Messages¶
Every message the tools print, in English, as the code writes it, with
<n> for a number, <name> for a name or a text and <x> for a
value. What the common ones mean and how to get past them is in
Messages.
cmmcomp (C±)¶
Errors¶
Syntax error on line
. You’re a confused soul! Error: yo, where’s the main() function?
Error: couldn’t open the file ‘
’ for the life of me!! Error on line
: the symbol ‘i’ is reserved for the imaginary part of a complex constant. Hands off! Error on line
: that name is way too long, characters tops. Error on line
: #NUGAIN is not a power of two. norm(x) is x / NUGAIN in hardware: a power of two becomes a shift; anything else becomes a divider, the critical path of the whole ALU. Error on line
: #NUBITS is out of range. A SAPHO word has 2 to 32 bits. Error on line
: #FFTSIZ is out of range. v[k) reverses 1 to 30 low bits of the index. Error on line
: #FROUND only takes 0 (legacy), 1 (truncation + saturation) or 2 (round to nearest even). Error on line
: the biggest integer this thing can hold is ! Error on line
: ‘ ’ starts with 0, so it is octal, and octal has no 8 or 9. Error on line
: the smallest float this thing can hold is ! Error on line
: the biggest float this thing can hold is ! Error on line
: the variable ‘ ’ already exists. Go grab an Ω³! Error on line
: is ‘ ’ in the assembly, and that name is already . Rename one of them. Error on line
: is ‘ ’ in the assembly, which is an instruction of the processor. Rename it. Error on line
: : ‘ ’ is an instruction of the processor. Rename it. Error on line
: a function can’t be called ‘ ’: . Error on line
: ‘ ’: names starting with ‘__’ belong to the compiler. Error on line
: I’d really appreciate it if you declared the variable ‘ ’. Error on line
: c’mon dude, declare the variable ‘ ’ properly! Error on line
: gotta declare ‘ ’ first! Error on line
: where’s the array index for the variable ‘ ’? Error on line
: ‘ ’ is an array. Where’s the index, dude? Error on line
: ‘ ’ isn’t an array. Error on line
: ‘ ’ ain’t an array, mate! Error: the source scanner stopped:
Error on line
: an expression nested too deep for the compiler. Split it with intermediate variables. Error on line
: case appears twice in this switch. Error on line
: this switch already has a default. Error on line
: the character ‘ ’ is not part of C±. Error on line
: the index of ‘ ’ is out of range, 0 to . Error on line
: the array ‘ ’ has elements; it can have 1 to . Error on line
: array ‘ ’ has two dimensions! Error on line
: array ‘ ’ has just one dimension! Error on line
: what were you smoking when you tried to increment a complex number? Error on line
: that’s not how you use ‘ ’! Error on line
: there’s no input port , dude! Error on line
: there’s no output port , dude! Error on line
: #PRNAME is gone. The processor takes the name of the file (filtro.cmm becomes the processor filtro): delete the line and, for another name, rename the file. Error on line
: the link number is an integer from 0 to , and I found . Error on line
: link was already used _sync. A link has one hardware: either the partner is on the same clock (toma_sync/cade_sync/fcade_sync in every use) or it is not (toma/cade/fcade in every use). Error on line
: cade() brings one word, and ‘ ’ is complex (two words). Take the real and the imaginary parts separately. Error on line
: first pick which info you want from this complex number! Error on line
: using sign(.,.) with complex numbers makes zero sense! Error on line
: using the ‘pset(.)’ function with complex numbers makes zero sense! Error on line
: nope! norm(.) is integer-only! Error on line
: (.) doesn’t take a comp. Error on line
: a complex number raised to a power? Dream on! Didn’t build that yet, you’re on your own! Error on line
: argument of real(.) has gotta be complex! Error on line
: argument of imag(.) has gotta be complex! Error on line
: argument of mod2(.) has gotta be complex! Error on line
: argument of fase(.) has gotta be complex! Error on line
: arguments of complex(.,.) can’t themselves be complex! Error on line
: haven’t coded this up for complex numbers yet. You’re on your own! Error on line
: ‘ ’ ain’t even a vector, mate! Error on line
: ‘ ’ ain’t even a matrix, mate! Error on line
: you can only shift a vector into itself, mate! Error on line
: adding and ? Those shapes don’t fit, mate! Error on line
: times ? No way: the columns of the first must match the rows of the second! Error on line
: ‘ ’ is and the expression gives . That won’t fit, pal! Error on line
: I wanted a number here and you gave me ! Error on line
: out() sends a number or a vector, not . Slice it up first! Error on line
: this Dirac combo is still in the oven. Hang tight! Error on line
: complex numbers have no order, so <, >, <= and >= don’t work with a comp. Compare mod2(z), real(z) or imag(z) instead. Error on line
: what’s the point of computing a remainder with non-integer numbers? Get a grip! Error on line
: logical operation with a comp (int and float only)! Error on line
: bad use of the ‘~’ operator. Gotta pass an int. You tripping? Error on line
