helpful hints for c programming paul stuyvesant. clarity

Helpful Hints for C Programming

Paul Stuyvesant

CLARITY

CLARITY

CLARITY

CLARITY

CLARITY

CLARITY

CLARITYCLARITY

CLARITY

CLARITYCLARITYCLARITY

CLARITYCLARITY

CLARITYCLARITY

CLARITY

CLARITY

CLARITYCLARITY

CLARITYCLARITY

CLARITY

CLARITY CLARITYCLARITY

CLARITY


CLARITY

CLARITYCLARITY

CLARITYCLARITY

CLARITY

CLARITY

CLARITY

CLARITYCLARITY

CLARITYCLARITY CLARITY

CLARITYCLARITY

CLARITYCLARITYCLARITY CLARITY


CLARITY

CLARITY

CLARITY

CLARITY

CLARITYCLARITY

CLARITYCLARITY

CLARITY

CLARITY

CLARITYCLARITY

CLARITY

CLARITY

CLARITY CLARITYCLARITY

CLARITYCLARITY

CLARITYCLARITY

CLARITY

CLARITY

CLARITY CLARITY

CLARITY

CLARITY

CLARITY

Clarity

memcpy(&buffer[3],&ecbptr()->ebw000,InputLen);

if(*ptr==InputChar)

• Use space/blank lines to make readable

memcpy( &buffer[3], &ecbptr()->ebw000, InputLen);

if( *ptr == InputChar )

Clarity

int CVZZ( void )

{

int LoopCnt = 0;

ret = 0;

char buffer[20] = { 0 };

if( levtest( D4 ) )

{

ret = -1;

• Indent your code

Clarity

if( ++num == 10 )

{

num++; // or ++num;

if( num == 10 )

• Just because its less source doesn’t mean less object.

#include <stdio.h>

main(t,_,a ) char * a; { return!

0<t? t<3? main(-79,-13,a+ main(-87,1-_,

main(-86, 0, a+1 ) +a)):

1, t<_?

main( t+1, _, a ) :3,

main ( -94, -27+t, a ) &&t == 2 ?_

<13 ? main ( 2, _+1, "%s %d %d\n" )

:9:16: t<0?t<-72? main( _, t,

"@n'+,#'/*{}w+/w#cdnr/+,{}r/*de}+,/*{*+,/w{%+,/w#q#n+,

/#{l,+,/n{n+,/+#n+,/#;#q#n+,/+k#;*+,/'r :'d*'3,}{w+K w'K:'+}

e#';dq#'l q#'+d'K#!/+k#;q#'r}eKK#}w'r}eKK{nl]'/#;#q#n'){)#}w'){){nl]'/+#n';d}rw' i;# ){nl]!/n{n#'; r{#w'r nc{nl]'/#{l,+'K{rw' iK{;[{nl]'/w#q#n'wk nw' iwk{KK{nl]!/w{%'l##w#' i; :{nl]'/*{q#'ld;r'}{nlwb!/*de}'c ;;{nl'-{}rw]'/+,}

##'*}#nc,',#nw]'/+kd'+e}+;#'rdq#w! nr'/ ') }+}{rl#'{n' ')# }'+}##(!!/")

: t<-50? _==*a ?

putchar(31[a]): main(-65,_,a+1)

: main((*a == '/') + t, _, a + 1 )

: 0<t?

main ( 2, 2 , "%s") :*a=='/'||

main(0, main(-61,*a, "!ek;dc i@bK'(q)-[w]*%n+r3#l,{}:

\nuwloca-O;m .vpbks,fxntdCeghiry") ,a+1);}

On the first day of Christmas my true love gave to me

a partridge in a pear tree.

On the second day of Christmas my true love gave to me

two turtle doves

and a partridge in a pear tree.

