmain__ 
  
  
 Warn if the __main__ function is declared or defined with 
-a suspicious type. Typically, it is a func- tion with  
-external linkage, returning __int__, and tak- ing zero or 
+a suspicious type. Typically, it is a function with  
+external linkage, returning __int__, and taking zero or 
 two arguments. 
  
  
 __-Wreturn-type__ 
  
  
-Warn whenever a function is defined with a re- turn -type 
+Warn whenever a function is defined with a return -type 
 that defaults to __int__. Also warn about any 
 __return__ statement with no return-value in a function 
 whose return-type is not __void__. 
  
@@ -1395,9 +1394,9 @@
 __-Wchar-subscripts__ 
  
  
 Warn if an array subscript has type __char__. This is a 
-common cause of error, as programmers often for- get that 
+common cause of error, as programmers often forget that 
 this type is signed on some machines. 
  
  
 __-Wuninitialized__ 
@@ -1407,9 +1406,9 @@
 initialized. 
  
  
 These warnings are possible only in optimizing com- 
-pilation, because they require data flow informa- tion that 
+pilation, because they require data flow information that 
 is computed only when optimizing. If you don't specify 
 `__-O__', you simply won't get these 
 warnings. 
  
@@ -1423,9 +1422,9 @@
  
  
 Note that there may be no warning about a variable that is 
 used only to compute a value that itself is never used, 
-because such computations may be delet- ed by data flow 
+because such computations may be deleted by data flow 
 analysis before the warnings are printed. 
  
  
 These warnings are made optional because GNU CC is not smart 
@@ -1457,9 +1456,9 @@
 
 This has no bug because __save_y__ is used only if it is set. 
  
  
-Some spurious warnings can be avoided if you de- clare as 
+Some spurious warnings can be avoided if you declare as 
 __volatile__ all the functions you use that never 
 return. 
  
  
@@ -1472,17 +1471,17 @@
  
 __-Wtemplate-debugging__ 
  
  
-When using templates in a C++ program, warn if de- bugging  
+When using templates in a C++ program, warn if debugging  
 is not yet fully available (C++ only). 
  
  
-__-Wall__ 
+__[ -Wall] __ 
  
  
 All of the above `__-W__' options combined. These are all 
-the options which pertain to usage that we rec- ommend  
+the options which pertain to usage that we recommend  
 avoiding and that we believe is easy to avoid, even in 
 conjunction with macros. 
  
  
@@ -1494,15 +1493,15 @@
  
 __-Wtraditional__ 
  
  
-Warn about certain constructs that behave differ- ently in 
+Warn about certain constructs that behave differently in 
 traditional and ANSI C. 
  
  
 Macro arguments occurring within string constants in the 
-macro body. These would substitute the ar- gument in  
-traditional C, but are part of the con- stant in ANSI 
+macro body. These would substitute the argument in  
+traditional C, but are part of the constant in ANSI 
 C. 
  
  
 A function declared external in one block and then used 
@@ -1515,9 +1514,9 @@
  
 __-Wshadow__ 
  
  
-Warn whenever a local variable shadows another lo- cal  
+Warn whenever a local variable shadows another local  
 variable. 
  
  
 __-Wid-clash-__''len'' 
@@ -1549,12 +1548,12 @@
  
 __-Wcast-align__ 
  
  
-Warn whenever a pointer is cast such that the re- quired  
+Warn whenever a pointer is cast such that the required  
 alignment of the target is increased. For example, warn if a 
-__char__ is cast to an __int__ on ma- chines where  
-integers can only be accessed at two- or four-byte 
+__char__ is cast to an __int__ on machines where  
+integers can only be accessed at twoor four-byte 
 boundaries. 
  
  
 __-Wwrite-strings__ 
@@ -1574,10 +1573,10 @@
 __-Wconversion__ 
  
  
 Warn if a prototype causes a type conversion that is 
-different from what would happen to the same ar- gument in  
-the absence of a prototype. This in- cludes conversions of 
+different from what would happen to the same argument in  
+the absence of a prototype. This includes conversions of 
 fixed point to floating and vice versa, and conversions 
 changing the width or signedness of a fixed point argument 
 except when the same as the default promotion. 
  
@@ -1586,34 +1585,34 @@
  
  
 Warn if any functions that return structures or unions are 
 defined or called. (In languages where you can return an 
-array, this also elicits a warn- ing .) 
+array, this also elicits a warning .) 
  
  
 __-Wstrict-prototypes__ 
  
  
 Warn if a function is declared or defined without specifying 
-the argument types. (An old-style func- tion definition is 
+the argument types. (An old-style function definition is 
 permitted without a warning if preceded by a declaration 
-which specifies the argu- ment types.) 
+which specifies the argument types.) 
  
  
 __-Wmissing-prototypes__ 
  
  
-Warn if a global function is defined without a pre- vious  
-prototype declaration. This warning is is- sued even if the  
-definition itself provides a pro- totype . The aim is to 
+Warn if a global function is defined without a previous  
+prototype declaration. This warning is issued even if the  
+definition itself provides a prototype . The aim is to 
 detect global functions that fail to be declared in header 
 files. 
  
  
 __-Wmissing-declarations__ 
  
  
-Warn if a global function is defined without a pre- vious  
+Warn if a global function is defined without a previous  
 declaration. Do so even if the definition itself provides a 
 prototype. Use this option to detect global functions that 
 are not declared in header files. 
  
@@ -1621,9 +1620,9 @@
 __-Wredundant-decls__ 
  
  
 Warn if anything is declared more than once in the same 
-scope, even in cases where multiple declara- tion is valid 
+scope, even in cases where multiple declaration is valid 
 and changes nothing. 
  
