Tcl Source Code

Check-in [f4efe03192]
Login
Bounty program for improvements to Tcl and certain Tcl packages.
Tcl 2019 Conference, Houston/TX, US, Nov 4-8
Send your abstracts to [email protected]
or submit via the online form by Sep 9.

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

Overview
Comment:TIP 488: Remove tcl_precision
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | tip-488
Files: files | file ages | folders
SHA3-256: f4efe03192ee11b5abcb44249a3905aa14f5723144cefbc7a6163f544e544dd6
User & Date: jan.nijtmans 2017-12-01 11:55:28
Context
2017-12-01
12:24
remove many test-cases related to tcl_precision Closed-Leaf check-in: 421523d6cd user: jan.nijtmans tags: tip-488
11:55
TIP 488: Remove tcl_precision check-in: f4efe03192 user: jan.nijtmans tags: tip-488
11:03
Remove stub entries for Tcl_EvalObj()/Tcl_GlobalEvalObj(). Since those functions are actually macro'... check-in: e247a8da3a user: jan.nijtmans tags: trunk
Changes
Hide Diffs Unified Diffs Ignore Whitespace Patch

Changes to doc/PrintDbl.3.

14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
..
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
.nf
\fB#include <tcl.h>\fR
.sp
\fBTcl_PrintDouble\fR(\fIinterp, value, dst\fR)
.SH ARGUMENTS
.AS Tcl_Interp *interp out
.AP Tcl_Interp *interp in
Before Tcl 8.0, the \fBtcl_precision\fR variable in this interpreter
controlled the conversion.  As of Tcl 8.0, this argument is ignored and
the conversion is controlled by the \fBtcl_precision\fR variable
that is now shared by all interpreters.
.AP double value in
Floating-point value to be converted.
.AP char *dst out
Where to store the string representing \fIvalue\fR.  Must have at
least \fBTCL_DOUBLE_SPACE\fR characters of storage.
.BE
.SH DESCRIPTION
................................................................................
.QW .
or an
.QW e
so that it does not look like an integer.  Where \fB%g\fR would
generate an integer with no decimal point, \fBTcl_PrintDouble\fR adds
.QW .0 .
.PP
If the \fBtcl_precision\fR value is non-zero, the result will have
precisely that many digits of significance.  If the value is zero
(the default), the result will have the fewest digits needed to
represent the number in such a way that \fBTcl_NewDoubleObj\fR
will generate the same number when presented with the given string.
IEEE semantics of rounding to even apply to the conversion.
.SH KEYWORDS
conversion, double-precision, floating-point, string






<
|
<
<







 







<
<
|





14
15
16
17
18
19
20

21


22
23
24
25
26
27
28
..
34
35
36
37
38
39
40


41
42
43
44
45
46
.nf
\fB#include <tcl.h>\fR
.sp
\fBTcl_PrintDouble\fR(\fIinterp, value, dst\fR)
.SH ARGUMENTS
.AS Tcl_Interp *interp out
.AP Tcl_Interp *interp in

This argument is ignored.


.AP double value in
Floating-point value to be converted.
.AP char *dst out
Where to store the string representing \fIvalue\fR.  Must have at
least \fBTCL_DOUBLE_SPACE\fR characters of storage.
.BE
.SH DESCRIPTION
................................................................................
.QW .
or an
.QW e
so that it does not look like an integer.  Where \fB%g\fR would
generate an integer with no decimal point, \fBTcl_PrintDouble\fR adds
.QW .0 .
.PP


The result will have the fewest digits needed to
represent the number in such a way that \fBTcl_NewDoubleObj\fR
will generate the same number when presented with the given string.
IEEE semantics of rounding to even apply to the conversion.
.SH KEYWORDS
conversion, double-precision, floating-point, string

Changes to doc/tclvars.n.