On the third day of Christmas my true love gave to me

three french hens, two turtle doves


On the fourth day of Christmas my true love gave to me

four calling birds, three french hens, two turtle doves


On the fifth day of Christmas my true love gave to me

five gold rings;

four calling birds, three french hens, two turtle doves


Expertise

• New to C or C++ for TPF?– Get other people involved– Mix TPFers with C programmers

• Peer review a must for new C programmers– There is still a learning curve

• Online newsgroups are a fantastic resource

C or C++• Up to you!

• Spend lots of time arguing about it…

• Compile C programs as C++– Better type checking– Can use C++ features such as function

overloading– Saves a lot of time later on if you move to full C+

+ development.

C or C++

#ifdef __cplusplus

extern "C" {

#endif

int function1( char *, int );

void function2( char, int * );

#ifdef __cplusplus

}

#endif

• Header files should always have conditional compilation in them

• Will save you a lot of hassle later

C & C++ DLL• Very useful for common routines

• Easy for the programmer to create

• However …– Make sure programmers know about them– Make them easy to use– Just because you can doesn’t mean you should!

ECB

• Still hanging on in there

• Defined in c$eb0eb.h

• If you want to understand unions this is the place to look

• Not necessarily addressed by R9

• Accessed as required by ecbptr() macro

ecbptr()->

#define ecbptr() (*(struct eb0eb **)0x00000514ul)

00042 | * ecbptr()->ebw000 = 'Z'; 0003B2 5810 0514 00042 | L r1,1300 0003B6 92E9 1008 00042 | MVI (*)eb0eb.eb(r1,8),233

ecbptr()->

5810 0514 L R1,1300• Load from Base Register 0, offset X’514’• Address 0 + Prefixing = Prefix Page• Each I/Stream has its own Prefix Page

92E9 1008 MVI 8(R1),233• Use offset from R1 for ebw000

Useful Compiler Options

• AGGR/NOAGGR– Map of structures in compilation– Shows offsets, similar to PRINT GEN

• XREF/NOXREF– Shows information for each symbol– Displayed by source line/file/line number


• ARCH( )– Specify machine architecture– Generates different object code depending on

setting

• ROSTRINGS– Place R/O strings in program instead of static.– Similar to DC statement


• OPTIMISATION– Code is optimised by compiler– Can give unusual looking results– Make sure you test optimised code

• INLINE– Functions may be moved inline– Saves overhead of epilog/prolog

00007 | * {

00008 | * int a = 0,

00009 | * b = 2;

00010 | *

00011 | * a += 2;

00012 | *

00013 | * a = b + a;

00014 | *

00015 | * a++;

00016 | *

00017 | * b++;

00018 | *

00019 | * return a + b;

000060 41F0 0008 00019 | LA r15,8

00020 | * }

000064 00020 | @1L3 DS 0H

typedef

• Use and enforce company wide– It will be a pain when you start– In the long run it will make things easier

• Use names that are logical and easy to remember

• Place in ‘company’ header file

• Do NOT use assembler-ish names!

typedef unsigned char Uchar;

typedef unsigned short int Uint16; typedef signed short int Sint16; typedef unsigned long int Uint32; typedef signed long int Sint32; typedef int Sint32;

DO NOT!

typedef unsigned char BYTE; typedef unsigned short int HALFWORD;

typedef - example

typedef

• Worth being aware of standard typedefs

typedef char dft_alg;

typedef char dft_are;

typedef unsigned int dft_fad;

typedef char dft_fid;

typedef struct sw00sr dft_fil;

typedef char dft_hdr;

typedef struct id_list dft_idl;

typedef struct sw01sr dft_kyl;

Pointers

• The power of C language

• Use correct type of pointer

• Calling a function to check a field in specific structure?– Pass in address of field not address of structure.– Function is not then tied to a specific structure.

int Validate_Whatever1 ( struct wa0aa * ); int Validate_Whatever2 ( char * );

void CVZZ(void) { int ret = 0; struct wa0aa *AaaPtr = ecbptr()->ebccr1; // Validate_Whatever is tied to wa0aa structure ret = Validate_Whatever1( AaaPtr ); // Validate_Whatever2 can be used for any char pointer ret = Validate_Whatever2( &(AaaPtr->wa0xxx) );

