Fatal error executing: awk -f /opt/GCBASIC/preprocess.awk program1.2.3.gcb

[bed]

A special problem, maybe I am the only one who concern about it :-)
Symptom: Source can not compiled:
Fatal error executing:
echo ' awk -f /opt/GCBASIC/preprocess.awk r1.1.gcb'
awk -f /opt/GCBASIC/preprocess.awk r1.1.gcb
exit

After digging around a while I found the reason.
The reason is that my source code has several dots "." in the file name.
I noticed by chance.
The error comes from the fact that makehex.sh can only handle one dot in the file name.
I have adapted makehex accordingly and ask other Linux and MAC (?) users to try it.

in makehex.sh begin with line 22

  # process any header files newer than current HEX file
  # ! There can be only one dot in the filename, or this will fail
  # out_name=$(basename "$1" | cut -d. -f1).hex # Note HEX suffix
  # this should work with more than 1 dots also
  out_name=$(basename "$1"|rev|cut -d. -f2-|rev).hex # Note HEX suffix

[mkstevo]

Hello.

At some point, I had noticed this. A kind GCB member who understands the LINUX command line far better than I do, was helping me out in writing a script that extracted the #Chip directive in source code so that it could be passed to the NSDSP programmer when programming the device. When writing this script I came across this very problem and managed to remedy it, with a similar solution to yours.

# ! There can be only one dot in the filename, or this will fail
# hex_file=$(basename "${1}" | cut -d. -f1).hex
# Changed as I sometimes have more than one dot in the filename

hex_file=$(basename "$1" | rev | cut -d"." -f2-  | rev).hex

# ! There can be only one dot in the filename, or this will fail
# html_file=$(basename "${1}" | cut -d. -f1).html
# Changed as I sometimes have more than one dot in the filename

html_file=$(basename "$1" | rev | cut -d"." -f2-  | rev).html

As for various reasons the 'awk' processes failed to work on the Mac, I have that section commented out any way.

[Anobium]

An insight. The AWK is there to inspect for errors that the compiler did not/was not picking up and to ensure the libraries where valid.

I will look to remove the AWK processing at the 0.98.08. I think that is possible.

So, if you do not execute the AWK, today, what is the impact? You may get strange error messages relating to the following (as the AWK preprocessing issues proper messages):

• Unmatched Braces
• If - Else errors
• Casting errors
• Loops that are not terminated errors
• Subs and Function that are not terminated errors

All these errors are masked today. If someone wants to help move these checks into the compiler... please let us know.

Evan

[mkstevo]

That's a question there...

So, if you do not execute the AWK, today, what is the impact? You may get strange error messages relating to the following (as the AWK preprocessing issues proper messages):

I am sure that in more recent macOS versions, AWK doesn't run correctly? Which is why I initially stopped using it...

Unmatched Braces

Tested, I get the error: CS_V2_11.gcb (956): Error: Brackets do not match up

If - Else errors

Tested, I get the error: CS_V2_11.gcb (972): Error: Else outside of If ... End If block
Or: CS_V2_11.gcb (953): Error: If without matching End If

Casting errors

Untested

Loops that are not terminated errors

Tested, I get the error: CS_V2_11.gcb (1397): Error: Do without matching Loop
Or: CS_V2_11.gcb (1420): Error: Loop without matching Do

Subs and Function that are not terminated error

Tested, I get the error: None, no error raised for Un-terminated Sub [Though looking at the .ASM it appears as though the sub did terminate normally?]

When compiling this Sub with the "End Sub" commented out:

Sub PrintAnimA
    Let D_Lat  = 0                'Start output data latch
    ShiftData(254) '-
    ShiftData(255) '
    ShiftData(255) '
    ShiftData(247) '-
    Let D_Lat  = 1                'End output data latch
'End Sub

The .ASM looks like this:

PRINTANIMA
;Let D_Lat  = 0                'Start output data latch
    banksel LATC
    bcf LATC,4
;ShiftData(254) '-
    movlw   254
    banksel DATAOUT
    movwf   DATAOUT
    call    SHIFTDATA
;ShiftData(255) '
    movlw   255
    movwf   DATAOUT
    call    SHIFTDATA
;ShiftData(255) '
    movlw   255
    movwf   DATAOUT
    call    SHIFTDATA
;ShiftData(247) '-
    movlw   247
    movwf   DATAOUT
    call    SHIFTDATA
;Let D_Lat  = 1                'End output data latch
    banksel LATC
    bsf LATC,4
;End Sub
    banksel STATUS
    return

Remove the commented out "'End Sub" completely and the .ASM looks like this:

PRINTANIMA
;Let D_Lat  = 0                'Start output data latch
    banksel LATC
    bcf LATC,4
;ShiftData(254) '-
    movlw   254
    banksel DATAOUT
    movwf   DATAOUT
    call    SHIFTDATA
;ShiftData(255) '
    movlw   255
    movwf   DATAOUT
    call    SHIFTDATA
;ShiftData(255) '
    movlw   255
    movwf   DATAOUT
    call    SHIFTDATA
;ShiftData(247) '-
    movlw   247
    movwf   DATAOUT
    call    SHIFTDATA
;Let D_Lat  = 1                'End output data latch
    banksel LATC
    bsf LATC,4
    banksel STATUS
    return

And with the "End Sub" included as normal I get:

PRINTANIMA
;Let D_Lat  = 0                'Start output data latch
    banksel LATC
    bcf LATC,4
;ShiftData(254) '-
    movlw   254
    banksel DATAOUT
    movwf   DATAOUT
    call    SHIFTDATA
;ShiftData(255) '
    movlw   255
    movwf   DATAOUT
    call    SHIFTDATA
;ShiftData(255) '
    movlw   255
    movwf   DATAOUT
    call    SHIFTDATA
;ShiftData(247) '-
    movlw   247
    movwf   DATAOUT
    call    SHIFTDATA
;Let D_Lat  = 1                'End output data latch
    banksel LATC
    bsf LATC,4
    banksel STATUS
    return

The only "Odd" errors I have had have been with my own "included" files. If I have made an error within one of my header files, the program fails to compile, throwing an unknown error:
An error has been found: [With no further information shown]

As I now know that this indicates an error in an "included" file, I know approximately where to look for this one.

[mkstevo]

Just spent a few hours returning the preprocessor options to the script. I did get the commands working to some degree. I have them reporting errors in the header files, though as I can't persuade the script to stop executing by the time the rest of the compilation has completed they are all but invisible, being way off the compiler output window.

They still seem to not report where any error might be. Although they do if you scroll right to the top of the output window note an error of some sort in a header file. They still fail to make any note of any unterminated sub routines.

I am far from an expert in shell scripting and I probably have not done things correctly, but I am struggling to see what, if anything, I've acheived?

[bed]

In Linux build .. makehex.sh isn't html_file anaway.
What does MAC do with it?

[mkstevo]

The script uses the "filename.html" file to extract the #Chip parameter from it on compile so that it can be passed to the command line of the NSDSP programmer

[Anobium]

When we release 0.98.07 you may want to use the new compiler switch that outputs only the chipname and the config.

Evan

[mkstevo]

But does that work in the macOS version, and it needs to be processed in a way that gives the result similar to:
PIC16F1825

From the #Chip 16F1829, 32 directive.

This value then needs passing to the NSDSP programmer command line.

[bed]

Ohh, My posted script has a typo $1 not $2 , sorry.
My Script is ok. Wonder why here stands $1, I am a messy.
(corrected in Post above)