error 2 error #6633

[ali_p_]

Hello all 

im a physics student and i am working on a simulation project so my doctor gave me an old fortran script for the simulation which he had written few years ago the problem is when i try to debug it,it shows error #6633: the type of the actual argument differs from the type of the dummy argument.[0]

several times this is one of them  :

call moveto_w(0,0,wxy)
       dummy=setcolor(9)
       ekm=abs(phif(2,2))
       do j=1,ngx    
	ekm=amax1(ekm,phif(1,j),phif(2,j),phif(3,j))
       end do
       call moveto_w(500/ngx,phif(1,1)*100*.8/ekm,wxy) 
       do j=1,ngx 
	   dummy=lineto_w(j*500/ngx,phif(1,j)*100*.8/ekm)    
       end do
       call moveto_w(500/ngx,phif(2,1)*100*.8/ekm,wxy) 
       do  j=1,ngx 
	   dummy=lineto_w(j*500/ngx,phif(2,j)*100*.8/ekm)    
       end do

and the declarations of this subroutine goes like this:

subroutine fieldf (ith)
**************************************************************************      
       include 'fgraph.fd'
	common/param/ nsp,lx,ly,dx,dy,dt,nt
	common/cfield/ ngx,ngy ,iw,rho0,g1(34,34),ex(34,34),ey(34,34)
     $,bz(34,34)   
	common /xvp/ x(25602),y(25602),vx(25602),vy(25602)
	common/cntrl/it,time,ithlx,ithly,iex,iey,ixvx,iyvy,
     $  ivxvy,ifvx,ifvy          
	integer*2 dummy
       record /wxycoord/ wxy  
	integer ith,l,ng2,hdx,hdy,ik,jk 
	real lx,ly,pi,erun1,erun2,kdx2,sm(34),ksq(34),eset,phif(34,34)
     $,rhof(34,34),sc(34),rhok(34),lxlyi
	data ng2,pi/0,3.128/

ps:im not very familiar with fortran coding and im using vs2012 

[Les_Neilson]

I would guess that the subroutine moveto_w and lineto_w are expecting floating point numbers.

For example:

subroutine moveto_w(x, y, wxy)

real x, y, wxy

and similar for lineto_w

The call moveto_w(0, 0, wxy) is passing integer zero actual values to the floating point dummy arguments x and y. Hence the mismatch.

If my guess is correct then you need to change 0 to 0.0

Also expressions such as 500/ngx are integer division. 500/1000 will give the result 0 not 0.5.

If a floating point value is expected then you need to change 500 to 500.0

Hope this helps

Les

[ali_p_]

thanks for responding 

i tried 0.0 instead of 0 but the same error appear:error #6633: the type of the actual argument differs from the type of the dummy argument.[0.0]

and when trying x,y :error #6633: the type of the actual argument differs from the type of the dummy argument.[x]

error #6633: the type of the actual argument differs from the type of the dummy argument.

[mecej4]

Instead of blindly trying different types of arguments to the routine moveto_w, consult the interface declaration to the routine in file ifqwin.f90, which is located in the compiler's include subdirectory, and make sure that the actual arguments are of the required type and kind. 

[Arjen_Markus]

To solve this, you will need to retrieve the interface definition of the routines. Apparently the compiler knows about them via an interface block or a module, or, it uses the first call to moveto_w and lineto_w to determine what other calls should look like. Chances are the include file "fgraph.fd" holds the secret.

That said, I do notice a few things in the code that you may want to look at more closely:

• You have 500/ngx in there. As these are two integers, the outcome is again an integer according to the rules in Fortran. That means no fraction. As long as ngx is smaller than 500, the result will be non-zero, but for instance 500/1000 is rounded to 0.
• Your calls mix integers and reals - call moveto_w(500/ngx,phif(1,1)*100*.8/ekm,wxy) for instance. The first argument is an integer (see first bullet0 and the second is a real - at th every least because of the factor .8 (again the rules that are used in Fortran)

This is going to upset the compiler as it wants the argument lists to be uniform in type (unless routine overloading is defined, but given the "age" of the code, I doubt that is the case).

[ali_p_]

arjenmarkus wrote:

To solve this, you will need to retrieve the interface definition of the routines. Apparently the compiler knows about them via an interface block or a module, or, it uses the first call to moveto_w and lineto_w to determine what other calls should look like. Chances are the include file "fgraph.fd" holds the secret.