Pointers

struct xxx *pXData = calloc( 1, sizeof(struct xxx) );

Later a change is made to structure type

struct yyy *pXData = calloc( 1, sizeof(struct xxx) );

Oops

Pointers

struct xxx *pXData = calloc( 1, sizeof( *pXData ) );

Later a change is made to structure type

struct yyy *pXData = calloc( 1, sizeof( *pXData ) );

Pointers

• Pointer and array notation are interchangeable

• Don’t mix and match within same function– Causes confusion– Is just not good practice

• Array notation can sometimes be clearer

void functionX( char *pMi0acc ) { char option = *pMi0acc; option = pMi0acc[0];

if( *(pMi0acc+1) == ‘A’ )

if( pMi0acc[1] == ‘A’ )

Pointers

• All pointer types can be dereferenced

• Use this facility to move data around

struct wa0aa *pWa0aa = ecbptr()->ebccr1, *pScrWa0aa = NULL; pScrWa0aa = calloc( 1, sizeof( *pWa0aa));

memcpy( pScrWa0aa, pWa0aa, sizeof( *pWa0aa));

OR

*pScrWa0aa = *pWa0aa;

Converting DSects

• Dsect are for Assembler Programs• Does exact equivalent help with C or C++?• Try to identify separate structures and declare

individually.• Place in C/C++ source file when developing

– Move to header when completed

• Use AGGR option in compile to view offsets• More later in TPFDF discussion

Functions

• Too many parameters - consider parameter list structure.

• Try to keep functions simple, easier said than done.

• Avoid overuse of #define to ‘simplify’ function calls.

struct param_list { char *pImsg; char *pOutput; };int Validate_Input( struct param_list * );

void CVZZ(void){ struct param_list parms = { 0 }; struct mi0mi *miptr = ecbptr()->ebccr0; int ret = 0; parms.pOutput = calloc( 1, 10000 ); parms.pImsg = miptr->mi0acc;

ret = Validate_Input( &parms );

struct param_list { char *pImsg; char *pOutput; int CharCount; };int Validate_Input( struct param_list * );

void CVZZ(void){ struct param_list parms = { 0 }; struct mi0mi *miptr = ecbptr()->ebccr0; int ret = 0; parms.CharCount = miptr->mi0cct - 4; parms.pOutput = calloc( 1, 10000 ); parms.pImsg = miptr->mi0acc;

ret = Validate_Input( &parms );

#define & functionsPCS_Link ( p_PCS_Ent, d_PCS_Buff, p_PCS_Last_Ent, p_PCS_Next_Ent_Item, st_PCS_Ent_Item );

#define PCS_Link(NEW, BUFFER, PREVIOUS, NEXTFIELD, STRUCT) \ ( PCS_BuffGet(NEW,BUFFER,sizeof(STRUCT)), \ *PREVIOUS = NEW, \ PREVIOUS = (STRUCT **) &(NEW->NEXTFIELD) )

#define PCS_BuffGet(NEW,BUFFER,SIZE) \ ( ( (NEW = XXBuffGet(BUFFER,SIZE)) != NULL ) ? \ ( TRUE ) : \ ( snapc (SNAPC_RETURN, 0xB00B00, "NO MORE BUFFER", \ NULL, 'U', SNAPC_REGS, SNAPC_ECB, "PCS000"), \ exit (1), \ FALSE ))

Default Values

• Most functions have default valuesptr = dfopn( FNAME, FID, 0 );

• What does the 0 in dfopn actually mean?