  
 __-Wnested-externs__ 
@@ -1659,17 +1658,17 @@
 request warnings when a derived class declares a function 
 that may be an erroneous attempt to define a virtual 
 function: that is, warn when a function with the same name 
 as a virtual function in the base class, but with a type 
-signa- ture that doesn't match any virtual functions from 
+signature that doesn't match any virtual functions from 
 the base class. 
  
  
 __-Winline__ 
  
  
 Warn if a function can not be inlined, and either it was 
-declared as inline, or else the __-fin- line -functions__ 
+declared as inline, or else the __-finline -functions__ 
 option was given. 
  
  
 __-Werror__ 
@@ -1686,17 +1685,17 @@
  
 __-g__ 
  
  
-Produce debugging information in the operating sys- tem 's 
+Produce debugging information in the operating system 's 
 native format (stabs, COFF, XCOFF, or DWARF). GDB can work 
 with this debugging information. 
  
  
 On most systems that use stabs format, `__-g__' enables 
 use of extra debugging information that only GDB can use; 
 this extra information makes debugging work better in GDB 
-but will probably make other de- buggers crash or refuse to 
+but will probably make other debuggers crash or refuse to 
 read the program. If you want to control for certain whether 
 to generate the extra information, use `__-gstabs+__', 
 `__-gstabs__', `__-gxcoff+__', `__-gxcoff__', 
 `__-gdwarf+__', or `__-gdwarf__' (see 
@@ -1704,19 +1703,19 @@
  
  
 Unlike most other C compilers, GNU CC allows you to use 
 `__-g__' with `__-O__'. The shortcuts taken by opti- 
-mized code may occasionally produce surprising re- sults
+mized code may occasionally produce surprising results
 some variables you declared may not exist at all; flow of 
 control may briefly move where you did not expect it; some 
 statements may not be executed because they compute constant 
-results or their val- ues were already at hand; some  
-statements may exe- cute in different places because they 
+results or their values were already at hand; some  
+statements may execute in different places because they 
 were moved out of loops. 
  
  
 Nevertheless it proves possible to debug optimized output. 
-This makes it reasonable to use the opti- mizer for programs 
+This makes it reasonable to use the optimizer for programs 
 that might have bugs. 
  
  
 The following options are useful when GNU CC is generated 
@@ -1811,14 +1810,14 @@
  
 Level 1 produces minimal information, enough for making 
 backtraces in parts of the program that you don't plan to 
 debug. This includes descriptions of functions and external 
-variables, but no informa- tion about local variables and no 
+variables, but no information about local variables and no 
 line numbers. 
  
  
 Level 3 includes extra information, such as all the macro 
-definitions present in the program. Some de- buggers support 
+definitions present in the program. Some debuggers support 
 macro expansion when you use `__-g3__'. 
  
  
 __-p__ 
@@ -1842,9 +1841,9 @@
 blocks, which will record the number of times each basic 
 block is executed. This data could be analyzed by a program 
 like __tcov__. Note, however, that the format of the data 
 is not what __tcov__ expects. Eventually GNU __gprof__ 
-should be ex- tended to process this data. 
+should be extended to process this data. 
  
  
 __-ax__ 
  
@@ -1871,18 +1870,18 @@
 __-d__''letters'' 
  
  
 Says to make debugging dumps during compilation at times 
-specified by ''letters''. This is used for de- bugging  
+specified by ''letters''. This is used for debugging  
 the compiler. The file names for most of the dumps are made 
 by appending a word to the source file name (e.g. 
 `__foo.c.rtl__' or `__foo.c.jump__'). 
  
  
 __-dM__ 
  
  
-Dump all macro definitions, at the end of prepro- cessing
+Dump all macro definitions, at the end of preprocessing
 and write no output. 
  
  
 __-dN__ 
@@ -1894,16 +1893,16 @@
  
 __-dD__ 
  
  
-Dump all macro definitions, at the end of prepro- cessing
+Dump all macro definitions, at the end of preprocessing
 in addition to normal output. 
  
  
 __-dy__ 
  
  
-Dump debugging information during parsing, to stan- dard  
+Dump debugging information during parsing, to standard  
 error. 
  
  
 __-dr__ 
@@ -1915,9 +1914,9 @@
  
 __-dx__ 
  
  
-Just generate RTL for a function instead of compil- ing it. 
+Just generate RTL for a function instead of compiling it. 
 Usually used with `__r__'. 
  
  
 __-dj__ 
@@ -2028,9 +2027,9 @@
  
 __-dp__ 
  
  
-Annotate the assembler output with a comment indi- cating  
+Annotate the assembler output with a comment indicating  
 which pattern and alternative was used. 
  
  
 __-fpretend-float__ 
@@ -2040,17 +2039,17 @@
 machine uses the same floating point format as the host 
 machine. This causes incorrect output of the actual floating 
 constants, but the actual instruction sequence will probably 
 be the same as GNU CC would make when running on the target 
-ma- chine
+machine
  
  
 __-save-temps__ 
  
  
 Store the usual ``temporary'' intermediate files 
 permanently; place them in the current directory and name 
-them based on the source file. Thus, com- piling  
+them based on the source file. Thus, compiling  
 `__foo.c__' with `__-c -save-temps__' would produce 
 files `__foo.cpp__' and `__foo.s__', as well as 
 `__foo.o__'. 
  
@@ -2088,18 +2087,18 @@
 __-O1__ 
  
  
 Optimize. Optimizing compilation takes somewhat more time, 
-and a lot more memory for a large func- tion
+and a lot more memory for a large function
  
  
 Without `__-O__', the compiler's goal is to reduce the 
 cost of compilation and to make debugging produce the 
 expected results. Statements are independent: if you stop 
 the program with a breakpoint between statements, you can 
 then assign a new value to any variable or change the 
 program counter to any other statement in the function and 
-get exactly the re- sults you would expect from the source 
+get exactly the results you would expect from the source 
 code. 
  
