C/C++ Users Group Library 1996 July/C-C++ Users Group Library July 1996.iso/vol_100/108_01/datedemo.doc

Text File | 1985-11-13 | 12KB | 398 lines

.OP









                    SOME_USEFUL_C_TIME_FUNCTIONS


                             By Bill Bolton

                             Software Tools,

                             P.O. Box 80,

                             Newport Beach,

                             NSW, 2106

                             AUSTRALIA



                             Phone [phone]


.HESOME USEFUL C TIME FUNCTIONS, By Bill Bolton              PAGE #

.PA
                    SOME_USEFUL_C_TIME_FUNCTIONS

ABSTRACT

     Rather  than show why you SHOULD use a programming  language

(in this case C) this article endeavours to show why I HAVE  used

it.  An example program demonstrates some aspects of the language

and is directly useful to owners of the Mountain Hardware 100,000

day  clock board as well as being generally useful to others with

real time clock facilities in their systems. A basic knowledge of

C  will  be required to understand the  example  program,  it  is

structured   ,commented  and  brief  extra  descriptions  of  the

functions are given.

WHY_C?

     While  I  have admired the C language from afar for  a  long

time  it is only relatively recently that I have  discovered  its

true delights. Because PASCAL was supposed to be THE language for

structured programming and there was a need for more structure in

my  programming  efforts,  I  had given PASCAL a  good  try.  The

system utilities which I was interested in writing just would not

come together properly under PASCAL,  I had the distinct  feeling

that  the  language  was  fighting  me  rather  than  helping  my

programming efforts.  My problems with PASCAL were not limited to

one  implementation,  PASCAL/M,  PASCAL  MT+,  PASCAL/Z and  UCSD

PASCAL all presented similar generic problems.  Before the PASCAL

phreaks  rise  up again,  let me stress that there are a  lot  of

things  about PASCAL which seem very worthwhile,  but I  couldn't

get  I to do what the things that I need done,  in a way I  could

grasp.

     So  for a time I retreated to assembler and FORTRAN for  the

serious  stuff  and BASIC for the quickies.  I found that it  was

possible to write reasonably structured programs in BASIC-80  (1)

and  FORTRAN  80  if  a moderate amount of care  was  taken   but

assembler was still the mainstay.  A brief flirtation with  FORTH

followed.  FORTH  showed  great  promise but  the  temptation  to

produce  code  which even I couldn't understand a week later  was

too  great.   Still,  the  idea  of  an  extensible  language  is

fascinating !

      In  the  meantime  I had been reading  "The  C  Programming

Language"(2)  and  it was clear that C was a  language  that  was

structured  but  not to verbose or inflexible (my major  problems

with  PASCAL).  It also appeared to have an extensible aspect  in

that the overhead in writing,  maintaining and using a library of

functions was quite small.

     At  about this time a major software project of the AWA Data

Group was started in C.  I was able to look over the programmer's

shoulder as the project progressed and I liked what I saw. What's

more  when  the project finished the programmers  involved  still

seemed  sane.  I  had  a  chance to  see  a  working  C  compiler

(Whitesmith's  for  the  PDP-11)  in  action  and  to  study  the

documentation.  While  the  Whitesmith's compiler was out  of  my

financial range for starters,  Leor Zolman had his BDS C compiler

available  at a very reasonable price (on reflection I'd say  its

an absolute steal) so that was my choice.

     As   soon  as  the compiler arrived I was  thrown  into  the

situation  that forced me to jump in at the deep end and  use  it

for a commercial project.  I had quoted a client 4 hours to write

a  file conversion program in assembler,  it soon became  obvious

that  I  had  made  a major error with my time  prediction  so  I

decided  to try and salvage the situation by writing the  program

in  C.  This first programming effort in C went so smoothly  that

the project was finished within the budgeted time and the  client

got  a  product with error handling an order of magnitude  better

than he would have received from an assembler program. Since then

I  have  abandoned BASIC almost entirely,  only use  FORTRAN  for

maintenance  and  have greatly reduced the  amount  of  assembler

programming.  The description of C as "a PASCAL that's not afraid

to get its hands dirty"(3) is very, very apt.

     If  there  is one single thing that delayed my entry in  the

world of C it is probably the sparsity of references to C in  the

microcomputer  journals.  There  have  been only  a   handful  of

programs  published and often they are examples that don't really

do anything useful (Small C in an obvious  exception).  Following

is one small effort to correct this situation.

THE_PROGRAM

     DATEDEMO  is a collection of C functions which will format a

string  with  the current date and current time  fetched  from  a

Mountain Hardware 100,000 day clock S-100 board. Several date and

time  formats  are  available  by specification  as  one  of  the

arguments to the primary functions.

     The  date formats are those used by the nearly all "Western"

countries outside North America (this is an appropriate point  to

complain  loudly about software packages that lock the user  into

MM/DD/YY date formats !!!) but are coded in a very general manner

so  that  nearly  any  imaginable date  format  could  be  easily

generated  and extra formats are easily added.  Time formats  are

somewhat  arbritary  and  only support 24 hour types  (perhaps  if

someone is keen they will do an AM/PM 12 hour type and contribute

it to DDJ).

      In practice I use the functions to put date and time on  in

the  page  headings  of  hardcopy output  and  in  some  business

packages.  The  functions live in my standard C function  library

and are linked in at compile time.  Normally the "#define" state

ments are loaded from a header file CLOCK.H but the are listed in

full here.  While the program is written for compilation with the

BDS  C compiler (Version 1.4X),  it should also compile with  the

other  C  compilers.  The only non-standard library  function  is

'initw' which is the BDS kludge to allow array initialisation, as

BDS  C does not support initialisers.  As far as I know all other

library  functions are supported by other C compilers (maybe  not

Small C).

THE_FUNCTIONS

'main' -      is  simply  an executable demonstration program  to

              tie the actual working functions together,  all the

              presently  supported  formats are displayed  and  a

              typical  use of the formatted strings is  shown  on

              the last line displayed.

'date' -      Attempts  to get basic data to format a date string

              form the clock,  if this fails an error message  is

              displayed.   Else  the  month  and  week  name  are

              determined and the string formatted.

'get_date' -  tests  if  the clock board is in the  system.  Then

              collects digits from the clock to get the number of

              days  since  31/Dec/1977  (N.B.   this  day  was  a

              Saturday,  so a simple MOD 7 on the total number of

              days since then makes Sunday day 1).  This value is

              then  turned  into the year, month and day  of  the

              month.

'month_day'-  Uses  an  initialised array to determine month  and

              day of the month from year and day of the year.

'ndays' -     Returns number of days in the specified year.

'leap' -      Thorough test for leap years.

'name_month'- Fills  a string with a month name g