That said, I do notice a few things in the code that you may want to look at more closely:

• You have 500/ngx in there. As these are two integers, the outcome is again an integer according to the rules in Fortran. That means no fraction. As long as ngx is smaller than 500, the result will be non-zero, but for instance 500/1000 is rounded to 0.
• Your calls mix integers and reals - call moveto_w(500/ngx,phif(1,1)*100*.8/ekm,wxy) for instance. The first argument is an integer (see first bullet0 and the second is a real - at th every least because of the factor .8 (again the rules that are used in Fortran)

This is going to upset the compiler as it wants the argument lists to be uniform in type (unless routine overloading is defined, but given the "age" of the code, I doubt that is the case).

as i mentioned i have no fortran experience but from what i understood the fgraph.fd is a missing file or subroutine ?!

[Arjen_Markus]

No, it is a file on disk somewhere:
 

subroutine fieldf (ith)
**************************************************************************     
include 'fgraph.fd'
common/param/ nsp,lx,ly,dx,dy,dt,nt

This is bound to contain some useful definitions. Its location is given by compiler options like -I<name of a directory>.

If the compiler were not able to find it, it would complain about that, so it is definitely somewhere around.

[ali_p_]

arjenmarkus wrote:

No, it is a file on disk somewhere:
 

subroutine fieldf (ith)
**************************************************************************     
include 'fgraph.fd'
common/param/ nsp,lx,ly,dx,dy,dt,nt

This is bound to contain some useful definitions. Its location is given by compiler options like -I<name of a directory>.

If the compiler were not able to find it, it would complain about that, so it is definitely somewhere around.

well i dont see it i have only two files .for and .dat 

but i've found somthing on the internet about fgraph.fd :

*fgraph.fd - declare constants and functions for graphics library
*
*   Copyright (c) 1987-1989 Microsoft Corporation.  All rights reserved.
*
*Purpose:
*   This file declares the graphics library functions and
*   the manifest constants that are used with them.
*
*******************************************************************************

$NOTRUNCATE       ! required for some names to be significant
$NOTSTRICT         ! uses structures which are non-standard conforming

* user-visible declarations for FORTRAN Graphics Library

* structure for getvideoconfig() as visible to user
      STRUCTURE/videoconfig/
        INTEGER*2 numxpixels       ! number of pixels on X axis
        INTEGER*2 numypixels       ! number of pixels on Y axis
        INTEGER*2 numtextcols      ! number of text columns available
        INTEGER*2 numtextrows      ! number of text rows available
        INTEGER*2 numcolors        ! number of actual colors
        INTEGER*2 bitsperpixel     ! number of bits per pixel
        INTEGER*2 numvideopages    ! number of available video pages
        INTEGER*2 mode             ! current video mode
        INTEGER*2 adapter          ! active display adapter
        INTEGER*2 monitor          ! active display monitor
        INTEGER*2 memory           ! adapter video memory in K bytes
      END STRUCTURE

* return value of setlogorg(), etc.
      STRUCTURE/xycoord/
        INTEGER*2 xcoord
        INTEGER*2 ycoord
      END STRUCTURE

* structure for text position
      STRUCTURE/rccoord/
        INTEGER*2 row
        INTEGER*2 col
      END STRUCTURE