  
 Without `__-O__', only variables declared __register__ 
@@ -2114,20 +2113,20 @@
  
 When you specify `__-O__', the two options 
 `__-fthread-jumps__' and `__-fdefer-pop__' are turned 
 on. On machines that have delay slots, the `__-fde- 
-layed-branch__' option is turned on. For those ma- chines  
+layed-branch__' option is turned on. For those machines  
 that can support debugging even without a frame pointer, the 
 `__-fomit-frame-pointer__' option is turned on. On some 
 machines other flags may also be turned on. 
  
  
 __-O2__ 
  
  
-Optimize even more. Nearly all supported optimiza- tions  
+Optimize even more. Nearly all supported optimizations  
 that do not involve a space-speed tradeoff are performed. 
-Loop unrolling and function inlin- ing are not done, for 
+Loop unrolling and function inlining are not done, for 
 example. As compared to __-O__, this option increases 
 both compilation time and the performance of the generated 
 code. 
  
@@ -2163,16 +2162,16 @@
 __-ffloat-store__ 
  
  
 Do not store floating point variables in registers. This 
-prevents undesirable excess precision on ma- chines such as  
-the 68000 where the floating regis- ters (of the 68881) keep  
-more precision than a __dou- ble __ is supposed to 
+prevents undesirable excess precision on machines such as  
+the 68000 where the floating registers (of the 68881) keep  
+more precision than a __double __ is supposed to 
 have. 
  
  
 For most programs, the excess precision does only good, but 
-a few programs rely on the precise defi- nition of IEEE 
+a few programs rely on the precise definition of IEEE 
 floating point. Use `__-ffloat-store__' for such 
 programs. 
  
  
@@ -2205,34 +2204,34 @@
  
  
 makes six passes through all three steps. By using a 
 software cache, a ``hit'' significantly reduces this cost. 
-Unfortunately, using the cache intro- duces another layer of  
-mechanisms which must be im- plemented , and so incurs its  
-own overhead. `__-fmem- oize -lookups__' enables the 
+Unfortunately, using the cache introduces another layer of  
+mechanisms which must be implemented , and so incurs its  
+own overhead. `__-fmemoize -lookups__' enables the 
 software cache. 
  
  
 Because access privileges (visibility) to members and member 
 functions may differ from one function context to the next, 
 __g++__ may need to flush the cache. With the 
 `__-fmemoize-lookups__' flag, the cache is flushed after 
-every function that is com- piled . The `-fsave-memoized' 
+every function that is compiled . The `-fsave-memoized' 
 flag enables the same software cache, but when the compiler 
 determines that the context of the last function compiled 
 would yield the same access privileges of the next function 
 to compile, it preserves the cache. This is most helpful 
 when defining many member functions for the same class: with 
 the exception of member functions which are friends of other 
 classes, each member function has exactly the same access 
-privi- leges as every other, and the cache need not be 
+privileges as every other, and the cache need not be 
 flushed. 
  
  
 __-fno-default-inline__ 
  
  
-Don't make member functions inline by default mere- ly  
+Don't make member functions inline by default merely  
 because they are defined inside the class scope (C++ 
 only). 
  
  
@@ -2240,9 +2239,9 @@
  
  
 Always pop the arguments to each function call as soon as 
 that function returns. For machines which must pop arguments 
-after a function call, the com- piler normally lets 
+after a function call, the compiler normally lets 
 arguments accumulate on the stack for several function calls 
 and pops them all at once. 
  
  
@@ -2250,12 +2249,12 @@
  
  
 Force memory operands to be copied into registers before 
 doing arithmetic on them. This may produce better code by 
-making all memory references poten- tial common  
-subexpressions. When they are not com- mon subexpressions, 
+making all memory references potential common  
+subexpressions. When they are not common subexpressions, 
 instruction combination should eliminate the separate 
-register-load. I am inter- ested in hearing about the 
+register-load. I am interested in hearing about the 
 difference this makes. 
  
  
 __-fforce-addr__ 
@@ -2270,19 +2269,19 @@
 __-fomit-frame-pointer__ 
  
  
 Don't keep the frame pointer in a register for functions 
-that don't need one. This avoids the in- structions to save,  
-set up and restore frame point- ers ; it also makes an extra 
+that don't need one. This avoids the instructions to save,  
+set up and restore frame pointers ; it also makes an extra 
 register available in many functions. ''It also makes 
 debugging impossible on most machines''. 
  
  
 On some machines, such as the Vax, this flag has no effect, 
-because the standard calling sequence auto- matically  
+because the standard calling sequence automatically  
 handles the frame pointer and nothing is saved by pretending 
-it doesn't exist. The ma- chine -description macro  
-__FRAME_POINTER_REQUIRED__ con- trols whether a target 
+it doesn't exist. The machine -description macro  
+__FRAME_POINTER_REQUIRED__ controls whether a target 
 machine supports this flag. 
  
  
 __-finline-functions__ 
@@ -2293,9 +2292,9 @@
 enough to be worth integrating in this way. 
  
  
 If all calls to a given function are integrated, and the 
-function is declared __static__, then GCC nor- mally does 
+function is declared __static__, then GCC normally does 
 not output the function as assembler code in its own 
 right. 
  
  
@@ -2303,23 +2302,23 @@
  
  
 Enable values to be allocated in registers that will be 
 clobbered by function calls, by emitting extra instructions 
-to save and restore the regis- ters around such calls. Such 
+to save and restore the registers around such calls. Such 
 allocation is done only when it seems to result in better 
 code than would otherwise be produced. 
  
  
-This option is enabled by default on certain ma- chines
+This option is enabled by default on certain machines
 usually those which have no call-preserved registers to use 
 instead. 
  
  
 __-fkeep-inline-functions__ 
  
  
-Even if all calls to a given function are integrat- ed , and  
-the function is declared __static__, neverthe- less  
+Even if all calls to a given function are integrated , and  
+the function is declared __static__, nevertheless  
 output a separate run-time callable version of the 
 function. 
  
