drygdryg/calc
1
0
Fork 0
mirror of https://github.com/lcn2/calc.git synced 2025-08-16 01:03:29 +03:00
Code Issues Packages Projects Releases Wiki Activity

improve calc error code system

Browse Source 
Changed calc_errno a global int variable so that is may be directly
accessed by libcalc users.

Further improve help files for help/errno, help/error, help/newerror,
help/stoponerror and help/strerror by adding to documentation
of the calc error code system as well as libcalc interface
where applicable.
This commit is contained in:
Landon Curt Noll
2023-09-12 16:06:27 -07:00
parent 19819340ff
commit 229a60e4d5
8 changed files with 162 additions and 80 deletions
Download Patch File Download Diff File Expand all files Collapse all files

50
help/errno
View File

@@ -5,19 +5,35 @@ SYNOPSIS
errno([errnum])
TYPES
errnum integer, 0 <= errnum <= 32767
errnum integer
return integer
DESCRIPTION
Whenever an operation or evaluation of function returns an error-value,
the numerical code for that value is stored as calc_errno.
the numerical code for that value is stored as "global calc_errno value".
errno() returns the current value of calc_errno.
The term "global calc_errno value" is an internal calc state.
This is NOT a calc variable called calc_errno.
errno(errnum) sets calc_errno to the value errnum and returns its
previous value.
NOTE:
The errno() builtin should not be confused with the errno
used by libc in C. The range of errnum extend beyond the
C's libc errno. See help error for information on the mean
of various errnum ranges.
Unlike error() calling errno() does NOT increment the global
calc error count (see help errcount).
errno() returns the current value of "global calc_errno value".
Unlike error() calling errno() does NOT increment the global
calc error count (see help errcount).
errno(errnum) sets "global calc_errno value" to the value errnum
and returns its previous value. Unlike error(errnum) calling
errno(errnum) does NOT increment the global calc error count
(see help errcount).
To detect whether an error occurs during some sequence of operations,
one may immediately before that sequence set the stored error-number
@@ -34,29 +50,29 @@ DESCRIPTION
EXAMPLE
; ## Assuming there is no file with name "not_a_file"
; errno(0)
0
; errmax(errcount()+4)
20
; errcount(0),; errmax(20),;
; print errcount(), errmax()
0 20
; badfile = fopen("not_a_file", "r")
; print errno(), error(), strerror()
2 System error 2 No such file or directory
; print errcount(), errno(), strerror()
1 2 No such file or directory
; a = 1/0
; print errno(), error(), strerror()
10001 Error 10001 Division by zero
; print errcount(), errno(), strerror()
2 10001 Division by zero
LIMITS
none
0 <= errnum <= 32767
LINK LIBRARY
none
int calc_errno; /* global calc_errno value */
SEE ALSO
errmax, errcount, error, strerror, iserror, newerror, errorcodes,
stoponerror
## Copyright (C) 1999-2006,2021 Landon Curt Noll
## Copyright (C) 1999-2006,2021,2023 Landon Curt Noll
##
## Calc is open software; you can redistribute it and/or modify it under
## the terms of the version 2.1 of the GNU Lesser General Public License