* SETUP AND CONFIGURATION

* arguments to setvideomode()
      INTEGER*2, $MAXRESMODE, $MAXCOLORMODE, $DEFAULTMODE,$TEXTBW40,
     +           $TEXTC40,$TEXTBW80,$TEXTC80, $MRES4COLOR,$MRESNOCOLOR,
     +           $HRESBW,$TEXTMONO,$HERCMONO, $MRES16COLOR,$HRES16COLOR,
     +           $ERESNOCOLOR,$ERESCOLOR, $VRES2COLOR,$VRES16COLOR,
     +           $MRES256COLOR,$ORESCOLOR
      PARAMETER($MAXRESMODE   =-3)    ! graphics mode with highest resolution
      PARAMETER($MAXCOLORMODE =-2)    ! graphics mode with most colors
      PARAMETER($DEFAULTMODE  =-1)    ! restore screen to original mode
      PARAMETER($TEXTBW40      =0)    ! 40 x 25 text, 16 grey
      PARAMETER($TEXTC40       =1)    ! 40 x 25 text, 16/8 color
      PARAMETER($TEXTBW80      =2)    ! 80 x 25 text, 16 grey
      PARAMETER($TEXTC80       =3)    ! 80 x 25 text, 16/8 color
      PARAMETER($MRES4COLOR    =4)    ! 320 x 200, 4 color
      PARAMETER($MRESNOCOLOR   =5)    ! 320 x 200, 4 grey
      PARAMETER($HRESBW        =6)    ! 640 x 200, BW
      PARAMETER($TEXTMONO      =7)    ! 80 x 25 text, BW
      PARAMETER($HERCMONO      =8)    ! 720 x 348, BW for HGC
      PARAMETER($MRES16COLOR   =13)   ! 320 x 200, 16 color
      PARAMETER($HRES16COLOR   =14)   ! 640 x 200, 16 color
      PARAMETER($ERESNOCOLOR   =15)   ! 640 x 350, BW 
      PARAMETER($ERESCOLOR     =16)   ! 640 x 350, 4 or 16 color 
      PARAMETER($VRES2COLOR    =17)   ! 640 x 480, BW 
      PARAMETER($VRES16COLOR   =18)   ! 640 x 480, 16 color 
      PARAMETER($MRES256COLOR  =19)   ! 320 x 200, 256 color
      PARAMETER($ORESCOLOR     =64)   ! 640 x 400, 1 of 16 colors (Olivetti)

* videoconfig adapter values 
* these manifest constants can be used to test adapter values for
* a particular adapter using the bitwise-AND operator
      INTEGER*4 $MDPA,$CGA,$EGA,$MCGA,$VGA,$HGC,$OCGA,$OEGA,$OVGA
      PARAMETER($MDPA     =#0001)    ! Monochrome Display Adapter (MDPA) 
      PARAMETER($CGA      =#0002)    ! Color Graphics Adapter     (CGA)
      PARAMETER($EGA      =#0004)    ! Enhanced Graphics Adapter  (EGA)
      PARAMETER($VGA      =#0008)    ! Video Graphics Array       (VGA) 
      PARAMETER($MCGA     =#0010)    ! MultiColor Graphics Array  (MCGA)
      PARAMETER($HGC      =#0020)    ! Hercules Graphics Card     (HGC)
      PARAMETER($OCGA     =#0042)    ! Olivetti Color Graphics Adapter (OCGA)
      PARAMETER($OEGA     =#0044)    ! Olivetti Enhanced Graphics Adapter (OEGA)
      PARAMETER($OVGA     =#0048)    ! Olivetti Video Graphics Array (OVGA)

* videoconfig monitor values
* these manifest constants can be used to test monitor values for
* a particular monitor using the bitwise-AND operator
      INTEGER*4 $MONO,$COLOR,$ENHCOLOR,$ANALOGMONO,$ANALOGCOLOR,$ANALOG
      PARAMETER($MONO       =#0001)     ! Monochrome
      PARAMETER($COLOR      =#0002)     ! Color (or Enhanced emulating color)
      PARAMETER($ENHCOLOR   =#0004)     ! Enhanced Color
      PARAMETER($ANALOGMONO =#0008)     ! Analog Monochrome only
      PARAMETER($ANALOGCOLOR=#0010)     ! Analog Color only
      PARAMETER($ANALOG     =#0018)     ! Analog

* COORDINATE SYSTEMS

* OUTPUT ROUTINES

* control parameters for Rectangle, Ellipse and Pie
      INTEGER*2 $GBORDER,$GFILLINTERIOR,
     +          $GCLEARSCREEN, $GVIEWPORT,$GWINDOW

      PARAMETER($GBORDER       =2)      ! draw outline only
      PARAMETER($GFILLINTERIOR =3)      ! fill using current fill mask

      PARAMETER($GCLEARSCREEN=0)
      PARAMETER($GVIEWPORT   =1)
      PARAMETER($GWINDOW     =2)

* TEXT
      INTEGER*4 $GCURSOROFF,$GCURSORON,$GWRAPOFF,$GWRAPON
      PARAMETER($GCURSOROFF=0)
      PARAMETER($GCURSORON =1)

      PARAMETER($GWRAPOFF  =0)
      PARAMETER($GWRAPON   =1)