  
@@ -2348,17 +2347,17 @@
  
  
 This option allows GCC to violate some ANSI or IEEE 
 rules/specifications in the interest of optimizing code for 
-speed. For example, it allows the compil- er to assume 
+speed. For example, it allows the compiler to assume 
 arguments to the __sqrt__ function are non-negative 
 numbers. 
  
  
 This option should never be turned on by any `__-O__' 
 option since it can result in incorrect output for programs 
 which depend on an exact implementation of IEEE or ANSI 
-rules/specifications for math func- tions
+rules/specifications for math functions
  
  
 The following options control specific optimizations. The 
 `__-O2__' option turns on all of these optimizations 
@@ -2379,19 +2378,19 @@
  
 __-fstrength-reduce__ 
  
  
-Perform the optimizations of loop strength reduc- tion and 
+Perform the optimizations of loop strength reduction and 
 elimination of iteration variables. 
  
  
 __-fthread-jumps__ 
  
  
 Perform optimizations where we check to see if a jump 
-branches to a location where another compari- son subsumed 
+branches to a location where another comparison subsumed 
 by the first is found. If so, the first branch is redirected 
-to either the destina- tion of the second branch or a point 
+to either the destination of the second branch or a point 
 immediately following it, depending on whether the condition 
 is known to be true or false. 
  
  
@@ -2442,9 +2441,9 @@
 __-felide-constructors__ 
  
  
 Elide constructors when this seems plausible (C++ only). 
-With this flag, GNU C++ initializes __y__ di- rectly from 
+With this flag, GNU C++ initializes __y__ directly from 
 the call to __foo__ without going through a temporary in 
 the following code: 
  
  
@@ -2475,17 +2474,17 @@
  
 __-fdelayed-branch__ 
  
  
-If supported for the target machine, attempt to re- order  
+If supported for the target machine, attempt to reorder  
 instructions to exploit instruction slots available after 
 delayed branch instructions. 
  
  
 __-fschedule-insns__ 
  
  
-If supported for the target machine, attempt to re- order  
+If supported for the target machine, attempt to reorder  
 instructions to eliminate execution stalls due to required 
 data being unavailable. This helps machines that have slow 
 floating point or memory load instructions by allowing other 
 instructions to be issued until the result of the load or 
@@ -2495,9 +2494,9 @@
 __-fschedule-insns2__ 
  
  
 Similar to `__-fschedule-insns__', but requests an ad- 
-ditional pass of instruction scheduling after reg- ister  
+ditional pass of instruction scheduling after register  
 allocation has been done. This is especially useful on 
 machines with a relatively small number of registers and 
 where memory load instructions take more than one 
 cycle. 
@@ -2506,10 +2505,10 @@
  
 By default, GNU CC compiles code for the same type of ma- 
 chine that you are using. However, it can also be in- 
 stalled as a cross-compiler, to compile for some other type 
-of machine. In fact, several different configura- tions of  
-GNU CC, for different target machines, can be in- stalled  
+of machine. In fact, several different configurations of  
+GNU CC, for different target machines, can be installed  
 side by side. Then you specify which one to use with the 
 `__-b__' option. 
  
  
@@ -2522,9 +2521,9 @@
 __-b__ ''machine'' 
  
  
 The argument ''machine'' specifies the target machine for 
-compilation. This is useful when you have in- stalled GNU CC 
+compilation. This is useful when you have installed GNU CC 
 as a cross-compiler. 
  
  
 The value to use for ''machine'' is the same as was 
@@ -2550,17 +2549,17 @@
 2.0. 
  
  
 The default version, when you do not specify `__-V__', is 
-controlled by the way GNU CC is installed. Nor- mally , it 
+controlled by the way GNU CC is installed. Normally , it 
 will be a version that is recommended for general 
 use. 
 !!MACHINE DEPENDENT OPTIONS 
  
  
 Each of the target machine types can have its own special 
 options, starting with `__-m__', to choose among various 
-hard- ware models or configurations--for example, 68010 vs 
+hardware models or configurations--for example, 68010 vs 
 68020, floating coprocessor or none. A single installed 
 version of the compiler can compile for any model or con- 
 figuration, according to the options specified. 
  
@@ -2636,14 +2635,14 @@
  
 __-msoft-float__ 
  
  
-Generate output containing library calls for float- ing  
+Generate output containing library calls for floating  
 point. ''WARNING:'' the requisite libraries are not part 
 of GNU CC. Normally the facilities of the machine's usual C 
 compiler are used, but this can't be done directly in 
 cross-compilation. You must make your own arrangements to 
-provide suitable li- brary functions for 
+provide suitable library functions for 
 cross-compilation. 
  
  
 __-mshort__ 
@@ -2671,9 +2670,9 @@
 __-mrtd__ 
  
  
 Use a different function-calling convention, in which 
-functions that take a fixed number of argu- ments return 
+functions that take a fixed number of arguments return 
 with the __rtd__ instruction, which pops their arguments 
 while returning. This saves one instruction in the caller 
 since there is no need to pop the arguments 
 there. 
@@ -2713,16 +2712,16 @@
  
 __-mgnu__ 
  
  
-Do output those jump instructions, on the assump- tion that 
+Do output those jump instructions, on the assumption that 
 you will assemble with the GNU assembler. 
  
  
 __-mg__ 
  
  
-Output code for g-format floating point numbers in- stead of 
+Output code for g-format floating point numbers instead of 
 d-format. 
  
  
 These `__-m__' switches are supported on the 
@@ -2734,9 +2733,9 @@
  
 __-mhard-float__ 
  
  
-Generate output containing floating point instruc- tions
+Generate output containing floating point instructions
 This is the default. 
  