: how am I supposed to do bitwise ops with a complex number? You tripping? Error on line
: how am I supposed to shift bits of a complex number? You tripping? Error on line
: using comp to shift bits? You’re a confused soul! Error on line
: array ‘ ’ holds and does not take the value between braces. Error on line
: array ‘ ’ has positions and the list between braces reaches position . Error on line
: array ‘ ’ has one dimension; braces inside the list only mark the rows of a 2D array. Error on line
: in array ‘ ’, inner braces mark the rows of a 2D array: every row between braces, or none, each with up to values. Error on line
: a return value in a void function? You tripping! Error in function
: some path gets to the end without a return, and then it’d run straight into the next function. Put a return at the end. Error on line
: where’s the function return value? Error on line
: where’s this function ‘ ’ supposed to be? Error on line
: function ‘ ’ is where exactly? Error on line
: check again how many parameters the function ‘ ’ actually takes. Error on line
: take a look at function ‘ ’, you’ll see it doesn’t return anything. Error on line
: parameter list of function ‘ ’ doesn’t match the original. Error on line
: that break is just hanging out lost! Error on line
: a continue outside a loop goes nowhere! Error on line
: function ‘ ’ calls itself ( ). Recursion doesn’t fly in C±. Really want recursion? Use SAPHO’s C++ compiler. Error on line
: there’s already an interrupt going on somewhere else before this one! Error on line
: duplicate #TOAQUI marker (only one allowed) Error on line
: #WAVE takes a variable, and ‘ ’ isn’t one Error: option ‘
’ needs a value. Error: unknown option ‘
’. Error: the file ‘
’ names the processor, and ‘ ’ . Rename the file. Error: where’s the
folder? It has to sit next to cmmcomp’s bin/ folder. Install YANC again. Error: missing required option ‘
’.
Warnings¶
Heads up on line
: the array has positions, but an int index of bits only reaches . Heads up on line
: if you don’t want this warning, use ‘fout’. Heads up on line
: if you don’t want this warning, use ‘out’. Warning on line
: conj of a real number? It’s just itself with a zero imaginary part, but ok! Heads up on line
: the second operand of the shift is a float. Now you’re killin’ me! Heads up on line
: variable ‘ ’ is int, but getting a float. Heads up on line
: local array ‘ ’ gets the values between braces only once, when the memory is loaded; from one call to the next it keeps what the function wrote into it (like a C static). Heads up on line
: local array ‘ ’ is initialized from file only once, when the memory is loaded; from one call to the next it keeps what the function wrote into it (like a C static). Heads up on line
: variable ‘ ’ is float, but getting an int. Heads up on line
: variable ‘ ’ is comp, but getting an int. Heads up on line
: variable ‘ ’ is comp, but getting a float. Heads up on line
: in this conversion I’m truncating the real part, just so you know! Heads up on line
: in this conversion I’m only grabbing the real part, just so you know! Heads up on line
: converting float to int on parameter of function ‘ ’. Heads up on line
: converting comp to int on parameter of function ‘ ’. Heads up on line
: converting int to float on parameter of function ‘ ’. Heads up on line
: converting comp to float on parameter of function ‘ ’. Heads up on line
: converting int to comp on parameter of function ‘ ’. Heads up on line
: converting float to comp on parameter of function ‘ ’. Heads up on line
: gonna convert float to int on the return of function ‘ ’? Code away then! Heads up on line
: return is float, but getting an int. Heads up on line
: function return is comp, but getting an int. Heads up on line
: function return is comp, but getting a float. Heads up on line
: case index came out float! Gonna truncate it. Heads up on line
: case index came out comp! Gonna truncate the real part. Heads up on line
: see how the array index came out a float? Gonna truncate it! Heads up on line
: see how the first array index came out a float? Gonna truncate it! Heads up on line
: see how the second array index came out a float? Gonna truncate it! Heads up on line
: see how the array indices came out floats? Gonna truncate ‘em! Heads up on line
: a complex array index? Seriously?! Gonna truncate the real part. Heads up on line
: a complex array index? Seriously?! Gonna grab the real part. Heads up on line
: these array indices are all over the place. You’re a confused soul! Heads up on line
: array index is coming out float. Gonna generate a ton of code to truncate it! Heads up: global variable ‘
’ is never read, so it’s out of the program and the waveform (#WAVE keeps it). Heads up: variable ‘
’ in function ‘ ’ is never read, so it’s out of the program and the waveform (#WAVE , in the function, keeps it). Heads up: function ‘
’ isn’t being used. Save some memory, will ya! Heads up on line
: constant is below the smallest normal float ( ) and becomes 0 at #FROUND . Use #FROUND 0 or raise #NBEXPO.