68
help/error
View File

@@ -2,72 +2,82 @@ NAME
error - generate a value of specified error type
SYNOPSIS
error([n])
error([errnum])
TYPES
n integer, defaults to errno()
errnum integer, defaults to errno()
return null value or error value
DESCRIPTION
If n == 0:
If errnum == 0:
error(n) returns the null value.
error(errnum) returns the null value.
The "global calc_errno value" is set to 0.
The global calc error count is not changed.
For 0 < n <= 32767:
For 0 < errnum <= 32767:
error(n) returns a value that is of a special type called
error(errnum) returns a value that is of a special type called
"error". The "error" return value, when given to iserror()
will cause that function to return a true value that happens
to match the "global calc_errno value".
The "global calc_errno value" is then set to n.
The "global calc_errno value" is then set to errnum.
The global calc error count (see help errcount) is incremented.
If the global calc error count exceeds the maximum count
(see help errmax), any ongoing calc execution is aborted.
All other values of n are reserved for future use and currently
All other values of errnum are reserved for future use and currently
will generate an error.
NOTE:
Calling error() without an argument will return the current
"global calc_errno value" AND will also increment the
global calc error count.
Use errno() instead of error() to return the "global calc_errno
value" without incrementing the global calc error count.
The term "global calc_errno value" is an internal calc state.
This is NOT a calc variable called calc_errno.
This is NOT a calc variable called calc_errno. The value of the
"global calc_errno value" may be returned by errno() (see help errno).
The "global calc_errno value" be set to some other non-zero
value by other builtin functions and calculations, especially
if they result in an error. Thus, while calling errno(n) may
if they result in an error. Thus, while calling errno(errnum) may
set calc_errno, some other error may later overwrite the
"global calc_errno value".
By convention, the "global calc_errno value" has the following ranges:
n < 0 Reserved for future use
errnum < 0 Reserved for future use
n == 0 calc_errno cleared: libc errno codes above here
errnum == 0 calc_errno cleared: libc errno codes above here
1 <= n <= 9999 system error and libc errno codes
1 <= errnum <= 9999 system error and libc errno codes
n == 10000 Reserved for "No error" calc internal state
errnum == 10000 Reserved for "No error" calc internal state
10001 <= n <= E__HIGHEST calc computation error codes
10001 <= errnum <= E__HIGHEST calc computation error codes
E__HIGHEST < n < 20000 Reserved for future calc error codes
E__HIGHEST < errnum < 20000 Reserved for future calc error codes
20000 < n <= 32767 User defined error codes start here
20000 < errnum <= 32767 User defined error codes start here
n >= 32768 Reserved for future use
errnum >= 32768 Reserved for future use
The following constants are part of the calc error code mechanism:
E__BASE == 10000
E_USERDEF == 20000
E_USERMAX == 32767
E__BASE < E__HIGHEST < E_USERDEF
E__COUNT == E__HIGHEST - E__BASE
E__BASE == 10000 # 1 less than the 1st calc computation error code
E_USERDEF == 20000 # User defined error codes start
E_USERMAX == 32767 # maximum user defined error code
E__BASE < E__HIGHEST < E_USERDEF # How E__HIGHEST is bounded
E__COUNT == E__HIGHEST - E__BASE # Number of calc computation error codes
See help errorcodes for the actual values of the above constants.
@@ -84,8 +94,10 @@ EXAMPLE
Error 10009
; print iserror(a)
10009
; strerror(a)
"Bad argument for inverse"
; print strerror(a)
"Bad argument for inverse"
; print errno()
10009
; b = error(0)
@@ -93,12 +105,14 @@ EXAMPLE
1 20
; print isnull(b)
1
; print errno()
0
LIMITS
0 <= n <= 32767
0 <= errnum <= 32767
LINK LIBRARY
none
int calc_errno; /* global calc_errno value */
SEE ALSO
errcount, errmax, errorcodes, iserror, errno, strerror, newerror,

10
help/newerror
View File

@@ -15,6 +15,11 @@ DESCRIPTION
that any future use of newerror(str) with the same str will return
the same error-value.
When this function is called, the global calc error count (see
help errcount) is incremented. If the global calc error count
exceeds the maximum count (see help errmax), any ongoing calc
execution is aborted.
If x = newerror(str), both strerror(x) and strerror(iserro(x)) will
return str and iserror(x) will return the error code value of the
new error.
@@ -61,7 +66,8 @@ EXAMPLE
triangle side length <= 0: 20000
LIMITS
The number of new described error-values is not to exceed 12767.
The number of new described error-values is not to exceed E_USERMAX-E_USERDEF
(which is usually 12767 calls to the newerror function).
LINK LIBRARY
none
@@ -70,7 +76,7 @@ SEE ALSO
errmax, errcount, error, strerror, iserror, errno, errorcodes,
stoponerror
## Copyright (C) 1999-2006 Landon Curt Noll
## Copyright (C) 1999-2006,2023 Landon Curt Noll
##
## Calc is open software; you can redistribute it and/or modify it under
## the terms of the version 2.1 of the GNU Lesser General Public License

