- Jun 20, 2005
-
-
rtoy authored
in case we're given a logical pathname.
-
- Jun 19, 2005
-
-
rtoy authored
micro-optimization.
-
- Jun 15, 2005
-
-
rtoy authored
Make sure the position arg to flonum-to-string is non-negative in this case.
-
- Jun 14, 2005
-
-
rtoy authored
cmucl-imp, on or around 2005/06/13. Some useful tests: (format nil "~9,3,2,0,'%G" 0.0314159) Expected: "0.314e-01" (format nil "~9,3,2,-2,'%@e" 3.14159) Expected: "+.003E+03" (format nil "~6,2,1,'*F" 3.14159) Expected: " 31.42" (format nil "~9,0,6f" 3.14159) Expected: " 3141590." (let (x) (dotimes (k 13 x) (setq x (cons (format nil "~%Scale factor ~2D: |~13,6,2,VE|" (- k 5) (- k 5) 3.14159) x)))) (" Scale factor 7: | 3141590.e-06|" " Scale factor 6: | 314159.0e-05|" " Scale factor 5: | 31415.90e-04|" " Scale factor 4: | 3141.590e-03|" " Scale factor 3: | 314.1590e-02|" " Scale factor 2: | 31.41590e-01|" " Scale factor 1: | 3.141590e+00|" " Scale factor 0: | 0.314159e+01|" " Scale factor -1: | 0.031416e+02|" " Scale factor -2: | 0.003142e+03|" " Scale factor -3: | 0.000314e+04|" " Scale factor -4: | 0.000031e+05|" " Scale factor -5: | 0.000003e+06|") code/format.lisp: o If the scale factor (k) is negative, the min number of digits to print is 1, not (- 1 k) because that prints too many if the field is too short. Setting fmin to fdig is ok if k >= 0. (See scale factor test above.) o If flonum-to-string returns with a trailing decimal point, we don't need to decrement spaceleft because that deletes a white-space character. (See first scale factor 7 test above.) code/print.lisp: o We need to adjust the number of digits to be printed to include the scale factor. See tests above.
-
- Jun 13, 2005
-
-
rtoy authored
This is done by faking it. The file being compiled is compiled as usual, but we append fake forms to the file as if they came from the file. These fake forms insert the necessary information into the xref databases when the fasl is loaded. To support this feature, we also updated COMPILE-FILE to recognize the :xref keyword arg. Set this to non-NIL to enable computing and saving xref information. code/exports.lisp: o Update XREF exports compiler/fndb.lisp: o Update with new definition of COMPILE-FILE. compiler/main.lisp: o Append fake forms to the file being compiled to save xref information to the fasl. This clears out any xref info we might have for the file, and inserts the necessary xref information into the database. o Add :XREF keyword arg to COMPILE-FILE. Default value of :XREF is C::*RECORD-XREF-INFO*. compiler/xref.lisp: o Add function to invalidate xref info for a given namestring, so we can reset the info when a fasl with xref info is loaded. o Add a function to find all xref information for a given pathname. Used for saving xref info to a fasl.
-
- Jun 01, 2005
-
-
rtoy authored
-
- May 26, 2005
- May 23, 2005
-
-
rtoy authored
string.
-
rtoy authored
Description: In (peek-char nil s nil foo), if foo happens to be the same character that peek-char returns, the character is removed from the input stream, as if read by read-char. Examples: * (with-input-from-string (s "123") (list (peek-char nil s nil #\1) (read-char s) (read-char s))) (#\1 #\2 #\3) This fix is based on the version proposed by Rudi Schlatte, with minor changes in naming.
-
- May 19, 2005
-
-
rtoy authored
short and fit on one line.
-
- May 17, 2005
-
-
rtoy authored
extend over multiple lines.
-
- May 12, 2005
-
-
rtoy authored
(declare (type fixnum n) (type bar x)) We now get (declare (type fixnum n) (type bar x))
-
- May 11, 2005
-
-
rtoy authored
:absolute or :wild-inferiors followed by :up or :back. Not exactly sure what pathname to use here since we don't accept these pathnames. So we use the given pathname with the :up/:back deleted. And the format strings gives more info about what happened. Fixes ansi-test MAKE-PATHNAME-ERROR-RELATIVE-WILD-INFERIORS-UP and friends.