Information¶
Info on line
: ‘ ’ is read at one spot and written at another, so the result takes a detour through the temporary area ( words). Info on line
: the product in the middle of the chain goes to the temporary area ( words), or it’d be computed all over again for every element. Info: the Dirac temporary area (__dirac_tmp) takes
words of data memory. To save them: write the result to an array that isn’t on the right-hand side, or compute the middle product into an array of your own, on a line before. Who used it: Info on line
: #TOAQUI marker registered (drives the cheguei pin) Info: interruption directive found at line
Info:
assembly instructions generated Info: constant
on line approximated to (error = ) Info: adding assembly macro for root square computation
Info: adding assembly macro for arc-tangent computation
Info: adding assembly macro for sin computation
Info: adding assembly macro for tangent computation
Info: adding assembly macro for exponential computation
Info: adding assembly macro for natural-logarithm computation
Info: array initialization with file
for variable ‘ ’ at line Info: Dirac notation for shift register in vector
detected at line
Other¶
Internal error of cmmcomp (
), on line . Not your program’s fault: tell the maintainers and send the program along. Sweet: it compiled! Now you just gotta figure out why it doesn’t work.
cppcomp (C++)¶
Errors¶
error:
: ‘ ’ is an instruction of the processor’s assembly, rename it error:
is ‘ ’ in the assembly, which is an instruction of the processor: rename it error:
: ‘global’ names the global variables in the simulation log, rename the function error:
: ‘ ’ is reserved, it is ; rename it error:
and are both ‘ ’ in the assembly: rename one of them error: _Generic: no association matches the controlling type
error: undefined ‘
’ error: subscript on non-array/non-pointer
error: undefined function ‘
’ error: deref on non-pointer
error: ‘.’ on non-struct
error: no such field ‘
’ error: ‘->’ on non-pointer-to-struct
error: expression is not an lvalue
error: no field ‘
’ error: bad -> target
error:
(): the link number comes first and is an integer literal from 0 to error: link
was already used _sync: a link has one hardware, the partner is on my clock (the _sync forms in every use) or it is not (toma/cade/fcade in every use) error: cade() brings one word: it can only go into a one-word variable
error: assignment to const ‘
’ error: T(args) on non-class type
error: method call on non-class value
error: no method ‘
’ in class ‘ ’ error: in(port) takes 1 arg
error: in() port must be a literal int
error: out(port,val) takes 2 args
error: out() port must be a literal int
error:
(link, value) takes 2 args error: toma() carries one word: this value takes
error:
(link) takes 1 arg, the link number error: malloc(nwords) takes 1 arg
error: free(ptr) takes 1 arg
error: __zerofill takes 2 args
error: field designator in array initializer
error: array designator in struct initializer
error: no field ‘
’ in initializer error: too many initializers
error: array/struct local in a recursive function is not supported
error: break outside of loop/switch
error: continue outside of loop
error: return value in void function
error: case/default outside switch
error: cannot open
error: program has no main() function
error: a lambda is only supported as a global initializer (immediately-invoked, no captures)
error: cannot deduce the lambda’s return type (return a local variable or a literal)
error: range-for over ‘
’: it takes an array or an object with size() and operator[] error: _Static_assert requires a constant integer expression
error: static assertion failed
error: redefinition of global ‘
’ (if this comes from a header, guard it with #ifndef or #pragma once) error: unknown type ‘
’ error: ‘
’ is not a class template error: unknown base class ‘