6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
...
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
'\" of this file, and for a DISCLAIMER OF ALL WARRANTIES.
'\"
.TH tclvars n 8.0 Tcl "Tcl Built-In Commands"
.so man.macros
.BS
'\" Note:  do not modify the .SH NAME line immediately below!
.SH NAME
argc, argv, argv0, auto_path, env, errorCode, errorInfo, tcl_interactive, tcl_library, tcl_nonwordchars, tcl_patchLevel, tcl_pkgPath, tcl_platform, tcl_precision, tcl_rcFileName, tcl_traceCompile, tcl_traceExec, tcl_wordchars, tcl_version \- Variables used by Tcl
.BE
.SH DESCRIPTION
.PP
The following global variables are created and managed automatically
by the Tcl library.  Except where noted below, these variables should
normally be treated as read-only by application-specific code and by users.
.TP
................................................................................
.TP
\fBwordSize\fR
.
This gives the size of the native-machine word in bytes (strictly, it
is same as the result of evaluating \fIsizeof(long)\fR in C.)
.RE
.TP
\fBtcl_precision\fR
.
This variable controls the number of digits to generate
when converting floating-point values to strings.  It defaults
to 0.  \fIApplications should not change this value;\fR it is
provided for compatibility with legacy code.
.PP
.RS
The default value of 0 is special, meaning that Tcl should
convert numbers using as few digits as possible while still
distinguishing any floating point number from its nearest
neighbours.  It differs from using an arbitrarily high value
for \fItcl_precision\fR in that an inexact number like \fI1.4\fR
will convert as \fI1.4\fR rather than \fI1.3999999999999999\fR
even though the latter is nearer to the exact value of the
binary number.
.RE
.PP
.RS
If \fBtcl_precision\fR is not zero, then when Tcl converts a floating
point number, it creates a decimal representation of at most
\fBtcl_precision\fR significant digits; the result may be shorter if
the shorter result represents the original number exactly. If no
result of at most \fBtcl_precision\fR digits is an exact representation
of the original number, the one that is closest to the original
number is chosen.
If the original number lies precisely between two equally accurate
decimal representations, then the one with an even value for the least
significant digit is chosen; for instance, if \fBtcl_precision\fR is 3, then
0.3125 will convert to 0.312, not 0.313, while 0.6875 will convert to
0.688, not 0.687. Any string of trailing zeroes that remains is trimmed.
.RE
.PP
.RS
a \fBtcl_precision\fR value of 17 digits is
.QW perfect
for IEEE floating-point in that it allows
double-precision values to be converted to strings and back to
binary with no loss of information. For this reason, you will often
see it as a value in legacy code that must run on Tcl versions before
8.5. It is no longer recommended; as noted above, a zero value is the
preferred method.
.RE
.PP
.RS
All interpreters in a thread share a single \fBtcl_precision\fR value:
changing it in one interpreter will affect all other interpreters as
well.  Safe interpreters are not allowed to modify the
variable.
.RE
.PP
.RS
Valid values for \fBtcl_precision\fR range from 0 to 17.
.RE
.TP
\fBtcl_rcFileName\fR
.
This variable is used during initialization to indicate the name of a
user-specific startup file.  If it is set by application-specific
initialization, then the Tcl startup code will check for the existence
of this file and \fBsource\fR it if it exists.  For example, for \fBwish\fR
the variable is set to \fB~/.wishrc\fR for Unix and \fB~/wishrc.tcl\fR
for Windows.
.TP
\fBtcl_traceCompile\fR
.
The value of this variable can be set to control
how much tracing information
is displayed during bytecode compilation.
By default, \fBtcl_traceCompile\fR is zero and no information is displayed.
Setting \fBtcl_traceCompile\fR to 1 generates a one-line summary in \fBstdout\fR






|







 







<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<







6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
...
352
353
354
355
356
357
358
































































359
360
361
362
363
364
365
'\" of this file, and for a DISCLAIMER OF ALL WARRANTIES.
'\"
.TH tclvars n 8.0 Tcl "Tcl Built-In Commands"
.so man.macros
.BS
'\" Note:  do not modify the .SH NAME line immediately below!
.SH NAME
argc, argv, argv0, auto_path, env, errorCode, errorInfo, tcl_interactive, tcl_library, tcl_nonwordchars, tcl_patchLevel, tcl_pkgPath, tcl_platform, tcl_rcFileName, tcl_traceCompile, tcl_traceExec, tcl_wordchars, tcl_version \- Variables used by Tcl
.BE
.SH DESCRIPTION
.PP
The following global variables are created and managed automatically
by the Tcl library.  Except where noted below, these variables should
normally be treated as read-only by application-specific code and by users.
.TP
................................................................................
.TP
\fBwordSize\fR
.
This gives the size of the native-machine word in bytes (strictly, it
is same as the result of evaluating \fIsizeof(long)\fR in C.)
.RE
.TP
































































\fBtcl_traceCompile\fR
.
The value of this variable can be set to control
how much tracing information
is displayed during bytecode compilation.
By default, \fBtcl_traceCompile\fR is zero and no information is displayed.
Setting \fBtcl_traceCompile\fR to 1 generates a one-line summary in \fBstdout\fR

Changes to generic/tcl.h.

944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
#define Tcl_DStringValue(dsPtr) ((dsPtr)->string)
#ifndef TCL_NO_DEPRECATED
#   define Tcl_DStringTrunc Tcl_DStringSetLength
#endif /* !TCL_NO_DEPRECATED */

/*
 * Definitions for the maximum number of digits of precision that may be
 * specified in the "tcl_precision" variable, and the number of bytes of
 * buffer space required by Tcl_PrintDouble.
 */

#define TCL_MAX_PREC		17
#define TCL_DOUBLE_SPACE	(TCL_MAX_PREC+10)

/*
 * Definition for a number of bytes of buffer space sufficient to hold the






|
|







944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
#define Tcl_DStringValue(dsPtr) ((dsPtr)->string)
#ifndef TCL_NO_DEPRECATED
#   define Tcl_DStringTrunc Tcl_DStringSetLength
#endif /* !TCL_NO_DEPRECATED */

/*
 * Definitions for the maximum number of digits of precision that may be
 * produced by Tcl_PrintDouble, and the number of bytes of buffer space
 * required by Tcl_PrintDouble.
 */

#define TCL_MAX_PREC		17
#define TCL_DOUBLE_SPACE	(TCL_MAX_PREC+10)

/*
 * Definition for a number of bytes of buffer space sufficient to hold the

Changes to generic/tclBasic.c.

