concat weirdness

[rikky]

hello

following errors in both the latest stable and the latest fossil.sh


help$ = NL$ & \
 "TOKEN [option] haystack$ n [delimiter$]" & NL$ & \
 NL$ & \
 "haystack$ can be replaced with piped input." & NL$ & \
 NL$ & \
 "Returns the nth token in haystack$ split by delimiter$. The delimiter$ is " & NL$ & \
 "optional. If it is omitted, then the definition from OPTION DELIM is assumed." & NL$ & \
 "When specified, it may consist of multiple characters. If delimiter$ occurs " & NL$ & \
 "between double quotes in haystack$, then it is ignored. This behavior can be" & NL$ & \
 "changed by setting OPTION QUOTED to FALSE. Use the OPTION COLLAPSE TRUE to " & NL$ & \
 "collapse multiple delimiters into one. If the indicated position is outside" & NL$ & \
 "a valid range, the result will be an empty string. Use the FLATTEN$ function to " & NL$ & \
 "flatten out the returned token. See also ISTOKEN, AMOUNT and SPLIT. " & NL$ & \
 NL$  & \
 "options :" & NL$ & \
 " -v --version     print version." & NL$ & \
 " -h --help        print help." & NL$ & \
 NL$ & \
 "example:" & NL$ & \
 NL$ & \
 "TOKEN 'a b c d " & CHR$(34) & "e f" & CHR$(34) & " g h i j' 6" & NL$ & \
 "cat file1 | TOKEN 4 $'\n'" & NL$ & \
 NL$ & \
 "export QUOTED=FALSE" & NL$ & \
 "TOKEN 'a b c d " & CHR$(34) & "e f" & CHR$(34) & " g h i j' 6"  & NL$ & \
 "FLATTEN $(TOKEN 'a b c d " & CHR$(34) & "e f" & CHR$(34) & " g h i j' 6)" & NL$ & \
 ""  
result:

Converting 'Playroom/test.bac'... done, 34 lines were processed in 21 seconds.
Creating lexical analyzer... done.
Compiling 'Playroom/test.bac'... cc  -c test.bac.c
Makefile.bacon:6: recipe for target 'test.bac.o' failed

test.bac:27:1889: error: expected ')' before ';' token

INFO: see full error report (y/[n])? y
test.bac: In function 'main':
test.bac:27:1889: error: expected ')' before ';' token
test.bac:29:1: error: expected ';' before '}' token
make: *** [test.bac.o] Error 1

The error lies somewhere in the lines:

 "export QUOTED=FALSE" & NL$ & \
 "TOKEN 'a b c d " & CHR$(34) & "e f" & CHR$(34) & " g h i j' 6"  & NL$ & \
For if I replace these texts with empty strings, it compiles well.
So following compiles well:

help$ = NL$ & \
 "TOKEN [option] haystack$ n [delimiter$]" & NL$ & \
 NL$ & \
 "haystack$ can be replaced with piped input." & NL$ & \
 NL$ & \
 "Returns the nth token in haystack$ split by delimiter$. The delimiter$ is " & NL$ & \
 "optional. If it is omitted, then the definition from OPTION DELIM is assumed." & NL$ & \
 "When specified, it may consist of multiple characters. If delimiter$ occurs " & NL$ & \
 "between double quotes in haystack$, then it is ignored. This behavior can be" & NL$ & \
 "changed by setting OPTION QUOTED to FALSE. Use the OPTION COLLAPSE TRUE to " & NL$ & \
 "collapse multiple delimiters into one. If the indicated position is outside" & NL$ & \
 "a valid range, the result will be an empty string. Use the FLATTEN$ function to " & NL$ & \
 "flatten out the returned token. See also ISTOKEN, AMOUNT and SPLIT. " & NL$ & \
 NL$  & \
 "options :" & NL$ & \
 " -v --version     print version." & NL$ & \
 " -h --help        print help." & NL$ & \
 NL$ & \
 "example:" & NL$ & \
 NL$ & \
 "TOKEN 'a b c d " & CHR$(34) & "e f" & CHR$(34) & " g h i j' 6" & NL$ & \
 "cat file1 | TOKEN 4 $'\n'" & NL$ & \
 NL$ & \
 "" & NL$ & \
 ""  & NL$ & \
 "FLATTEN $(TOKEN 'a b c d " & CHR$(34) & "e f" & CHR$(34) & " g h i j' 6)" & NL$ & \
 ""
Putting back either one of these lines errors.

I tried but failed to find the cause.

 Rik

PS. It did work, for I only did a recompile after a minor improvement somewhere completely else.

[Pjot]

Hi rikky,

Your string contains a special character which is interpreted by the C compiler during compilation. I am talking about the "\n" in the part:

"cat file1 | TOKEN 4 $'\n'" & NL$ &


You have to escape the '\' here. So your code should look like this:

help$ = NL$ & \
 "TOKEN [option] haystack$ n [delimiter$]" & NL$ & \
 NL$ & \
 "haystack$ can be replaced with piped input." & NL$ & \
 NL$ & \
 "Returns the nth token in haystack$ split by delimiter$. The delimiter$ is " & NL$ & \
 "optional. If it is omitted, then the definition from OPTION DELIM is assumed." & NL$ & \
 "When specified, it may consist of multiple characters. If delimiter$ occurs " & NL$ & \
 "between double quotes in haystack$, then it is ignored. This behavior can be" & NL$ & \
 "changed by setting OPTION QUOTED to FALSE. Use the OPTION COLLAPSE TRUE to " & NL$ & \
 "collapse multiple delimiters into one. If the indicated position is outside" & NL$ & \
 "a valid range, the result will be an empty string. Use the FLATTEN$ function to " & NL$ & \
 "flatten out the returned token. See also ISTOKEN, AMOUNT and SPLIT. " & NL$ & \
 NL$  & \
 "options :" & NL$ & \
 " -v --version     print version." & NL$ & \
 " -h --help        print help." & NL$ & \
 NL$ & \
 "example:" & NL$ & \
 NL$ & \
 "TOKEN 'a b c d " & CHR$(34) & "e f" & CHR$(34) & " g h i j' 6" & NL$ & \
 "cat file1 | TOKEN 4 $'\\n'" & NL$ & \   <-------------------------------------- added '\' to '\n'
 NL$ & \
 "export QUOTED=FALSE" & NL$ & \
 "TOKEN 'a b c d " & CHR$(34) & "e f" & CHR$(34) & " g h i j' 6"  & NL$ & \
 "FLATTEN $(TOKEN 'a b c d " & CHR$(34) & "e f" & CHR$(34) & " g h i j' 6)" & NL$ & \
 ""  


Now it works for me.

HTH
Peter

[rikky]

Ahj, I see.

Well then I probably did not check if it really did compile, the last time that I made this change in the help$

Thanks

 Rik