’ error: member ‘
’ takes one initializer error: array field size must be constant
error: array size must be a constant expression
error: template argument must be a constant
error: typedef cannot have an initializer
error: an array of function pointers takes a { … } initializer
error: ‘
’ redeclared as a function (already a variable) error: redefinition of function ‘
’ error: out-of-class definition for unknown class ‘
’ error: qualified id ‘
:: ’ is not a constructor error: out-of-class ctor for unknown class ‘
’ error: qualified id ‘
::~ ’ is not a destructor error: out-of-class dtor for unknown class ‘
’ error: ‘
’: a name with ‘__’ in it is reserved (C++ keeps them for the compiler, and YANC uses them in the assembly) error: #pragma yanc prname is gone: the processor takes the name of the file (filtro.cpp becomes the processor filtro); delete the line and, for another name, rename the file
error: #pragma yanc fftsiz does nothing in C++ (the bit-reversed addressing is C± only); delete the line
error: #pragma yanc itradd does nothing in C++ (interrupts are C± only); delete the line
error: ‘
’: a number starting with 0 is octal, and has no 8 or 9 error: unexpected character ‘
’ (0x ) error: unknown option ‘
’ error: the file ‘
’ names the processor, and ‘ ’ ; rename the file error: parse failed
error: cannot open ‘
’ for writing error: include too deep
error: cannot open ‘
’ error: can’t find #include “
” error: can’t find #include <
> error: #ifdef nesting too deep
error: #if nesting too deep
error: stray #elif
error: stray #else
error: stray #endif
error: too many -I dirs
error: syntax error
error: memory exhausted
Warnings¶
warning: out() sends an integer word; this float is output as its raw bit pattern (wrong value) – cast to int, e.g. out(port, (int)x)
warning: ‘
’ asks for bits, but a YANC word has : the requested size is ignored and ‘ ’ is a -bit warning: unknown #pragma yanc setting ‘
’
asmcomp¶
Errors¶
Error: out of memory while growing the symbol table!
Error: the program has no instruction, so there is no processor to make.
Error: where the heck is the file
? Error: couldn’t open the file ‘
’ for the life of me!! Error: couldn’t create the file
/app_log.txt. Error: the constant ‘
’ is not a number. Error: the constant
does not fit the processor’s float (NBMANT , NBEXPO ): its exponent is past the maximum. Error: #NUGAIN
does not fit a signed -bit word (at most ): norm(x), which is x / NUGAIN, would divide by a negative number. Use a smaller #NUGAIN. Error: #NUGAIN
is not a power of two. norm(x) is x / NUGAIN in hardware: a power of two becomes a shift (~12 LUT levels); anything else becomes a constant divider (38 to 70 levels, measured), the critical path of the whole ALU. Error: floating-point setup doesn’t add up. You gotta have NUBITS = NBMANT + NBEXPO + 1 (NUBITS
, NBMANT , NBEXPO add up to ). A program without float needs no #NBMANT or #NBEXPO. Error: the exponent is too short for the mantissa (NBMANT
, NBEXPO ). You need NBMANT < 2^(NBEXPO-1): with no hidden bit, 1.0 is 2^(NBMANT-1) x 2^-(NBMANT-1), and that exponent has to fit. Widen NBEXPO or narrow NBMANT. Error: #FFTSIZ
is out of range. v[k) reverses 1 to 30 low bits of the index. Error: #NUBITS
is past 32. A SAPHO word is at most 32 bits. Error:
: the operand of LDI/ILI/STI/ISI is an array name or an integer base address. Error:
takes words: it must be followed by NOP and the read of its result (expected , found ). The compilers write the sequence; in a hand-written .asm, write it out. Error: the operand of
is the link number, an integer from 0 to (found ‘ ’). Error: #LKSYNC
, but the program has no TOM or CAD on link . Error: the program ends inside the sequence of
(NOPs and are missing). Error: #FROUND only takes 0 (legacy), 1 (truncation + saturation) or 2 (round to nearest even).