933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
    /*
     * Set up other variables such as tcl_version and tcl_library
     */

    Tcl_SetVar2(interp, "tcl_patchLevel", NULL, TCL_PATCH_LEVEL, TCL_GLOBAL_ONLY);
    Tcl_SetVar2(interp, "tcl_version", NULL, TCL_VERSION, TCL_GLOBAL_ONLY);
    Tcl_TraceVar2(interp, "tcl_precision", NULL,
	    TCL_GLOBAL_ONLY|TCL_TRACE_READS|TCL_TRACE_WRITES|TCL_TRACE_UNSETS,
	    TclPrecTraceProc, NULL);
    TclpSetVariables(interp);

#ifdef TCL_THREADS
    /*
     * The existence of the "threaded" element of the tcl_platform array
     * indicates that this particular Tcl shell has been compiled with threads
     * turned on. Using "info exists tcl_platform(threaded)" a Tcl script can






<
<
<







933
934
935
936
937
938
939



940
941
942
943
944
945
946
    /*
     * Set up other variables such as tcl_version and tcl_library
     */

    Tcl_SetVar2(interp, "tcl_patchLevel", NULL, TCL_PATCH_LEVEL, TCL_GLOBAL_ONLY);
    Tcl_SetVar2(interp, "tcl_version", NULL, TCL_VERSION, TCL_GLOBAL_ONLY);



    TclpSetVariables(interp);

#ifdef TCL_THREADS
    /*
     * The existence of the "threaded" element of the tcl_platform array
     * indicates that this particular Tcl shell has been compiled with threads
     * turned on. Using "info exists tcl_platform(threaded)" a Tcl script can

Changes to generic/tclInt.decls.

354
355
356
357
358
359
360

361
362
363
364

365
366
367
368
369
370
371
#  declare 86 {
#      int TclParseQuotes(Tcl_Interp *interp, char *str, int termChar,
#  	    int flags, char **termPtr, ParseValue *pvPtr)
#  }
#  declare 87 {
#      void TclPlatformInit(Tcl_Interp *interp)
#  }

declare 88 {
    char *TclPrecTraceProc(ClientData clientData, Tcl_Interp *interp,
	    const char *name1, const char *name2, int flags)
}

declare 89 {
    int TclPreventAliasLoop(Tcl_Interp *interp, Tcl_Interp *cmdInterp,
	    Tcl_Command cmd)
}
# Removed in 8.1 (only available if compiled with TCL_COMPILE_DEBUG):
#  declare 90 {
#      void TclPrintByteCodeObj(Tcl_Interp *interp, Tcl_Obj *objPtr)






>
|
|
|
<
>







354
355
356
357
358
359
360
361
362
363
364

365
366
367
368
369
370
371
372
#  declare 86 {
#      int TclParseQuotes(Tcl_Interp *interp, char *str, int termChar,
#  	    int flags, char **termPtr, ParseValue *pvPtr)
#  }
#  declare 87 {
#      void TclPlatformInit(Tcl_Interp *interp)
#  }
# Removed in 9.0:
#declare 88 {
#    char *TclPrecTraceProc(ClientData clientData, Tcl_Interp *interp,
#	    const char *name1, const char *name2, int flags)

#}
declare 89 {
    int TclPreventAliasLoop(Tcl_Interp *interp, Tcl_Interp *cmdInterp,
	    Tcl_Command cmd)
}
# Removed in 8.1 (only available if compiled with TCL_COMPILE_DEBUG):
#  declare 90 {
#      void TclPrintByteCodeObj(Tcl_Interp *interp, Tcl_Obj *objPtr)

Changes to generic/tclIntDecls.h.