      INTEGER*4 $GSCROLLUP, $GSCROLLDOWN
      PARAMETER($GSCROLLUP   =1)
      PARAMETER($GSCROLLDOWN =-1)

C request maximum number of rows in _settextrows and _setvideomoderows
      INTEGER*4 $MAXTEXTROWS
      PARAMETER($MAXTEXTROWS =-1)

* "action verbs" for putimage()
      INTEGER*4 $GPSET,$GPRESET,$GAND,$GOR,$GXOR
      PARAMETER($GPSET         =3)
      PARAMETER($GPRESET       =2)
      PARAMETER($GAND          =1)
      PARAMETER($GOR           =0)
      PARAMETER($GXOR          =4)

* universal color values:
      INTEGER*4 $BLACK,$BLUE,$GREEN,$CYAN,$RED,$MAGENTA,$BROWN,
     +          $WHITE,$GRAY, $LIGHTBLUE,$LIGHTGREEN,$LIGHTCYAN,
     +          $LIGHTRED,$LIGHTMAGENTA, $LIGHTYELLOW,$BRIGHTWHITE
      PARAMETER($BLACK         =#000000)
      PARAMETER($BLUE          =#2a0000)
      PARAMETER($GREEN         =#002a00)
      PARAMETER($CYAN          =#2a2a00)
      PARAMETER($RED           =#00002a)
      PARAMETER($MAGENTA       =#2a002a)
      PARAMETER($BROWN         =#00152a)
      PARAMETER($WHITE         =#2a2a2a)
      PARAMETER($GRAY          =#151515)
      PARAMETER($LIGHTBLUE     =#3F1515)
      PARAMETER($LIGHTGREEN    =#153f15)
      PARAMETER($LIGHTCYAN     =#3f3f15)
      PARAMETER($LIGHTRED      =#15153f)
      PARAMETER($LIGHTMAGENTA  =#3f153f)
      PARAMETER($LIGHTYELLOW   =#153f3f)
      PARAMETER($BRIGHTWHITE   =#3f3f3f)

* mono mode F color values:
      INTEGER*4 $MODEFOFF,$MODEFOFFTOON,$MODEFOFFTOHI,$MODEFONTOOFF,
     +          $MODEFON,$MODEFONTOHI,$MODEFHITOOFF,$MODEFHITOON,
     +          $MODEFHI
      PARAMETER($MODEFOFF      =0)
      PARAMETER($MODEFOFFTOON  =1)
      PARAMETER($MODEFOFFTOHI  =2)
      PARAMETER($MODEFONTOOFF  =3)
      PARAMETER($MODEFON       =4)
      PARAMETER($MODEFONTOHI   =5)
      PARAMETER($MODEFHITOOFF  =6)
      PARAMETER($MODEFHITOON   =7)
      PARAMETER($MODEFHI       =8)

* mono mode 7 color values:
      INTEGER*4 $MODE7OFF,$MODE7ON,$MODE7HI
      PARAMETER($MODE7OFF      =0)
      PARAMETER($MODE7ON       =1)
      PARAMETER($MODE7HI       =2)

* external function declarations

      INTEGER*2 setvideomode[EXTERN]
      INTEGER*2 setvideomoderows[EXTERN]
      INTEGER*2 setactivepage[EXTERN]
      INTEGER*2 setvisualpage[EXTERN]
      INTEGER*2 getactivepage[EXTERN]
      INTEGER*2 getvisualpage[EXTERN]
      EXTERNAL getvideoconfig
      EXTERNAL setvieworg
      EXTERNAL getviewcoord
      EXTERNAL getphyscoord
      EXTERNAL setcliprgn
      EXTERNAL setviewport
      EXTERNAL clearscreen
      EXTERNAL moveto
      EXTERNAL getcurrentposition
      INTEGER*2 lineto[EXTERN]
      INTEGER*2 rectangle[EXTERN]
      INTEGER*2 ellipse[EXTERN]
      INTEGER*2 arc[EXTERN]
      INTEGER*2 pie[EXTERN]
      INTEGER*2 setpixel[EXTERN]
      INTEGER*2 getpixel[EXTERN]
      INTEGER*2 floodfill[EXTERN]
      INTEGER*2 setcolor[EXTERN]
      INTEGER*2 getcolor[EXTERN]