5
help/stoponerror
View File

@@ -22,6 +22,9 @@ DESCRIPTION
stoponerror value. When 1 argument is given, stoponerror() returns
the previous stoponerror value.
See help errcount for information on the current error count.
See help errmax for the maximum error-count before execution stops.
EXAMPLE
; stoponerror()
0
@@ -35,7 +38,7 @@ LINK LIBRARY
SEE ALSO
errcount, errmax, errorcodes, iserror, errno, strerror, newerror
## Copyright (C) 2006,2021 Landon Curt Noll
## Copyright (C) 2006,2021,2023 Landon Curt Noll
##
## Calc is open software; you can redistribute it and/or modify it under
## the terms of the version 2.1 of the GNU Lesser General Public License

90
help/strerror
View File

@@ -2,51 +2,85 @@ NAME
strerror - returns a string describing an error value
SYNOPSIS
strerror([x])
strerror([errnum])
TYPES
x error-value or integer in [0, 32767], defaults to errno()
errnum integer
return string
DESCRIPTION
If x is the error-value with index n, strerror(x) and strerror(n)
return one of:
If 0 <= errnum <= 32767 then strerror(errnum) will return one of:
a system-generated message,
a calc-generated description,
a user-defined description created by newerror(str),
the string "Error n",
* system error and libc errno value related string
* calc computation error code string
* user-defined description string created by newerror(str)
* "Unknown error errnum" (where errnum is in decimal)
where, in the last form, n is represented decimally.
Without a argument, strerror() returns as if it were called with:
strerror(errno())
That is, the string associated with the "global calc_errno value"
is returned.
The term "global calc_errno value" is an internal calc state.
This is NOT a calc variable called calc_errno.
The system error and libc errno value related string may be
different for different systems.
For users of libcalc:
The libcalc array error_table[] contains the hard coded
calc-generated error array of E__COUNT+2 pointers to strings
starting with E__BASE, ending with E__HIGHEST and terminated
by a NULL pointer.
The error_table[0] is for errnum of E__BASE, and is the
string "No error".
The error_table[E__COUNT] is string associated with the
highest calc-generated error message.
The error_table[E__COUNT+1] is a NULL pointer.
See help error for more information on the ranges of errnum and
the meaning of the above mentioned constants.
EXAMPLE
System error messages may be different for different systems.
; errmax(errcount()+3)
0
; strerror(2)
"No such file or directory"
; errcount(0),; errmax(20),;
; print errcount(), errmax()
0 20
; x = 3 * ("a" + "b")
; print strerror(x)
Bad arguments for +
; print strerror(2)
No such file or directory
; print errcount(), errmax()
0 20
; a = newerror("alpha")
; print strerror(a)
alpha
; print strerror(999)
Error 999
; a = 1/0
; x = 3 / 0;
; print errcount(), errmax()
1 20
; print strerror()
Division by zero
; a = newerror("alpha")
; print errcount(), errmax()
2 20
; print strerror()
alpha
; print strerror(999)
Unknown error: 999
; print errcount(), errmax()
2 20
LIMITS
none
0 <= errnum <= 32767
LINK LIBRARY
none
int calc_errno; /* global calc_errno value */
CONST char *error_table[E__COUNT+2]; /* calc coded error messages */
SEE ALSO
strcat, strcpy, strlen, strncmp, strncpy, strpos,
@@ -55,7 +89,7 @@ SEE ALSO
errcount, errmax, error, iserror, errno, newerror, errorcodes,
stoponerror
## Copyright (C) 1999-2006 Landon Curt Noll
## Copyright (C) 1999-2006,2023 Landon Curt Noll
##
## Calc is open software; you can redistribute it and/or modify it under
## the terms of the version 2.1 of the GNU Lesser General Public License