206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
...
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
...
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
EXTERN char *		TclpRealloc(char *ptr, unsigned int size);
/* Slot 82 is reserved */
/* Slot 83 is reserved */
/* Slot 84 is reserved */
/* Slot 85 is reserved */
/* Slot 86 is reserved */
/* Slot 87 is reserved */
/* 88 */
EXTERN char *		TclPrecTraceProc(ClientData clientData,
				Tcl_Interp *interp, const char *name1,
				const char *name2, int flags);
/* 89 */
EXTERN int		TclPreventAliasLoop(Tcl_Interp *interp,
				Tcl_Interp *cmdInterp, Tcl_Command cmd);
/* Slot 90 is reserved */
/* 91 */
EXTERN void		TclProcCleanupProc(Proc *procPtr);
/* 92 */
................................................................................
    char * (*tclpRealloc) (char *ptr, unsigned int size); /* 81 */
    void (*reserved82)(void);
    void (*reserved83)(void);
    void (*reserved84)(void);
    void (*reserved85)(void);
    void (*reserved86)(void);
    void (*reserved87)(void);
    char * (*tclPrecTraceProc) (ClientData clientData, Tcl_Interp *interp, const char *name1, const char *name2, int flags); /* 88 */
    int (*tclPreventAliasLoop) (Tcl_Interp *interp, Tcl_Interp *cmdInterp, Tcl_Command cmd); /* 89 */
    void (*reserved90)(void);
    void (*tclProcCleanupProc) (Proc *procPtr); /* 91 */
    int (*tclProcCompileProc) (Tcl_Interp *interp, Proc *procPtr, Tcl_Obj *bodyPtr, Namespace *nsPtr, const char *description, const char *procName); /* 92 */
    void (*tclProcDeleteProc) (ClientData clientData); /* 93 */
    void (*reserved94)(void);
    void (*reserved95)(void);
................................................................................
	(tclIntStubsPtr->tclpRealloc) /* 81 */
/* Slot 82 is reserved */
/* Slot 83 is reserved */
/* Slot 84 is reserved */
/* Slot 85 is reserved */
/* Slot 86 is reserved */
/* Slot 87 is reserved */
#define TclPrecTraceProc \
	(tclIntStubsPtr->tclPrecTraceProc) /* 88 */
#define TclPreventAliasLoop \
	(tclIntStubsPtr->tclPreventAliasLoop) /* 89 */
/* Slot 90 is reserved */
#define TclProcCleanupProc \
	(tclIntStubsPtr->tclProcCleanupProc) /* 91 */
#define TclProcCompileProc \
	(tclIntStubsPtr->tclProcCompileProc) /* 92 */






|
<
<
<







 







|







 







|
<







206
207
208
209
210
211
212
213



214
215
216
217
218
219
220
...
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
...
984
985
986
987
988
989
990
991

992
993
994
995
996
997
998
EXTERN char *		TclpRealloc(char *ptr, unsigned int size);
/* Slot 82 is reserved */
/* Slot 83 is reserved */
/* Slot 84 is reserved */
/* Slot 85 is reserved */
/* Slot 86 is reserved */
/* Slot 87 is reserved */
/* Slot 88 is reserved */



/* 89 */
EXTERN int		TclPreventAliasLoop(Tcl_Interp *interp,
				Tcl_Interp *cmdInterp, Tcl_Command cmd);
/* Slot 90 is reserved */
/* 91 */
EXTERN void		TclProcCleanupProc(Proc *procPtr);
/* 92 */
................................................................................
    char * (*tclpRealloc) (char *ptr, unsigned int size); /* 81 */
    void (*reserved82)(void);
    void (*reserved83)(void);
    void (*reserved84)(void);
    void (*reserved85)(void);
    void (*reserved86)(void);
    void (*reserved87)(void);
    void (*reserved88)(void);
    int (*tclPreventAliasLoop) (Tcl_Interp *interp, Tcl_Interp *cmdInterp, Tcl_Command cmd); /* 89 */
    void (*reserved90)(void);
    void (*tclProcCleanupProc) (Proc *procPtr); /* 91 */
    int (*tclProcCompileProc) (Tcl_Interp *interp, Proc *procPtr, Tcl_Obj *bodyPtr, Namespace *nsPtr, const char *description, const char *procName); /* 92 */
    void (*tclProcDeleteProc) (ClientData clientData); /* 93 */
    void (*reserved94)(void);
    void (*reserved95)(void);
................................................................................
	(tclIntStubsPtr->tclpRealloc) /* 81 */
/* Slot 82 is reserved */
/* Slot 83 is reserved */
/* Slot 84 is reserved */
/* Slot 85 is reserved */
/* Slot 86 is reserved */
/* Slot 87 is reserved */
/* Slot 88 is reserved */

#define TclPreventAliasLoop \
	(tclIntStubsPtr->tclPreventAliasLoop) /* 89 */
/* Slot 90 is reserved */
#define TclProcCleanupProc \
	(tclIntStubsPtr->tclProcCleanupProc) /* 91 */
#define TclProcCompileProc \
	(tclIntStubsPtr->tclProcCompileProc) /* 92 */

Changes to generic/tclObj.c.

2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
 
/*
 *----------------------------------------------------------------------
 *
 * UpdateStringOfDouble --
 *
 *	Update the string representation for a double-precision floating point
 *	object. This must obey the current tcl_precision value for
 *	double-to-string conversions. Note: This function does not free an
 *	existing old string rep so storage will be lost if this has not
 *	already been done.
 *
 * Results:
 *	None.
 *
 * Side effects:






<
|







2345
2346
2347
2348
2349
2350
2351

2352
2353
2354
2355
2356
2357
2358
2359
 
/*
 *----------------------------------------------------------------------
 *
 * UpdateStringOfDouble --
 *
 *	Update the string representation for a double-precision floating point

 *	object. Note: This function does not free an
 *	existing old string rep so storage will be lost if this has not
 *	already been done.
 *
 * Results:
 *	None.
 *
 * Side effects:

Changes to generic/tclStrToD.c.