      EXTERNAL setlinestyle
      INTEGER*2 getlinestyle[EXTERN]
      EXTERNAL setfillmask
      EXTERNAL getfillmask
      INTEGER*4 setbkcolor[EXTERN]
      INTEGER*4 getbkcolor[EXTERN]
      INTEGER*4 remappalette[EXTERN]
      INTEGER*2 remapallpalette[EXTERN]
      INTEGER*2 selectpalette[EXTERN]
      INTEGER*2 settextrows[EXTERN]
      EXTERNAL settextwindow
      EXTERNAL scrolltextwindow
      EXTERNAL outtext
      INTEGER*2 wrapon[EXTERN]
      INTEGER*2 displaycursor[EXTERN]
      INTEGER*2 settextcursor[EXTERN]
      INTEGER*2 gettextcursor[EXTERN]
      EXTERNAL settextposition
      EXTERNAL gettextposition
      INTEGER*2 settextcolor[EXTERN]
      INTEGER*2 gettextcolor[EXTERN]
      EXTERNAL getimage
      EXTERNAL putimage
      INTEGER*4 imagesize[EXTERN]

* WINDOW COORDINATE SYSTEM 

* structure for window coordinate pair 
      STRUCTURE/wxycoord/
        DOUBLE PRECISION wx	! window x coordinate
        DOUBLE PRECISION wy	! window y coordinate
      END STRUCTURE

       INTEGER*2 setwindow[EXTERN]

       EXTERNAL getwindowcoord
       EXTERNAL getviewcoord_w
       EXTERNAL getcurrentposition_w

* window coordinate entry points for graphics output routines

      INTEGER*2 arc_w[EXTERN]
      INTEGER*2 ellipse_w[EXTERN]
      INTEGER*2 floodfill_w[EXTERN]
      INTEGER*2 getpixel_w[EXTERN]
      INTEGER*2 lineto_w[EXTERN]
      EXTERNAL moveto_w
      INTEGER*2 pie_w[EXTERN]
      INTEGER*2 rectangle_w[EXTERN]
      INTEGER*2 setpixel_w[EXTERN]
      EXTERNAL getimage_w
      INTEGER*2 imagesize_w[EXTERN]
      EXTERNAL putimage_w

      STRUCTURE/fontinfo/
        INTEGER*2 type		! b0 set = vector,clear = bit map
        INTEGER*2 ascent	! pix dist from top to baseline
        INTEGER*2 pixwidth	! character width in pixels, 0=prop
        INTEGER*2 pixheight	! character height in pixels
        INTEGER*2 avgwidth	! average character width in pixels
        CHARACTER*81 filename   ! file name including path
        CHARACTER*32 facename   ! font name
      END STRUCTURE

* Font parameters
      
      INTEGER*2 $NO_SPACE, $FIXED_SPACE, $PROP_SPACE
      PARAMETER ($NO_SPACE    = 0)
      PARAMETER ($FIXED_SPACE = 1)
      PARAMETER ($PROP_SPACE  = 2)

      INTEGER*2 $NO_FONT_MAP, $VECTOR_MAP, $BIT_MAP
      PARAMETER ($NO_FONT_MAP = 0)
      PARAMETER ($VECTOR_MAP  = 1)
      PARAMETER ($BIT_MAP     = 2)
      
      INTEGER*2 registerfonts[EXTERN]
      EXTERNAL unregisterfonts
      INTEGER*2 setfont[EXTERN]
      INTEGER*2 getfontinfo[EXTERN]
      EXTERNAL outgtext
      INTEGER*2 getgtextextent[EXTERN]

now do i just create a file named "fgraph.fd" with this code in it and add it to the project's directory?

[Arjen_Markus · Accepted solution]

>> now do i just create a file named "fgraph.fd" with this code in it and add it to the project's directory?

No, need for that. Like I said, the compiler can find it somewhere, possibly in its own installation directory.

This does make it possible to guess what is going on though. The structure "wxycoord" uses double precision reals, so my guess is that all coordinates are expected to be double precision. You will have to change your declarations from "real" to "double precision". And lines like:

call moveto_w(0,0,wxy)

call moveto_w(500/ngx,phif(1,1)*100*.8/ekm,wxy)

to (the "d0" indicates a double precision constant)

call moveto_w(0.0d0,0.0d0,wxy)

call moveto_w(500.0d0/ngx,phif(1,1)*100*.8d0/ekm,wxy)

to force the coordinates to double precision.