-
rtoy authored
valid function name. o Make the setf function-name-syntax try return the function name even if it's not a valid setf function name. o Use the above changes to fill out more completely and correctly the simple-type-error in FDEFINITION-OBJECT, when the function name is not valid. This fixes ansi-tests for fboundp and friends where the datum was a type of expected-type or where the datum had the wrong name.
-
- May 10, 2005
-
-
rtoy authored
so it might actually be a list when we generate the error. But 2nd isn't a list. This is a bit of kludge, but it fixes ansi-test nreconc.error.4.
-
- May 09, 2005
-
-
rtoy authored
-
rtoy authored
o Add PROPER-LIST-P function to determine if a list is a proper list. o Use it as the expected type for errors from functions requiring a proper list code/symbol.lisp: o Add VALID-PROPERTY-LIST-P function. o Use it as the expected type for errors from malformed property lists. These changes fix a few ansi-test bugs where the test was checking if the the expected-type was correct.
-
rtoy authored
can intersect any array, Fixes ansi-test misc.537.
-
rtoy authored
vector (to prevent confusion when the vector has a fill-pointer). Also note if the vector is adjustable or not.
-
- May 06, 2005
-
-
rtoy authored
a list.
-
- May 03, 2005
-
-
rtoy authored
before updating the items. From Aleksandar Bakic, cmucl-imp, 2005/05/02.
-
- Apr 30, 2005
-
-
rtoy authored
o Remove some unused code.
-
- Apr 28, 2005
- Apr 24, 2005
-
-
rtoy authored
Fixes misc.60 ansi-test.
-
- Apr 22, 2005
-
-
rtoy authored
anymore, because these functions are supposed to take just 2 args.
-
- Apr 19, 2005
-
-
rtoy authored
(let ((s (make-two-way-stream (make-concatenated-stream) (make-broadcast-stream)))) (dotimes (i 10) (format s "Hello, i = ~a~%" i))) doesn't work. This is caused by TWO-WAY-MISC not handling the :CHARPOS and :LINE-LENGTH operations. Just call the appropriate versions for the output stream of the two-way stream.
-
rtoy authored
non-NIL object. Fixes the MAKE-DISPATCH-MACRO-CHARACTER ansi-tests.
-
- Apr 18, 2005
- Apr 15, 2005
-
-
rtoy authored
o Add function for ppc to compute the number of CPU cycles per timebase ticks. Save this in a global variable. code/time.lisp: o Adjust cycle-count/float for ppc to use the cycles-per-tick to compute the actual number of CPU cycles, so we can display cycles instead of ticks.
-
- Apr 14, 2005
-
-
rtoy authored
o Define EXT:INVALID-FASL condition o Signal such a condition when we try to load a fasl compiled for a different version of CMUCL. code/exports.lisp: o Export EXT:INVALID-FASL o Export a few other symbols we forgot.
-
- Apr 13, 2005
-
-
pwerkowski authored
AFAICT, ANSI does not specify any type of exceptional condition for array bounds errors, and a type-error doesn't seem appropriate here. Reverting to SIMPLE-ERROR.
-
- Apr 04, 2005
-
-
rtoy authored
is done, based on a suggestion from Lynn Quam. o Add BINARY-TEXT-STREAM structure for streams that support binary and character I/O. This is basically an FD-STREAM. o Export BINARY-TEXT-STREAM from the EXTENSIONS package. o OPEN will recognize :class 'ext:binary-text-stream to enable such streams. o Removed *FD-STREAM-ENABLE-CHARACTER-AND-BINARY-INPUT*, since we use OPEN to specify that now. Use boot-2005-03-2.lisp to bootstrap this change.
-
- Mar 30, 2005
-
-
rtoy authored
WRITE-VECTOR are supposed to be octet positions rather than element-positions. From Lynn Quam, cmucl-imp, 2005-03-29.
-
- Mar 22, 2005
- Mar 18, 2005
-
-
rtoy authored
-