  
 __-mno-fpu__ 
@@ -2744,12 +2743,12 @@
  
 __-msoft-float__ 
  
  
-Generate output containing library calls for float- ing  
+Generate output containing library calls for floating  
 point. ''Warning:'' there is no GNU floating-point 
 library for SPARC. Normally the facilities of the machine's 
-usual C compiler are used, but this can- not be done 
+usual C compiler are used, but this cannot be done 
 directly in cross-compilation. You must make your own 
 arrangements to provide suitable library functions for 
 cross-compilation. 
  
@@ -2850,11 +2849,11 @@
 __-margcount__ 
  
  
 Generate code which puts an argument count in the word 
-preceding each argument list. Some non- portable Convex and 
+preceding each argument list. Some nonportable Convex and 
 Vax programs need this word. (Debuggers don't, except for 
-functions with vari- able -length argument lists; this info 
+functions with variable -length argument lists; this info 
 is in the symbol table.) 
  
  
 __-mnoargcount__ 
@@ -2871,9 +2870,9 @@
 __-mdw__ 
  
  
 Generate code that assumes the DW bit is set, i.e., that 
-byte and halfword operations are directly sup- ported by the 
+byte and halfword operations are directly supported by the 
 hardware. This is the default. 
  
  
 __-mnodw__ 
@@ -2933,10 +2932,10 @@
 __-mkernel-registers__ 
  
  
 Generate references to registers __gr64-gr95__ instead of 
-__gr96-gr127__. This option can be used when com- piling  
-kernel code that wants a set of global reg- isters disjoint 
+__gr96-gr127__. This option can be used when compiling  
+kernel code that wants a set of global registers disjoint 
 from that used by user-mode code. 
  
  
 Note that when this option is used, register names in 
@@ -2958,9 +2957,9 @@
 justment. This is often used for kernel code. 
  
  
 These `__-m__' options are defined for Motorola 88K 
-architec- tures
+architectures
  
  
 __-m88000__ 
  
@@ -3008,11 +3007,11 @@
  
 Early models of the 88K architecture had problems with 
 division by zero; in particular, many of them didn't trap. 
 Use these options to avoid including (or to include 
-explicitly) additional code to de- tect division by zero and 
+explicitly) additional code to detect division by zero and 
 signal an exception. All GCC configurations for the 88K use 
-`__-mcheck-ze- ro -division__' by default. 
+`__-mcheck-zero -division__' by default. 
  
  
 __-mocs-debug-info__ 
  
@@ -3020,9 +3019,9 @@
 __-mno-ocs-debug-info__ 
  
  
 Include (or omit) additional debugging information (about 
-registers used in each stack frame) as spec- ified in the 
+registers used in each stack frame) as specified in the 
 88Open Object Compatibility Standard, ``OCS''. This extra 
 information is not needed by GDB. The default for DG/UX, 
 SVr4, and Delta 88 SVr3.2 is to include this information; 
 other 88k configurations omit this information by 
@@ -3037,9 +3036,9 @@
  
 Force (or do not require) register values to be stored in a 
 particular place in stack frames, as specified in OCS. The 
 DG/UX, Delta88 SVr3.2, and BCS configurations use 
-`__-mocs-frame-position__'; oth- er 88k configurations 
+`__-mocs-frame-position__'; other 88k configurations 
 have the default 
 `__-mno-ocs-frame-position__'. 
  
  
@@ -3050,9 +3049,9 @@
  
  
 Control how to store function arguments in stack frames. 
 `__-moptimize-arg-area__' saves space, but may break some 
-debuggers (not GDB). `__-mno-opti- mize -arg-area__' 
+debuggers (not GDB). `__-mno-optimize -arg-area__' 
 conforms better to standards. By default GCC does not 
 optimize the argument area. 
  
  
@@ -3063,9 +3062,9 @@
 relative to __r0__, which allows loading a value using a 
 single instruction (rather than the usual two). You control 
 which data references are affected by specifying ''num'' 
 with this option. For example, if you specify 
-`__-mshort-data-512__', then the data ref- erences  
+`__-mshort-data-512__', then the data references  
 affected are those involving displacements of less than 512 
 bytes. `__-mshort-data-__''num''' is not effective for 
 ''num'' greater than 64K. 
  
@@ -3075,31 +3074,31 @@
  
 __-mno-serialize-volatile__ 
  
  
-Do, or do not, generate code to guarantee sequen- tial  
+Do, or do not, generate code to guarantee sequential  
 consistency of volatile memory references. 
  
  
 GNU CC always guarantees consistency by default, for the 
 preferred processor submodel. How this is done depends on 
 the submodel. 
  
  
-The m88100 processor does not reorder memory refer- ences  
-and so always provides sequential consisten- cy . If you use 
+The m88100 processor does not reorder memory references  
+and so always provides sequential consistency . If you use 
 `__-m88100__', GNU CC does not generate any special 
-instructions for sequential consisten- cy
+instructions for sequential consistency
  
  
 The order of memory references made by the m88110 processor 
 does not always match the order of the instructions 
-requesting those references. In par- ticular , a load 
+requesting those references. In particular , a load 
 instruction may execute before a preceding store 
-instruction. Such reordering vio- lates sequential 
+instruction. Such reordering violates sequential 
 consistency of volatile memory references, when there are 
 multiple processors. When you use `__-m88000__' or 
-`__-m88110__', GNU CC gener- ates special instructions 
+`__-m88110__', GNU CC generates special instructions 
 when appropriate, to force execution in the proper 
 order. 
  
  
@@ -3109,9 +3108,9 @@
 `__-mno-serialize-volatile__'. 
  
  
 If you use the `__-m88100__' option but require sequen- 
-tial consistency when running on the m88110 proces- sor , you 
+tial consistency when running on the m88110 processor , you 
 should use `__-mserialize-volatile__'. 
  