3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
....
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
 *	endPtr to point to the terminating '\0' byte of the string. Sets *sign
 *	to 1 if a minus sign should be printed with the number, or 0 if a plus
 *	sign (or no sign) should appear.
 *
 * This function is a service routine that produces the string of digits for
 * floating-point-to-decimal conversion. It can do a number of things
 * according to the 'flags' argument. Valid values for 'flags' include:
 *	TCL_DD_SHORTEST - This is the default for floating point conversion if
 *		::tcl_precision is 0. It constructs the shortest string of
 *		digits that will reconvert to the given number when scanned.
 *		For floating point numbers that are exactly between two
 *		decimal numbers, it resolves using the 'round to even' rule.
 *		With this value, the 'ndigits' parameter is ignored.
 *	TCL_DD_STEELE - This value is not recommended and may be removed in
 *		the future. It follows the conversion algorithm outlined in
 *		"How to Print Floating-Point Numbers Accurately" by Guy
................................................................................
 *		L. Steele, Jr. and Jon L. White [Proc. ACM SIGPLAN '90,
 *		pp. 112-126]. This rule has the effect of rendering 1e23 as
 *		9.9999999999999999e22 - which is a 'better' approximation in
 *		the sense that it will reconvert correctly even if a
 *		subsequent input conversion is 'round up' or 'round down'
 *		rather than 'round to nearest', but is surprising otherwise.
 *	TCL_DD_E_FORMAT - This value is used to prepare numbers for %e format
 *		conversion (or for default floating->string if tcl_precision
 *		is not 0). It constructs a string of at most 'ndigits' digits,
 *		choosing the one that is closest to the given number (and
 *		resolving ties with 'round to even').  It is allowed to return
 *		fewer than 'ndigits' if the number converts exactly; if the
 *		TCL_DD_E_FORMAT|TCL_DD_SHORTEN_FLAG is supplied instead, it
 *		also returns fewer digits if the shorter string will still
 *		reconvert without loss to the given input number. In any case,
 *		strings of trailing zeroes are suppressed.






|
|







 







<
|







3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
....
4008
4009
4010
4011
4012
4013
4014

4015
4016
4017
4018
4019
4020
4021
4022
 *	endPtr to point to the terminating '\0' byte of the string. Sets *sign
 *	to 1 if a minus sign should be printed with the number, or 0 if a plus
 *	sign (or no sign) should appear.
 *
 * This function is a service routine that produces the string of digits for
 * floating-point-to-decimal conversion. It can do a number of things
 * according to the 'flags' argument. Valid values for 'flags' include:
 *	TCL_DD_SHORTEST - This is the default for floating point conversion.
 *		It constructs the shortest string of
 *		digits that will reconvert to the given number when scanned.
 *		For floating point numbers that are exactly between two
 *		decimal numbers, it resolves using the 'round to even' rule.
 *		With this value, the 'ndigits' parameter is ignored.
 *	TCL_DD_STEELE - This value is not recommended and may be removed in
 *		the future. It follows the conversion algorithm outlined in
 *		"How to Print Floating-Point Numbers Accurately" by Guy
................................................................................
 *		L. Steele, Jr. and Jon L. White [Proc. ACM SIGPLAN '90,
 *		pp. 112-126]. This rule has the effect of rendering 1e23 as
 *		9.9999999999999999e22 - which is a 'better' approximation in
 *		the sense that it will reconvert correctly even if a
 *		subsequent input conversion is 'round up' or 'round down'
 *		rather than 'round to nearest', but is surprising otherwise.
 *	TCL_DD_E_FORMAT - This value is used to prepare numbers for %e format

 *		conversion. It constructs a string of at most 'ndigits' digits,
 *		choosing the one that is closest to the given number (and
 *		resolving ties with 'round to even').  It is allowed to return
 *		fewer than 'ndigits' if the number converts exactly; if the
 *		TCL_DD_E_FORMAT|TCL_DD_SHORTEN_FLAG is supplied instead, it
 *		also returns fewer digits if the shorter string will still
 *		reconvert without loss to the given input number. In any case,
 *		strings of trailing zeroes are suppressed.

Changes to generic/tclStubInit.c.

372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
    TclpRealloc, /* 81 */
    0, /* 82 */
    0, /* 83 */
    0, /* 84 */
    0, /* 85 */
    0, /* 86 */
    0, /* 87 */
    TclPrecTraceProc, /* 88 */
    TclPreventAliasLoop, /* 89 */
    0, /* 90 */
    TclProcCleanupProc, /* 91 */
    TclProcCompileProc, /* 92 */
    TclProcDeleteProc, /* 93 */
    0, /* 94 */
    0, /* 95 */






|







372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
    TclpRealloc, /* 81 */
    0, /* 82 */
    0, /* 83 */
    0, /* 84 */
    0, /* 85 */
    0, /* 86 */
    0, /* 87 */
    0, /* 88 */
    TclPreventAliasLoop, /* 89 */
    0, /* 90 */
    TclProcCleanupProc, /* 91 */
    TclProcCompileProc, /* 92 */
    TclProcDeleteProc, /* 93 */
    0, /* 94 */
    0, /* 95 */

Changes to generic/tclUtil.c.