Error: line
of file ‘ ’ is empty! Kinda hard to work with that. Error: line
of file ‘ ’ isn’t a valid integer! Error: line
of file ‘ ’ busts the upper limit of ! Error: line
of file ‘ ’ is below the lower limit of ! Error: line
of file ‘ ’ isn’t a valid float! Error: line
of file ‘ ’ busts the upper limit of ! Error: line
of file ‘ ’ is below the lower limit of ! Error: line
of file ‘ ’ isn’t a valid comp! Error: real part of line
of file ‘ ’ busts the upper limit of ! Error: real part of line
of file ‘ ’ is below the lower limit of ! Error: imaginary part of line
of file ‘ ’ busts the upper limit of ! Error: imaginary part of line
of file ‘ ’ is below the lower limit of ! Error: you’re
lines short in file ‘ ’! Error: array ‘
’ has positions and initial values! Error: value
of array ‘ ’ ( ) does not fit a -bit word! Error: option ‘
’ needs a value. Error: unknown option ‘
’. Error: the file ‘
’ names the processor, and ‘ ’ . Rename the file. Error: the
folder is not next to asmcomp’s bin/ folder. Install YANC again. Error: missing required option ‘
’. Error: option ‘
’ expects a non-negative integer, got ‘ ’.
Warnings¶
Heads up: there’s
extra lines hanging around in file ‘ ’! Heads up: input port
isn’t being used. Burning hardware for nothing! Heads up: output port
isn’t being used. Burning hardware for nothing!
Information¶
Info: implementing interruption handling
Info:
instructions and variables were found Info: stack depths from the program: SDEPTH
, NDSTAC Info: filling array with
values read from file Info: largest approximation error in file ‘
’ is on line Info: largest approximation error, for real part, in file ‘
’, is on line Info: largest approximation error, for imaginary part, in file ‘
’, is on line Info: largest approximation error in the values of array ‘
’ is on value Info: using
% of the Assembly Instruction Set Info: using
% of the ULA operations Info: adding array handling
Info: adding bit-reverse array index (FFT)
Info: adding data input handling
Info: adding data output handling
Info: adding toma (TOM), which hands a value to the partner processor
Info: adding cade (CAD), which takes a value from the partner processor
Info: adding stack memory for function calls
Info: ULA resource -> int to float converter
Info: ULA resource -> float to int converter
Info: ULA resource -> integer adder
Info: ULA resource -> float-point adder
Info: ULA resource -> integer multiplier
Info: ULA resource -> float-point multiplier
Info: ULA resource -> integer divider (not recomended for high frequency opperation)
Info: ULA resource -> float-point divider (not recomended for high frequency opperation)
Info: ULA resource -> modulo computation (not recomended for high frequency opperation)
Info: ULA resource -> integer sign computation (standard library)
Info: ULA resource -> float-point sign computation (standard library)
Info: ULA resource -> integer negative operation
Info: ULA resource -> float-point negative operation
Info: ULA resource -> integer absolute operation
Info: ULA resource -> float-point absolute operation
Info: ULA resource -> integer pset computation (standard library)
Info: ULA resource -> float-point pset computation (standard library)
Info: ULA resource -> normalization computation (standard library)
Info: ULA resource -> & operator
Info: ULA resource -> | operator
Info: ULA resource -> ^ operator
Info: ULA resource -> ~ operator
Info: ULA resource -> && operator
Info: ULA resource -> || operator
Info: ULA resource -> ! operator
Info: ULA resource -> < operator for int
Info: ULA resource -> < operator for float
Info: ULA resource -> > operator for int
Info: ULA resource -> > operator for float
Info: ULA resource -> == operator
Info: ULA resource -> << operator
Info: ULA resource -> >> operator
Info: ULA resource -> >>> operator
Info: ULA resource -> root-square approximation for float
Info: ULA resource -> float scale by a power of two (2^k)
Info: ULA resource -> base-2 exponent extraction of float
Other¶
Bad format!
Sweet: the processor is good to go.