  
 __-msvr4__ 
@@ -3120,9 +3119,9 @@
 __-msvr3__ 
  
  
 Turn on (`__-msvr4__') or off (`__-msvr3__') compiler 
-ex- tensions related to System V release 4 (SVr4). This 
+extensions related to System V release 4 (SVr4). This 
 controls the following: 
  
  
 Which variant of the assembler syntax to emit (which you can 
@@ -3130,9 +3129,9 @@
 sion-03.00__'). 
  
  
 `__-msvr4__' makes the C preprocessor recognize 
-`__#prag- ma weak__' 
+`__#pragma weak__' 
  
  
 `__-msvr4__' makes GCC issue additional declaration di- 
 rectives used in SVr4. 
@@ -3157,9 +3156,9 @@
 __-muse-div-instruction__ 
  
  
 Very early models of the 88K architecture didn't have a 
-divide instruction, so GCC avoids that in- struction by 
+divide instruction, so GCC avoids that instruction by 
 default. Use this option to specify that it's safe to use 
 the divide instruction. 
  
  
@@ -3205,9 +3204,9 @@
  
 __-min-line-mul__ 
  
  
-Use an in-line code sequence for integer multi- plies . This 
+Use an in-line code sequence for integer multiplies . This 
 is the default. 
  
  
 __-mcall-lib-mul__ 
@@ -3218,10 +3217,10 @@
  
 __-mfull-fp-blocks__ 
  
  
-Generate full-size floating point data blocks, in- cluding  
-the minimum amount of scratch space recom- mended by IBM. 
+Generate full-size floating point data blocks, including  
+the minimum amount of scratch space recommended by IBM. 
 This is the default. 
  
  
 __-mminimum-fp-blocks__ 
@@ -3236,12 +3235,12 @@
 __-mfp-arg-in-fpregs__ 
  
  
 Use a calling sequence incompatible with the IBM calling 
-convention in which floating point argu- ments are passed in 
+convention in which floating point arguments are passed in 
 floating point registers. Note that __varargs.h__ and 
 __stdargs.h__ will not work with floating point operands 
-if this option is speci- fied
+if this option is specified
  
  
 __-mfp-arg-in-gregs__ 
  
@@ -3253,9 +3252,9 @@
 __-mhc-struct-return__ 
  
  
 Return structures of more than one word in memory, rather 
-than in a register. This provides compati- bility with the 
+than in a register. This provides compatibility with the 
 !MetaWare HighC (hc) compiler. Use 
 `__-fpcc-struct-return__' for compatibility with the 
 Portable C Compiler (pcc). 
  
@@ -3270,9 +3269,9 @@
 `__-mhc-struct-return__'. 
  
  
 These `__-m__' options are defined for the MIPS family of 
-com- puters
+computers
  
  
 __-mcpu=__''cpu-type'' 
  
@@ -3327,9 +3326,9 @@
 __mips-tfile__ to add normal debug information. This is 
 the default for all platforms except for the OSF/1 reference 
 platform, using the OSF/rose object format. If any of the 
 __-ggdb__, __-gstabs__, or __-gstabs+__ switches 
-are used, the __mips-tfile__ program will en- capsulate  
+are used, the __mips-tfile__ program will encapsulate  
 the stabs within MIPS ECOFF. 
  
  
 __-mgas__ 
@@ -3388,9 +3387,9 @@
  
  
 The __-mmemcpy__ switch makes all block moves call the 
 appropriate string function (__memcpy__ or __bcopy__) 
-in- stead of possibly generating inline code. 
+instead of possibly generating inline code. 
  
  
 __-mmips-tfile__ 
  
@@ -3411,39 +3410,39 @@
  
 __-msoft-float__ 
  
  
-Generate output containing library calls for float- ing  
+Generate output containing library calls for floating  
 point. ''WARNING:'' the requisite libraries are not part 
 of GNU CC. Normally the facilities of the machine's usual C 
 compiler are used, but this can't be done directly in 
 cross-compilation. You must make your own arrangements to 
-provide suitable li- brary functions for 
+provide suitable library functions for 
 cross-compilation. 
  
  
 __-mhard-float__ 
  
  
-Generate output containing floating point instruc- tions .  
-This is the default if you use the unmodi- fied  
+Generate output containing floating point instructions .  
+This is the default if you use the unmodified  
 sources. 
  
  
 __-mfp64__ 
  
  
 Assume that the __FR__ bit in the status word is on, and 
-that there are 32 64-bit floating point regis- ters , instead  
-of 32 32-bit floating point regis- ters . You must also 
+that there are 32 64-bit floating point registers , instead  
+of 32 32-bit floating point registers . You must also 
 specify the __-mcpu=r4000__ and __-mips3__ 
 switches. 
  
  
 __-mfp32__ 
  
  
-Assume that there are 32 32-bit floating point reg- isters
+Assume that there are 32 32-bit floating point registers
 This is the default. 
  
  
 __-mabicalls__ 
@@ -3464,14 +3463,14 @@
  
  
 The __-mhalf-pic__ switch says to put pointers to ex- 
 tern references into the data section and load them up, 
-rather than put the references in the text sec- tion . This 
+rather than put the references in the text section . This 
 option does not work at present. __-G__''num'' Put 
 global and static items less than or equal to ''num'' 
-bytes into the small data or bss sections in- stead of the  
-normal data or bss section. This al- lows the assembler to  
-emit one word memory refer- ence instructions based on the 
+bytes into the small data or bss sections instead of the  
+normal data or bss section. This allows the assembler to  
+emit one word memory reference instructions based on the 
 global pointer (__gp__ or __$28__), instead of the 
 normal two words used. By default, ''num'' is 8 when the 
 MIPS assembler is used, and 0 when the GNU assembler is 
 used. The __-G__''num'' switch is also passed to the 
@@ -3501,14 +3500,14 @@
  
 __-msoft-float__ 
  
  
-Generate output containing library calls for float- ing  
+Generate output containing library calls for floating  
 point. ''Warning:'' the requisite libraries are not part 
 of GNU CC. Normally the facilities of the machine's usual C 
 compiler are used, but this can't be done directly in 
 cross-compilation. You must make your own arrangements to 
-provide suitable li- brary functions for 
+provide suitable library functions for 
 cross-compilation. 
  