89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
....
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
....
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
....
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
....
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
#define COMPAT 1
#define CONVERT_NONE	0
#define CONVERT_BRACE	2
#define CONVERT_ESCAPE	4
#define CONVERT_MASK	(CONVERT_BRACE | CONVERT_ESCAPE)
#define CONVERT_ANY	16

/*
 * The following key is used by Tcl_PrintDouble and TclPrecTraceProc to
 * access the precision to be used for double formatting.
 */

static Tcl_ThreadDataKey precisionKey;

/*
 * Prototypes for functions defined later in this file.
 */

static void		ClearHash(Tcl_HashTable *tablePtr);
static void		FreeProcessGlobalValue(ClientData clientData);
static void		FreeThreadHash(ClientData clientData);
................................................................................
 *
 * Tcl_PrintDouble --
 *
 *	Given a floating-point value, this function converts it to an ASCII
 *	string using.
 *
 * Results:
 *	The ASCII equivalent of "value" is written at "dst". It is written
 *	using the current precision, and it is guaranteed to contain a decimal
 *	point or exponent, so that it looks like a floating-point value and
 *	not an integer.
 *
 * Side effects:
 *	None.
 *
 *----------------------------------------------------------------------
 */

void
Tcl_PrintDouble(
    Tcl_Interp *interp,		/* Interpreter whose tcl_precision variable
				 * used to be used to control printing. It's
				 * ignored now. */
    double value,		/* Value to print as string. */
    char *dst)			/* Where to store converted value; must have
				 * at least TCL_DOUBLE_SPACE characters. */
{
    char *p, c;
    int exponent;
    int signum;
    char *digits;
    char *end;
    int *precisionPtr = Tcl_GetThreadData(&precisionKey, sizeof(int));

    /*
     * Handle NaN.
     */

    if (TclIsNaN(value)) {
	TclFormatNaN(value, dst);
................................................................................
	return;
    }

    /*
     * Ordinary (normal and denormal) values.
     */

    if (*precisionPtr == 0) {
	digits = TclDoubleDigits(value, -1, TCL_DD_SHORTEST,
		&exponent, &signum, &end);
    } else {
	/*
	 * There are at least two possible interpretations for tcl_precision.
	 *
	 * The first is, "choose the decimal representation having
	 * $tcl_precision digits of significance that is nearest to the given
	 * number, breaking ties by rounding to even, and then trimming
	 * trailing zeros." This gives the greatest possible precision in the
	 * decimal string, but offers the anomaly that [expr 0.1] will be
	 * "0.10000000000000001".
	 *
	 * The second is "choose the decimal representation having at most
	 * $tcl_precision digits of significance that is nearest to the given
	 * number. If no such representation converts exactly to the given
	 * number, choose the one that is closest, breaking ties by rounding
	 * to even. If more than one such representation converts exactly to
	 * the given number, choose the shortest, breaking ties in favour of
	 * the nearest, breaking remaining ties in favour of the one ending in
	 * an even digit."
	 *
	 * Tcl 8.4 implements the first of these, which gives rise to
	 * anomalies in formatting:
	 *
	 *	% expr 0.1
	 *	0.10000000000000001
	 *	% expr 0.01
	 *	0.01
	 *	% expr 1e-7
	 *	9.9999999999999995e-08
	 *
	 * For human readability, it appears better to choose the second rule,
	 * and let [expr 0.1] return 0.1. But for 8.4 compatibility, we prefer
	 * the first (the recommended zero value for tcl_precision avoids the
	 * problem entirely).
	 *
	 * Uncomment TCL_DD_SHORTEN_FLAG in the next call to prefer the method
	 * that allows floating point values to be shortened if it can be done
	 * without loss of precision.
	 */

	digits = TclDoubleDigits(value, *precisionPtr,
		TCL_DD_E_FORMAT /* | TCL_DD_SHORTEN_FLAG */,
		&exponent, &signum, &end);
    }
    if (signum) {
	*dst++ = '-';
    }
    p = digits;
    if (exponent < -4 || exponent > 16) {
	/*
	 * E format for numbers < 1e-3 or >= 1e17.
................................................................................
	    *dst++ = '.';
	    while (c != '\0') {
		*dst++ = c;
		c = *++p;
	    }
	}

	/*
	 * Tcl 8.4 appears to format with at least a two-digit exponent;
	 * preserve that behaviour when tcl_precision != 0
	 */

	if (*precisionPtr == 0) {
	    sprintf(dst, "e%+d", exponent);
	} else {
	    sprintf(dst, "e%+03d", exponent);
	}
    } else {
	/*
	 * F format for others.
	 */

	if (exponent < 0) {
	    *dst++ = '0';
................................................................................
		c = *++p;
	    }
	}
	*dst++ = '\0';
    }
    ckfree(digits);
}
 
/*
 *----------------------------------------------------------------------
 *
 * TclPrecTraceProc --
 *
 *	This function is invoked whenever the variable "tcl_precision" is
 *	written.
 *
 * Results:
 *	Returns NULL if all went well, or an error message if the new value
 *	for the variable doesn't make sense.
 *
 * Side effects:
 *	If the new value doesn't make sense then this function undoes the
 *	effect of the variable modification. Otherwise it modifies the format
 *	string that's used by Tcl_PrintDouble.
 *
 *----------------------------------------------------------------------
 */

