- Sep 04, 2016
-
-
Raymond Toy authored
Fix #27: pathname-match-p infinite recursion When support for search-lists was added to `PATHNAME-MATCH-P`, support for logical pathnames was broken because `PATHNAME-MATCH-P` eventually calls `TRANSLATE-LOGICAL-PATHNAME` which calls `PATHNAME-MATCH-P` with logical pathnames. This caused infinite recursion. So add back the original `PATHNAME-MATCH-P`, but rename to `%PATHNAME-MATCH-P` and use that in `TRANSLATE-LOGICAL-PATHNAME` and friends. Add test for this case too. See merge request !10
-
Raymond Toy authored
-
Raymond Toy authored
When support for search-lists was added to PATHNAME-MATCH-P, support for logical pathnames was broken because PATHNAME-MATCH-P eventually calls TRANSLATE-LOGICAL-PATHNAME which calls PATHNAME-MATCH-P with logical pathnames. This caused infinite recursion. So add back the original PATHNAME-MATCH-P, but rename to %PATHNAME-MATCH-P and use that in TRANSLATE-LOGICAL-PATHNAME and friends. Add test for this case too.
-
- Sep 03, 2016
-
-
Raymond Toy authored
Fix #28: Recursive function definition Convert a `double-double` to a double before coercing to a `single-float` or `double-float`. See merge request !9
-
Raymond Toy authored
-
Raymond Toy authored
The recursive definition comes from %single-float and %double-float trying to coerce a double-double-float to a single or double. Not sure the best place to fix this, but added a special case here for %single-float and %double-float to convert the double-double-float to a double that can then be coerced to the appropriate type. (Could have added a deftransform for coerce to handle double-doubles, but doing it here makes it easier to follow the code.) Verified that x86->x86 and sparc->sparc cross-compiles no longer have the warning. Also verified that sparc->sparc actually cross-compiles and loads successfully and that the result will compile itself successfully.
-
- Aug 25, 2016
-
-
Raymond Toy authored
Fix #25: Handle unicode strings more consistently. When writing a string to the program too few octets were written because strings are now 16-bits wide. To fix this, only write the low 8-bits of each character. This matches what reading does. This pretty much implies that the caller should use STREAM:STRING-ENCODE and STREAM::STRING-DECODE on the strings. Add several tests to verify the expected results. See merge request !8
-
- Aug 21, 2016
-
-
Raymond Toy authored
When writing a string to the program too few octets were written because strings are now 16-bits wide. To fix this, only write the low 8-bits of each character. This matches what reading does. This pretty much implies that the caller should use STREAM:STRING-ENCODE and STREAM::STRING-DECODE on the strings. Add several tests to verify the expected results.
-
- May 30, 2016
-
-
Raymond Toy authored
-
Raymond Toy authored
- May 22, 2016
-
-
Raymond Toy authored
Update compiler option to specify 10.6 as the min version. (I personally no longer have an version earlier than 10.11.) Fix ticket #19.
-
- May 21, 2016
-
-
Raymond Toy authored
-
Raymond Toy authored
This is now handled by the runtime feature :relocatable-stacks.
-
Raymond Toy authored
This is now handled by the runtime feature :relocatable-stacks
-
Raymond Toy authored
-
Raymond Toy authored
This is now handled by the runtime feature :relocatable-stacks
-
Raymond Toy authored
Add this as a runtime feature so that it shows up in *features* and so that the C runtime has FEATURE_RELOCATABLE_STACKS defined. src/code/sparc-svr4-vm.lisp: src/code/x86-vm.lisp: o Make :relocatable-stacks a runtime feature. src/lisp/validate.h o Need to include internals.h here so constants are set up according to internals.h src/lisp/x86-validate-darwin.h: o Use FEATURE_RELOCATABLE_STACKS instead of RELOCATABLE_STACK_START.
-
Raymond Toy authored
Describe better the following: sys::*runtime-features* register-lisp-feature register-lisp-runtime-feature
-
Raymond Toy authored
Typo prevented enabling the relocatable stack for OSX.
-
Raymond Toy authored
-
Raymond Toy authored
-
Raymond Toy authored
Map control and binding stacks wherever the OS wants to put them. See merge request !3
-
- May 14, 2016
-
-
Raymond Toy authored
When -X is given (break) just before quiting the cross-compilation. This allows for some debugging of the cross-compiled result.
-
Raymond Toy authored
src/compiler/float.lisp: o The deftransform coerce was checking for a type of 'float and using %single-float to do the conversion. This is incorrect; it should only apply if the type is 'single-float. tests/issues.lisp o Add test for this. Verified that the test fails on the current snapshot and ix fixed by this change.
-
- May 13, 2016
-
-
Raymond Toy authored
The deftransform for converting an (unsigned-byte 32) to a double-float was using the wrong type of float constant. It should have been a double instead of single. This fix allows sparc to cross-compile itself now.
-
- May 12, 2016
-
-
Raymond Toy authored
-
Raymond Toy authored
-
Raymond Toy authored
-
Raymond Toy authored
-
- May 11, 2016
-
-
Raymond Toy authored
-
Raymond Toy authored
-
Raymond Toy authored
-
Raymond Toy authored
When RELOCATABLE_STACK_START is defined, the control stack, binding stack, and sigalt stack are located wherever mmap finds room for them. Only implemented with sparc_sunc config.
-
Raymond Toy authored
We want #ifndef, not #ifdef to enable the relocatable stacks.
-
Raymond Toy authored
When RELOCATABLE_STACK_START is defined, the control stack, binding stack, and sigalt stack are located wherever mmap finds room for them.
-
- May 10, 2016
-
-
Raymond Toy authored
When RELOCATABLE_STACK_START is defined, the control stack, binding stack, and sigalt stack are located wherever mmap finds room for them. Currently only implemented for Darwin, but should work for linux and solaris. We enable this by default on Darwin now.
-
- May 08, 2016
-
-
Raymond Toy authored
-
- Mar 26, 2016
-
-
Raymond Toy authored
-