  
 On machines where a function returns floating point results 
@@ -3524,9 +3523,9 @@
 functions. 
  
  
 The usual calling convention has functions return values of 
-types __float__ and __double__ in an FPU regis- ter
+types __float__ and __double__ in an FPU register
 even if there is no FPU. The idea is that the operating 
 system should emulate an FPU. 
  
  
@@ -3534,9 +3533,9 @@
 be returned in ordinary CPU registers instead. 
  
  
 These `__-m__' options are defined for the HPPA family of 
-com- puters
+computers
  
  
 __-mpa-risc-1-0__ 
  
@@ -3563,10 +3562,10 @@
 __-mshared-libs__ 
  
  
 Generate code that can be linked against HP-UX shared 
-libraries. This option is not fully func- tion yet, and is  
-not on by default for any PA tar- get . Using this option can 
+libraries. This option is not fully function yet, and is  
+not on by default for any PA target . Using this option can 
 cause incorrect code to be generated by the 
 compiler. 
  
  
@@ -3591,10 +3590,10 @@
 __-mdisable-fpregs__ 
  
  
 Prevent floating point registers from being used in any 
-manner. This is necessary for compiling ker- nels which  
-perform lazy context switching of float- ing point 
+manner. This is necessary for compiling kernels which  
+perform lazy context switching of floating point 
 registers. If you use this option and attempt to perform 
 floating point operations, the compiler will 
 abort. 
  
@@ -3649,11 +3648,11 @@
 Do (or do not) attempt to alter leaf procedures to be 
 callable with the ''bal'' instruction as well as 
 ''call''. This will result in more efficient code for 
 explicit calls when the ''bal'' instruction can be sub- 
-stituted by the assembler or linker, but less effi- cient  
-code in other cases, such as calls via func- tion pointers,  
-or using a linker that doesn't sup- port this 
+stituted by the assembler or linker, but less efficient  
+code in other cases, such as calls via function pointers,  
+or using a linker that doesn't support this 
 optimization. 
  
  
 __-mtail-call__ 
@@ -3679,9 +3678,9 @@
 Assume (or do not assume) that the use of a complex 
 addressing mode is a win on this implementation of the i960. 
 Complex addressing modes may not be worthwhile on the 
 K-series, but they definitely are on the C-series. The 
-default is currently __-mcom- plex -addr__ for all 
+default is currently __-mcomplex -addr__ for all 
 processors except the CB and CC. 
  
  
 __-mcode-align__ 
@@ -3727,38 +3726,38 @@
  
 __-mold-align__ 
  
  
-Enable structure-alignment compatibility with In- tel 's gcc 
+Enable structure-alignment compatibility with Intel 's gcc 
 release version 1.3 (based on gcc 1.37). Currently this is 
 buggy in that __#pragma align 1__ is always assumed as 
 well, and cannot be turned off. 
  
  
 These `__-m__' options are defined for the DEC Alpha 
-implemen- tations
+implementations
  
  
 __-mno-soft-float__ 
  
  
 __-msoft-float__ 
  
  
-Use (do not use) the hardware floating-point in- structions  
+Use (do not use) the hardware floating-point instructions  
 for floating-point operations. When __-msoft-float__ is 
 specified, functions in `__libgcc1.c__' will be used to 
 perform floating-point operations. Unless they are replaced 
 by routines that emulate the floating-point operations, or 
 compiled in such a way as to call such emulations routines, 
 these routines will issue floating-point operations. If you 
-are compiling for an Alpha without float- ing -point  
-operations, you must ensure that the li- brary is built so 
+are compiling for an Alpha without floating -point  
+operations, you must ensure that the library is built so 
 as not to call them. 
  
  
-Note that Alpha implementations without float- ing -point  
-operations are required to have float- ing -point 
+Note that Alpha implementations without floating -point  
+operations are required to have floating -point 
 registers. 
  
  
 __-mfp-reg__ 
@@ -3766,22 +3765,22 @@
  
 __-mno-fp-regs__ 
  
  
-Generate code that uses (does not use) the float- ing -point 
+Generate code that uses (does not use) the floating -point 
 register set. __-mno-fp-regs__ implies 
 __-msoft-float__. If the floating-point register set is 
 not used, floating point operands are passed in integer 
 registers as if they were integers and floating-point 
 results are passed in $0 instead of $f0. This is a 
 non-standard calling sequence, so any function with a 
-floating-point argument or re- turn value called by code 
+floating-point argument or return value called by code 
 compiled with __-mno-fp-regs__ must also be compiled with 
-that op- tion
+that option
  
  
 A typical use of this option is building a kernel that does 
-not use, and hence need not save and re- store , any 
+not use, and hence need not save and restore , any 
 floating-point registers. 
  
  
 These additional options are available on System V Release 4 
@@ -3802,9 +3801,9 @@
  
 __-Qy__ 
  
  
-Identify the versions of each tool used by the com- piler
+Identify the versions of each tool used by the compiler
 in a __.ident__ assembler directive in the out- 
 put. 
  
  
@@ -3851,19 +3850,19 @@
 (C++ only). 
  
  
 Normally, GNU C++ makes conservative assumptions about 
-objects reached through references. For ex- ample , the 
+objects reached through references. For example , the 
 compiler must check that __a__ is not null in code like 
 the following: 
  
  
 obj 
  
  
 Checking that references of this sort have non-null values 
-requires extra code, however, and it is un- necessary for  
-many programs. You can use `__-fnon- null -objects__' to 
+requires extra code, however, and it is unnecessary for  
+many programs. You can use `__-fnonnull -objects__' to 
 omit the checks for null, if your program doesn't require 
 checking. 
  