	/* ARGSUSED */
char *
TclPrecTraceProc(
    ClientData clientData,	/* Not used. */
    Tcl_Interp *interp,		/* Interpreter containing variable. */
    const char *name1,		/* Name of variable. */
    const char *name2,		/* Second part of variable name. */
    int flags)			/* Information about what happened. */
{
    Tcl_Obj *value;
    int prec;
    int *precisionPtr = Tcl_GetThreadData(&precisionKey, sizeof(int));

    /*
     * If the variable is unset, then recreate the trace.
     */

    if (flags & TCL_TRACE_UNSETS) {
	if ((flags & TCL_TRACE_DESTROYED) && !Tcl_InterpDeleted(interp)) {
	    Tcl_TraceVar2(interp, name1, name2,
		    TCL_GLOBAL_ONLY|TCL_TRACE_READS|TCL_TRACE_WRITES
		    |TCL_TRACE_UNSETS, TclPrecTraceProc, clientData);
	}
	return NULL;
    }

    /*
     * When the variable is read, reset its value from our shared value. This
     * is needed in case the variable was modified in some other interpreter
     * so that this interpreter's value is out of date.
     */


    if (flags & TCL_TRACE_READS) {
	Tcl_SetVar2Ex(interp, name1, name2, Tcl_NewIntObj(*precisionPtr),
		flags & TCL_GLOBAL_ONLY);
	return NULL;
    }

    /*
     * The variable is being written. Check the new value and disallow it if
     * it isn't reasonable or if this is a safe interpreter (we don't want
     * safe interpreters messing up the precision of other interpreters).
     */

    if (Tcl_IsSafe(interp)) {
	return (char *) "can't modify precision from a safe interpreter";
    }
    value = Tcl_GetVar2Ex(interp, name1, name2, flags & TCL_GLOBAL_ONLY);
    if (value == NULL
	    || Tcl_GetIntFromObj(NULL, value, &prec) != TCL_OK
	    || prec < 0 || prec > TCL_MAX_PREC) {
	return (char *) "improper value for precision";
    }
    *precisionPtr = prec;
    return NULL;
}
 
/*
 *----------------------------------------------------------------------
 *
 * TclNeedSpace --
 *
 *	This function checks to see whether it is appropriate to add a space






<
<
<
<
<
<
<







 







|
|
|
<









|
<
<









<







 







<
|
|
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<







 







<
<
<
<
<
<
|
<
<
<







 







<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<







89
90
91
92
93
94
95







96
97
98
99
100
101
102
....
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035

3036
3037
3038
3039
3040
3041
3042
3043
3044
3045


3046
3047
3048
3049
3050
3051
3052
3053
3054

3055
3056
3057
3058
3059
3060
3061
....
3079
3080
3081
3082
3083
3084
3085

3086
3087












































3088
3089
3090
3091
3092
3093
3094
....
3100
3101
3102
3103
3104
3105
3106






3107



3108
3109
3110
3111
3112
3113
3114
....
3134
3135
3136
3137
3138
3139
3140














































































3141
3142
3143
3144
3145
3146
3147
#define COMPAT 1
#define CONVERT_NONE	0
#define CONVERT_BRACE	2
#define CONVERT_ESCAPE	4
#define CONVERT_MASK	(CONVERT_BRACE | CONVERT_ESCAPE)
#define CONVERT_ANY	16








/*
 * Prototypes for functions defined later in this file.
 */

static void		ClearHash(Tcl_HashTable *tablePtr);
static void		FreeProcessGlobalValue(ClientData clientData);
static void		FreeThreadHash(ClientData clientData);
................................................................................
 *
 * Tcl_PrintDouble --
 *
 *	Given a floating-point value, this function converts it to an ASCII
 *	string using.
 *
 * Results:
 *	The ASCII equivalent of "value" is written at "dst". It is guaranteed
 *	to contain a decimal point or exponent, so that it looks like a
 *	floating-point value and not an integer.

 *
 * Side effects:
 *	None.
 *
 *----------------------------------------------------------------------
 */

void
Tcl_PrintDouble(
    Tcl_Interp *interp,		/* Not used */


    double value,		/* Value to print as string. */
    char *dst)			/* Where to store converted value; must have
				 * at least TCL_DOUBLE_SPACE characters. */
{
    char *p, c;
    int exponent;
    int signum;
    char *digits;
    char *end;


    /*
     * Handle NaN.
     */

    if (TclIsNaN(value)) {
	TclFormatNaN(value, dst);
................................................................................
	return;
    }

    /*
     * Ordinary (normal and denormal) values.
     */


    digits = TclDoubleDigits(value, -1, TCL_DD_SHORTEST,
	    &exponent, &signum, &end);












































    if (signum) {
	*dst++ = '-';
    }
    p = digits;
    if (exponent < -4 || exponent > 16) {
	/*
	 * E format for numbers < 1e-3 or >= 1e17.
................................................................................
	    *dst++ = '.';
	    while (c != '\0') {
		*dst++ = c;
		c = *++p;
	    }
	}







	sprintf(dst, "e%+d", exponent);



    } else {
	/*
	 * F format for others.
	 */

	if (exponent < 0) {
	    *dst++ = '0';
................................................................................
		c = *++p;
	    }
	}
	*dst++ = '\0';
    }
    ckfree(digits);
}














































































 
/*
 *----------------------------------------------------------------------
 *
 * TclNeedSpace --
 *
 *	This function checks to see whether it is appropriate to add a space

Changes to tests/basic.test.