• Take a look in c$cdfeq.h – Contains values used by the various commands– These are usually the dft_opt parameter

Default Values

#define DFOPN_ALG 0x00000001UL

#define DFOPN_FADDR 0x00000002UL

#define DFOPN_ORD 0x00000004UL

#define DFOPN_SPACE 0x00000008UL

#define DFOPN_TAPE 0x00000010UL

#define DFOPN_AREA 0x00000020UL

#define DFOPN_HOLD 0x01000000UL

#define DFOPN_NOHOLD 0x00000000UL

#define DFOPN_DLI 0x00000000UL

#define DFOPN_DETAC 0x00008000UL

#define DFOPN_NODET 0x00000000UL

Default Values

• Understanding can make problem solving easier.

find_record( D6, NULL, “PS”, 0, UNHOLD);

• Why does the above line compile OK but cause Control 8 at exit?

Default Values

find_record( D6, NULL, “PS”, 0, UNHOLD);

/* FIND/FILE constants */ enum t_act {NOHOLD, UNHOLD, NOREL}; #define HOLD UNHOLD

find_record( D6, NULL, “PS”, 0, HOLD);

TPFDF

• Error Handling – Use DF_SERRC macro

• Converting DSECT to Header– Separate structures for each LREC

• Understand the typedefs and defines– Check the cdf header files

Error Handling

• Use built in macros in C$CDFERR.h

• DF_SERRC can be particularly useful– Check if dump already taken by TPFDF such as

DB0100 or DB0102– Do not repeat dump if already taken

pk40lrec = dfred( PCSFile, 0 ); if( DF_OK( PCSFile ) ) { // Found Record process } else if( DF_NR( PCSFile ) ) { // Not Found process } else { serrc_op( SERRC_RETURN, 0xC0FFEE, "OH NO - NOT AGAIN", NULL); }

pk40lrec = dfred( PCSFile, 0 ); if( DF_OK( PCSFile ) ) { // Found Record process } else if( DF_NR( PCSFile ) ) { // Not Found process } else { if( !(DF_SERRC( PCSFile ))) { serrc_op( SERRC_RETURN, 0xC0FFEE, "OH NO - NOT AGAIN", NULL); } }

DSect Conversion

• Each Lrec type should be separate structure.

• No need to build ‘overall’ structure

• Makes header much easier to read

• Why pretend it is Assembler?

GR99REC&CG1 DS 0CL1 1ST RECORD START (1=VARIABLE,ELSE SIZE)

GR99SIZ&CG1 DS H SIZE OF LOGICAL RECORD

GR99KEY&CG1 DS X LOGICAL RECORD IDENTIFIER

*********************************************************************

* DATA DEFINITIONS *

*********************************************************************

AIF ('&GR99RR1' EQ '1').KEYEQ GO IF NOT FIRST ISSUE

#GR99K40 EQU X'40' LOGICAL RECORD KEY X'40'

#GR99L40 EQU GR99E40&CG1-GR99REC&CG1 LENGTH OF RECORD KEY X'40'

#GR99K50 EQU X'50' LOGICAL RECORD KEY X'50'

#GR99L50 EQU GR99E50&CG1-GR99REC&CG1 LENGTH OF RECORD KEY X'50'

&GR99RR1 SETB 1 INDICATE 1ST TIME THROUGH

.KEYEQ ANOP GENERATE DSECT NAME

GR99ORG&CG1 EQU * START OF LOGICAL RECORD DESCRIPTION

*********************************************************************

* DESCRIPTION OF X'40' LOGICAL RECORD TYPE

*********************************************************************

GR99FIX&CG1 DS CL8 NUMBER OF LAST MESSAGE IN TABLE

GR99FRX&CG1 DS CL8 NUMBER OF FIRST MESSAGE IN TABLE

GR99IND&CG1 DS X INDICATOR

GR99NBX&CG1 DS H NUMBER OF MESSAGES IN TABLE

GR99E40&CG1 EQU * END OF LOGICAL RECORD KEY '40'

