- Oct 21, 2005
-
-
rtoy authored
so signal an error when we try to generate the the namestring.
-
rtoy authored
condition, or defstruct, and vice versa, so signal an error if we try to do that.
-
rtoy authored
-
rtoy authored
normally.
-
rtoy authored
converted to an internal number, an error of type reader-error is signaled." So, disable underflow traps, convert the number. If the resulting float is zero, but the number was not, signal a reader-error.
-
rtoy authored
underflow traps are enabled and rehash-threshold is very small.
-
rtoy authored
sure that we adjust the stack pointer by a multiple of 16-bytes.
-
rtoy authored
sure FULL_FRAME_SIZE is a multiple of 16. (From a patch to sbcl from Cyrus Harmon.)
-
- Oct 20, 2005
-
-
rtoy authored
10), and (log 17 10) returned different single-float results. I think this is allowed by ANSI CL, but I think it's unfortunate. Therefore, try to apply float contagion to the arguments before computing the log function. Also, if both args are single-floats or rationals, we coerce them to double-floats before computing the result. This makes (log 17 10) = (log 17.0 10). There are other cases still to be considered.
-
- Oct 19, 2005
-
-
rtoy authored
destructive functions (such as nreverse) modify constant args and for warning when the result of destructive functions is not used.
-
rtoy authored
nreverse) modify constant args and for warning when the result of destructive functions is not used. Detecting modification of constant args is done by adding a new IR1 transformation that checks that a function is not destructively modifying constants or literals. A new IR1 attribute, important-result, is used to determine if the result of a function should be used instead of discarded. (Note: this means some functions are not detected. This should probably be implemented as another transform so the compiler can detect those cases.) code/error.lisp: o Add new condition CONSTANT-MODIFIED. compiler/fndb.lisp: o Note destructive functions that should not modify constant args o Note destructive functions whose results should be used. compiler/ir1opt.lisp: o Add new function CHECK-IMPORTANT-RESULT to check if the result is used. o Update IR1-OPTIMIZE-COMBINATION to check if a function destructively modifies constant args or if the result of a destructive function is not used. compiler/knownfun.lisp: o Add new IR1 attribute, IMPORTANT-RESULT, to indicate that the result of a function should be used. o Add new FUNCTION-INFO slot, DESTROYED-CONSTANT-ARGS. This holds a function for computing the constant or literal arguments which are destructively modified by the function. o Add support functions for DESTROYED-CONSTANT-ARGS functions.
-
- Oct 18, 2005
- Oct 14, 2005
- Oct 11, 2005
-
-
rtoy authored
double-floats correctly. (Double's were broken before.) o Fix MAKE-CALLBACK-TRAMPOLINE-SEGMENT which wasn't saving all the callee-saved registers and was not adjusting the stack properly when saving these registers. These changes should (Ha!) fix all of the issues reported by Luis Oliveira for ppc callbacks.
-
- Oct 10, 2005
-
-
rtoy authored
PROBE-FILE return a pathname with directory and name components set appropriately. This is done by having UNIX-RESOLVE-LINKS append a slash when the pathname actually names a directory and not a file.
-
rtoy authored
PROBE-FILE return a pathname with directory and name components set appropriately. This is done by having UNIX-RESOLVE-LINKS append a slash when the pathname actually names a directory and not a file.
-
- Oct 07, 2005
-
-
rtoy authored
default encapsulation from :default to t for ppc, until this is fixed.
-
- Oct 06, 2005
- Oct 05, 2005
-
-
rtoy authored
from Rex Dieter, for gcc4 compatibility.
-
rtoy authored
doesn't exist anymore.
-
rtoy authored
shortest namestring: (enough-namestring #p"/foo" #p"/") => "foo", not "./foo" (enough-namestring #p"foo/bar" #p"foo/") => "bar", not "foo/bar" We check for a common prefix for all cases first now, instead of just when the pathname is :absolute.
-
- Oct 03, 2005
-
-
rtoy authored
-
- Sep 30, 2005
-
-
rtoy authored
o Don't allow a namestring to be parsed as a search-list if a logical host with the same host name already exists. Allowing this causes confusing in printing such pathnames because, when read, the logical host takes precedence. unparse-unix-file: o Preserve version :NEWEST when printing out logical pathnames. This makes #p"host:foo.bar.newest" readable.
-
rtoy authored
:newest are equivalent in CMUCL. src/pathname.lisp: o Make EQUAL compare pathnames such that version nil and :newest are equal. src/hash-new.lisp: o Adjust sxhash to treat :version nil and :newest the same. (Version :newest is hashed as if it were nil.)
-
rtoy authored
o Remove unused function o Remove redundant stack-pointer variable o Clean up comments.
-
rtoy authored
them from getting clobbered because the linkage-table trampoline uses them. This was causing the callback qsort example from Luis Oliviera to fail. There might be other registers that need saving.
-
- Sep 29, 2005
-
-
rtoy authored
incorrectly handling everything, including things that were not integers. I don't know why this didn't fail at runtime. So add a case for alien::integer$ and adjust the default case.
-
rtoy authored
instruction so that the disassembler prints it as "BLRL" instead of "BCLRL :BO-U, 0"
-
rtoy authored
selecting the instruction printer. Thus, the mflr instruction was not correctly disassembled. However, the emitter for these instructions appears to be correct.
-
- Sep 27, 2005
-
-
rtoy authored
o A declarations for the path designators that were missing in PATHNAME-MATCH-P and ENOUGH-NAMESTRING.
-
rtoy authored
o Fix regression in CALLBACK-ACCESSOR-FORM in handling double-floats. We only want to compute the offset for integer type args. o MAKE-CALLBACK-TRAMPOLINE-SEGMENT didn't correctly handle void return types. Don't need to allocate any stack space for the return value because there is none.
-
- Sep 26, 2005
-
-
rtoy authored
components of a pathname object.
-
- Sep 25, 2005
-
-
rtoy authored
o Leave some debugging prints in, but commented out. o In the last case for EXPLICIT-VERSION, when looking for digits surrounded by ~'s, return version NIL if we don't find just digits. #p"foo.*" was returning version :newest. I think we really want :version nil. unparse-unix-enough: o If the pathname has no directory, it's relative to the defaults. Returning NIL is probably as good as returning '(:RELATIVE), and results in a shorter namestring.
-
rtoy authored
pathname, Adjust PARSE-NAMESTRING to accept pathname designators.
-
- Sep 24, 2005
-
-
rtoy authored
pathname object.
-