889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
        set leak [expr {$end - $tmp}]
    } -cleanup {
	unset end i tmp
	rename getbytes {}
	rename stress {}
} -result 0

test basic-48.17.$noComp {expansion: object safety} -setup {
        set old_precision $::tcl_precision
        set ::tcl_precision 4
    } -constraints $constraints -body {
            set third [expr {1.0/3.0}]
            set l [list $third $third]
            set x [run {list $third {*}$l $third}]
	    set res [list]
            foreach t $x {
                lappend res [expr {$t * 3.0}]
            }
            set res
    } -cleanup {
        set ::tcl_precision $old_precision
        unset old_precision res t l x third
} -result {1.0 1.0 1.0 1.0}

test basic-48.18.$noComp {expansion: list semantics} -constraints $constraints -body {
        set badcmd {
            list a b
            set apa 10
        }






|
<
<
<









<
|







889
890
891
892
893
894
895
896



897
898
899
900
901
902
903
904
905

906
907
908
909
910
911
912
913
        set leak [expr {$end - $tmp}]
    } -cleanup {
	unset end i tmp
	rename getbytes {}
	rename stress {}
} -result 0

test basic-48.17.$noComp {expansion: object safety} -constraints $constraints -body {



            set third [expr {1.0/3.0}]
            set l [list $third $third]
            set x [run {list $third {*}$l $third}]
	    set res [list]
            foreach t $x {
                lappend res [expr {$t * 3.0}]
            }
            set res
    } -cleanup {

        unset res t l x third
} -result {1.0 1.0 1.0 1.0}

test basic-48.18.$noComp {expansion: list semantics} -constraints $constraints -body {
        set badcmd {
            list a b
            set apa 10
        }

Changes to tests/util.test.

420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
test util-6.5 {Tcl_PrintDouble - make sure there's a decimal point} {
    concat x[expr 2.0]
} {x2.0}
test util-6.6 {Tcl_PrintDouble - make sure there's a decimal point} {
    concat x[expr 3.0e98]
} {x3e+98}

test util-7.1 {TclPrecTraceProc - unset callbacks} -setup {
    set old_precision $::tcl_precision
} -body {
    set tcl_precision 7
    set x $tcl_precision
    unset tcl_precision
    list $x $tcl_precision
} -cleanup {
    set ::tcl_precision $old_precision
} -result {7 7}
test util-7.2 {TclPrecTraceProc - read traces, sharing among interpreters}  -setup {
    set old_precision $::tcl_precision
} -body {
    set tcl_precision 12
    interp create child
    set x [child eval set tcl_precision]
    child eval {set tcl_precision 6}
    interp delete child
    list $x $tcl_precision
} -cleanup {
    set ::tcl_precision $old_precision
} -result {12 6}
test util-7.3 {TclPrecTraceProc - write traces, safe interpreters} -setup {
    set old_precision $::tcl_precision
} -body {
    set tcl_precision 12
    interp create -safe child
    set x [child eval {
	list [catch {set tcl_precision 8} msg] $msg
    }]
    interp delete child
    list $x $tcl_precision
} -cleanup {
    set ::tcl_precision $old_precision
} -result {{1 {can't set "tcl_precision": can't modify precision from a safe interpreter}} 12}
test util-7.4 {TclPrecTraceProc - write traces, bogus values} -setup {
    set old_precision $::tcl_precision
} -body {
    set tcl_precision 12
    list [catch {set tcl_precision abc} msg] $msg $tcl_precision
} -cleanup {
    set ::tcl_precision $old_precision
} -result {1 {can't set "tcl_precision": improper value for precision} 12}

# This test always succeeded in the C locale anyway...
test util-8.1 {TclNeedSpace - correct UTF8 handling} {
    # Bug 411825
    # Note that this test relies on the fact that
    # [interp target] calls on Tcl_AppendElement()
    # which calls on TclNeedSpace().  If [interp target]
    # is ever updated, this test will no longer test






<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<







420
421
422
423
424
425
426












































427
428
429
430
431
432
433
test util-6.5 {Tcl_PrintDouble - make sure there's a decimal point} {
    concat x[expr 2.0]
} {x2.0}
test util-6.6 {Tcl_PrintDouble - make sure there's a decimal point} {
    concat x[expr 3.0e98]
} {x3e+98}













































# This test always succeeded in the C locale anyway...
test util-8.1 {TclNeedSpace - correct UTF8 handling} {
    # Bug 411825
    # Note that this test relies on the fact that
    # [interp target] calls on Tcl_AppendElement()
    # which calls on TclNeedSpace().  If [interp target]
    # is ever updated, this test will no longer test