  
@@ -3873,9 +3872,9 @@
 Use the same convention for returning __struct__ and 
 __union__ values that is used by the usual C compiler on 
 your system. This convention is less efficient for small 
 structures, and on many machines it fails to be reentrant; 
-but it has the advantage of allow- ing intercallability 
+but it has the advantage of allowing intercallability 
 between GCC-compiled code and PCC-compiled 
 code. 
  
  
@@ -3916,20 +3915,20 @@
  
 Requests that the data and non-__const__ variables of 
 this compilation be shared data rather than private data. 
 The distinction makes sense only on certain operating 
-systems, where shared data is shared be- tween processes 
+systems, where shared data is shared between processes 
 running the same program, while private data exists in one 
 copy per process. 
  
  
 __-fno-common__ 
  
  
 Allocate even uninitialized global variables in the bss 
-section of the object file, rather than gener- ating them as 
+section of the object file, rather than generating them as 
 common blocks. This has the effect that if the same variable 
-is declared (without __ex- tern __) in two different 
+is declared (without __extern __) in two different 
 compilations, you will get an error when you link them. The 
 only reason this might be useful is if you wish to verify 
 that the program will work on other systems which always 
 work this way. 
@@ -3951,9 +3950,9 @@
 use a non-GNU linker, which also requires using the 
 __collect2__ program to make sure the system linker 
 includes constructors and destructors. (__collect2__ is 
 included in the GNU CC distribution.) For systems which 
-''must'' use __col- lect2 __, the compiler driver 
+''must'' use __collect2 __, the compiler driver 
 __gcc__ is configured to do this 
 automatically. 
  
  
@@ -3962,18 +3961,18 @@
  
 Don't output a __.size__ assembler directive, or any- 
 thing else that would cause trouble if the function is split 
 in the middle, and the two halves are placed at locations 
-far apart in memory. This op- tion is used when compiling 
+far apart in memory. This option is used when compiling 
 `__crtstuff.c__'; you should not need to use it for 
 anything else. 
  
  
 __-fverbose-asm__ 
  
  
 Put extra commentary information in the generated assembly 
-code to make it more readable. This op- tion is generally 
+code to make it more readable. This option is generally 
 only of use to those who actually need to read the generated 
 assembly code (perhaps while debugging the compiler 
 itself). 
  
@@ -4003,9 +4002,9 @@
 __-fPIC__ 
  
  
 If supported for the target machine, emit posi- 
-tion-independent code, suitable for dynamic link- ing , even 
+tion-independent code, suitable for dynamic linking , even 
 if branches need large displacements. 
  
  
 __-ffixed-__''reg'' 
@@ -4018,9 +4017,9 @@
  
  
 ''reg'' must be the name of a register. The register 
 names accepted are machine-specific and are defined in the 
-__REGISTER_NAMES__ macro in the machine descrip- tion  
+__REGISTER_NAMES__ macro in the machine description  
 macro file. 
  
  
 This flag does not have a negative form, because it 
@@ -4084,17 +4083,17 @@
  
  
 (C++ only.) Use this directive in header files that define 
 object classes, to save space in most of the object files 
-that use those classes. Nor- mally , local copies of certain 
+that use those classes. Normally , local copies of certain 
 information (backup copies of inline member functions, 
-debugging infor- mation , and the internal tables that  
-implement vir- tual functions) must be kept in each object 
+debugging information , and the internal tables that  
+implement virtual functions) must be kept in each object 
 file that includes class definitions. You can use this 
 pragma to avoid such duplication. When a header file 
 containing `__#pragma interface__' is included in a 
 compilation, this auxiliary information will not be 
-generated (unless the main input source file it- self uses 
+generated (unless the main input source file itself uses 
 `__#pragma implementation__'). Instead, the object files 
 will contain references to be resolved at link 
 time. 
  
@@ -4109,24 +4108,24 @@
 (C++ only.) Use this pragma in a main input file, when you 
 want full output from included header files to be generated 
 (and made globally visible). The included header file, in 
 turn, should use `__#pragma interface__'. Backup copies 
-of inline mem- ber functions, debugging information, and the  
-in- ternal tables used to implement virtual functions are 
+of inline member functions, debugging information, and the  
+internal tables used to implement virtual functions are 
 all generated in implementation files. 
  
  
 If you use `__#pragma implementation__' with no argu- 
 ment, it applies to an include file with the same basename 
-as your source file; for example, in `__all- class .cc__', 
+as your source file; for example, in `__allclass .cc__', 
 `__#pragma implementation__' by itself is equivalent to 
 `__#pragma implementation __'. 
 Use the string argument if you want a single implementation 
 file to include code from multiple header 
 files. 
  
  
-There is no way to split up the contents of a sin- gle  
+There is no way to split up the contents of a single  
 header file into multiple implementation files. 
 !!FILES 
  
  
@@ -4168,9 +4167,9 @@
 `__gcc__', `__cpp__', `__as__', `__ld__', and 
 `__gdb__' entries in __info__.'' 
 Using and Porting GNU CC (for version 2.0)'', Richard M. 
 Stallman; ''The C Preprocessor'', Richard M. Stallman; 
-''Debug- ging with GDB: the GNU Source-Level Debugger'', 
+''Debugging with GDB: the GNU Source-Level Debugger'', 
 Richard M. Stallman and Roland H. Pesch; ''Using as: the 
 GNU Assembler'', Dean Elsner, Jay Fenlason 
 ''ld: the GNU linker'', Steve Chamberlain and Roland 
 Pesch. 
This page is a man page (or other imported legacy content). We are unable to automatically determine the license status of this page.