ORG GR99ORG&CG1

*********************************************************************

* DESCRIPTION OF X'50' LOGICAL RECORD TYPE

*********************************************************************

GR99LNO&CG1 DS CL2 LANGUAGE CODE - NUMBER

GR99NAM&CG1 DS CL10 LANGUAGE NAME

GR99LNN&CG1 DS XL1 LANGUAGE CODE - BINARY

GR99LST&CG1 DS CL8 NUMBER OF LAST MESSAGE

GR99E50&CG1 EQU * END OF LOGICAL RECORD KEY '50'

ORG GRXXORG&CG1

DBRED REF=GR99PS,REG=R2,ERROR=ERRORLAB, *

KEY1=(PKY=#GR99K40)#IF SW00RTN,EQ,Z

MVC EBW000(L’GR99LNO),GR99LNO

struct gr99rec {

dft_siz gr99siz;

dft_key gr99key;

union lrecs {

struct k40lrec {

char _gr99fix[8];

char _gr99frx[8];

char _gr99ind;

unsigned short int _gr99nbx;

} Key40Lrec;

struct k50lrec {

char _gr99lno[2];

char _gr99nam[10];

char _gr99lnn;

char _gr99lst[8];

} Key50Lrec;

} LrecStructs;

};

#define gr99fix LrecStructs.Key40Lrec._gr99fix

#define gr99frx LrecStructs.Key40Lrec._gr99fix

#define gr99lnn LrecStructs.Key50Lrec._gr99lnn

dft_fil *pGr99fil = NULL;

dft_kyl keys = { 0 };

struct gr99rec *pLrec = NULL;

pGr99fil = dfopn( GR99FNAME, GR99FID, 0 );

SetUpKeys( pGr99fil, &keys );

pLrec = dfred( pGr99fil, 0 )

if( DF_OK( pGr99fil ) )

{

ecbptr()->ebw000 = pLrec->gr99lnn // Key 50 lrec

ecbptr()->ebw000 = pLrec->gr99ind // Key 40 lrec

struct k40lrec {

dft_siz gr99siz;

dft_key gr99key;

char gr99fix[8];

char gr99frx[8];

char gr99ind;

unsigned short int gr99nbx;

};

struct k50lrec {

dft_siz gr99siz;

dft_key gr99key;

char gr99lno[2];

char gr99nam[10];

char gr99lnn;

char gr99lst[8];

};

dft_fil *pGr99fil = NULL;

dft_kyl keys = { 0 };

struct k40lrec *pLrec = NULL;

pGr99fil = dfopn( GR99FNAME, GR99FID, 0 );

SetUpKeys( pGr99fil, &keys );

pLrec = dfred( pGr99fil, 0 )

if( DF_OK( pGr99fil ) )

{

ecbptr()->ebw000 = pLrec->gr99lnn // COMPILER ERROR

ecbptr()->ebw000 = pLrec->gr99ind // Key 40 lrec

printf/sprintf

• Understand the different modifiers and flags available

• Use a PC compiler for quick ‘what if’ scenarios

• The ‘*’ modifier is particularly useful for variable length fields

printf/sprintf

char name[] = "STUYVESANT"; int count = 100;

printf("%s\n", name ); STUYVESANTprintf("%5s\n", name ); STUYVESANTprintf("%.5s\n", name ); STUYV

printf("%d\n", count ); 100printf("%5d\n", count ); 100printf("%.5d\n", count ); 00100

printf/sprintf

int count = 171;int negone = -1;

printf("%x\n", count ); abprintf("%X\n", count ); ABprintf("%#X\n", count ); 0XABprintf("%#.4x\n", count ); 0x00ab

printf("%u\n", negone ); 4294967295printf("%d\n", negone ); -1

* modifier

• Length is not specified in string itself

• Replaced by integer value in argument

• Very useful for runtime replacement

• Can be used in printf/sprintf

• Has a totally different meaning in scanf

char name[] = “UNION", fname1[] = “CITY", fname2[] = “BLUES“;

int len = strlen(name);

printf("%s %s\n", name, fname1 );printf("%*s %s\n", len, " ", fname2 );