[ali_p_]

oh that worked but now the same error with this line 

call moveto_w((j-1)*500.0d0/(ngx-1),g1(1,j)*gmax,wxy)

[Arjen_Markus]

Right, the variables g1 (an array in a COMMON block) and gmax are single precision reals, so their product, g1(1,j)*gmax, is too. I suspect that you have a lot of routines that use these COMMON blocks, so the simplest solution is to leave these variables as they are and use the dble() function to convert the result to double precision. That is:

call moveto_w((j-1)*500.0d0/(ngx-1),g1(1,j)*gmax,wxy)

becomes:

call moveto_w((j-1)*500.0d0/(ngx-1), dble(g1(1,j)*gmax), wxy)

(Spaces added for clarity). You may have to do that in more than one location.

[ali_p_]

that worked too u r a fortran genius 

but now im having trouble with too long statement:

call moveto_w((i-1)*100.0d0/(ngy-1),(i-1)*50.0d0/(ngy-1)+g1(i,1)*gmax,wxy)  

i tried the continuation lines with "&"

it didn't work 

[mecej4]

The '&' character is used as a continuation character at the end of lines that are followed by continuation lines in free format source.

In fixed format source, which is what you are using, use a continuation character in column-6 of continuation lines.

[ali_p_]

mecej4 wrote:

The '&' character is used as a continuation character at the end of lines that are followed by continuation lines in free format source.

In fixed format source, which is what you are using, use a continuation character in column-6 of continuation lines.

the "$" worked thanx

[Arjen_Markus]

I suspect that you are using "fixed form" sources. That means:

• The first six positions on a line are reserved and lines can be no longer than 72 characters.
• If you want to use long lines, they have to broken up using a non-zero, non-space character in the sixth position of the next line

Alternatively you can use the "long lines" option of the compiler.

In the source code breaking up a line is done like this:

*234567890 - just showing the positions on the line
      call mysubroutine( with, a, very, long, list, of,
     &     arguments, so, that, I, need, to, break, 
     &     it, up )

(I used to use an & because it is a character that meant nothing in Fortran)

[ali_p_]

now all #6633 are gone and they were alot 

now im having just this one :error #6455: This structure name has not been declared. [XYCOORD]    

record /xycoord/ wxy    
          open(1,file='em2db.dat')

[Arjen_Markus]

Just a guess, but I think the structure is called WXYCOORD - at least that is what it is named in the file "fgraph.fd" you showed.

[ali_p_]

the first thing i did is to add a "w" cuz it appears in other parts of the code without a problem 

but nothing happens the same error stays with chang [WXYCOORD]

[Arjen_Markus]

This is confusing - I found the file "fgraph.fd" in my installation of Intel. It actually uses the module ifqwin (so that the exact interfaces are known). The structure is indeed called XYCOORD and in the code you posted first I see "WXYCOORD".

So, from the file "fgraph.fd" (or rather "ifqwin.f90", which defines the module) I conclude it should indeed be "XYCOORD".

Is the error coming from the routine that include "fgraph.fd" or not? That makes a difference. At this point you will have to check the error messages and their relation to the code carefully - we can see only a small part (what you post) and have to guess what is going on.

[ali_p_]

it's not in the routine that contain the include fgraph.fd 

but i solved it.

now i have 2 linking errors 

   desecription                                                                                                                                                               file

Error    1     error LNK2019: unresolved external symbol SETTEXTPOSITION referenced in function MAIN__    EM2DB.obj    
Error    2     fatal error LNK1120: 1 unresolved externals                                                                                        EM2DB.exe    

[mecej4]

The (nonstandard) graphics interfaces have been adopted and adapted since the days of Microsoft Fortran (16-bit). Older versions tend to use 16-bit integer arguments, and later versions use 32-bit integer arguments. You should be very careful that you do not give the compiler incorrect interfaces, which is what could happen if you retrieve a seemingly compatible fgraph.fd from some internet site. Using an incorrect interface could change compile time errors into run-time errors, which are more harmful as well as more difficult to isolate and fix.

All the interfaces are available in files provided with your Intel Fortran installation. Do not corrupt your installation with files from other sources or files from an older version of the compiler. Remember the Zeroth Law of Computing: "Don't make a mess!".