UNION CITY BLUES

printf("%*s %s\n", len, " ", fname2 );

printf("%5s %s\n", len, " ", fname2 );

char name[] = "WALES", fname1[] = "CHARLES", fname2[] = "CAMILLA", fname3[] = "HARRY", fname4[] = "WILLIAM";


printf("%s %s\n", name, fname1 );printf("%*s %s\n", len, " ", fname2 );printf("%*s %s\n", len, " ", fname3 );printf("%*s %s\n", len, " ", fname4 );

RESULT

WALES CHARLES CAMILLA HARRY WILLIAM

char name[] = “STEVENSON", fname1[] = "CHARLES", fname2[] = "CAMILLA", fname3[] = "HARRY", fname4[] = "WILLIAM";


printf("%s %s\n%*s %s\n%*s %s\n%*s %s\n", name, fname1, len, " ", fname2, len, " ", fname3, len, " ", fname4 );

RESULT

STEVENSON CHARLES CAMILLA HARRY WILLIAM

sprintf

• Parameters similar to printf

• Used to build output into a buffer

• Returns integer value– Count of number of bytes added to the buffer.

• Make sure buffer is big enough!

• Can pass buffer off to FMSG/WTOPC

• Try to avoid multiple calls to sprintf

pk40lrec = dfred( PCSFile, 0 ); while( DF_OK( PCSFile ) ) { ind += sprintf( buffer + ind, "%d %.8s %2X\n", pk40lrec->pp40cnt, pk40lrec->pp40fld, pk40lrec->pp40ind ); pk40lrec = dfred( PCSFile, 0 ); }

pk40lrec = dfred( PCSFile, 0 ); while( DF_OK( PCSFile ) ) { ind += sprintf( buffer + ind, "%d ", pk40lrec->pp40cnt ); ind += sprintf( buffer + ind, "%.8s ", pk40lrec->pp40fld ); ind += sprintf( buffer + ind, "%2X\n", pk40lrec->pp40ind ); pk40lrec = dfred( PCSFile, 0 ); }

#pragma map

• Replaces entry point function with supposedly ‘real world’ name.

• Adds confusion

• Removes clarity

• When I see it in a program it makes me feel like this

#pragma map

• Its 2AM and your friendly coverage analyst is woken up by a dump on system.

• Leaps into action

#pragma map

• Post processes dump to examine.

• Looking through stacks in dump finds call to Go_Print( ptr ).

• Searches through component for Go_Print function - no luck

#pragma map

• Pulls up Build Script for failing program– Finds 8 components making up the DLM.

• Searches through each for Go_Print function.– No luck.

• Thinks!– Perhaps it is a DLL function call.

• Sadly no - after 15 minutes searching imports

#pragma map

• At top of program finds #include <project.h>

• Inside finds 25 #include statements

• Starts searching each one.– Of course each nested include has nested

includes– And some are conditional

• Eventually - 2 hours later our hero finds ...

#pragma map

#pragma map (Go_Print, “ABCD” )

Which means the line:

Go_Print( ptr ); // Is really

ABCD( ptr );

#pragma map

• Having the entry point to a DLM/Assembler program clearly identifiable is a plus point.

• Having program calls easily identifiable from function calls is a plus point.

• ABCD is just as portable as GoPrint

• Do standard C programmers feel a need to get rid of main() function?

Dump Analysis

• Easier in C/C++ than Assembler– When you know the tricks

• Understand the DLM/DLL Layout– File Map in Link Output or Link Map in dump

• More information available to help resolve– Function Stacks– Static Stacks

Dump Analysis

• Know your Register Conventions– Stack and Static Pointer– Parameter Lists – C vs C++

• Understand the Pseudo Assembler Listing– What is #MXTEMP_2?

• How are DLL functions accessed?

THE END

• Thanks for staying awake

• Enjoy the rest of the conference

• A copy of presentation is available atwww.pcs-training.co.uk/downloads.htm

• Any questions drop me a line at [email protected]

helpful hints for c programming paul stuyvesant. clarity

Documents