- cmucl-cvs - mailman3.common-lisp.net

[Git][cmucl/cmucl] Pushed new branch issue-485-import-openlibm-single-float-functions
by Raymond Toy (＠rtoy) 01 Mar '26

01 Mar '26

Raymond Toy pushed new branch issue-485-import-openlibm-single-float-functions at cmucl / cmucl -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/tree/issue-485-import-openlibm… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-478-run-analyzer-link-chcker-on-schedule] Allow linux:install job to run when scheduled
by Raymond Toy (＠rtoy) 01 Mar '26

01 Mar '26

Raymond Toy pushed to branch issue-478-run-analyzer-link-chcker-on-schedule at cmucl / cmucl Commits: 874302fc by Raymond Toy at 2026-02-28T18:16:37-08:00 Allow linux:install job to run when scheduled Updated so that linux:install is run because linux:build needs it, which is needed by the analyzer when run on a schedule. - - - - - 1 changed file: - .gitlab-ci.yml Changes: ===================================== .gitlab-ci.yml ===================================== @@ -129,6 +129,15 @@ linux:install: variables: osname: "linux" CURL: "curl" + rules: + # 1. Ensure build runs for schedules so the analyzer can use its files + - if: $CI_PIPELINE_SOURCE == "schedule" + when: on_success + # 2. Ensure build runs if the analyzer variable is set + - if: $RUN_ANALYSIS + when: on_success + # 3. Default: run for standard pushes/MRs + - when: on_success linux:build: <<: *build_configuration View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/874302fc9fd3018fe49794c… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/874302fc9fd3018fe49794c… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-478-run-analyzer-link-chcker-on-schedule] Allow linux:build job to run when scheduled
by Raymond Toy (＠rtoy) 01 Mar '26

01 Mar '26

Raymond Toy pushed to branch issue-478-run-analyzer-link-chcker-on-schedule at cmucl / cmucl Commits: dc66629a by Raymond Toy at 2026-02-28T18:15:01-08:00 Allow linux:build job to run when scheduled Updated so that linux:build is run when the analyzer is run. - - - - - 1 changed file: - .gitlab-ci.yml Changes: ===================================== .gitlab-ci.yml ===================================== @@ -141,6 +141,15 @@ linux:build: # Fedora 41 must build with gcc because the clang build fails some # ansi-tests. See [#469]. CONFIG: "x86_linux" + rules: + # 1. Ensure build runs for schedules so the analyzer can use its files + - if: $CI_PIPELINE_SOURCE == "schedule" + when: on_success + # 2. Ensure build runs if the analyzer variable is set + - if: $RUN_ANALYSIS + when: on_success + # 3. Default: run for standard pushes/MRs + - when: on_success linux:cross-build: stage: build View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/dc66629acb3e2acde7ad627… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/dc66629acb3e2acde7ad627… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-478-run-analyzer-link-chcker-on-schedule] Allow link-checker to run on a schedule or if variable is set.
by Raymond Toy (＠rtoy) 01 Mar '26

01 Mar '26

Raymond Toy pushed to branch issue-478-run-analyzer-link-chcker-on-schedule at cmucl / cmucl Commits: 979ad570 by Raymond Toy at 2026-02-28T18:07:38-08:00 Allow link-checker to run on a schedule or if variable is set. - - - - - 1 changed file: - .gitlab-ci.yml Changes: ===================================== .gitlab-ci.yml ===================================== @@ -287,9 +287,6 @@ linux:static-analyzer: tags: - linux - # Allows the pipeline to proceed and show a "passed with warnings" - # status (orange) instead of "failed" (red) if this job exits with an error. - allow_failure: true artifacts: when: always paths: @@ -306,7 +303,9 @@ linux:static-analyzer: when: on_success # 3. Otherwise, the job remains a manual button in the pipeline. - when: manual - allow_failure: true + # Allows the pipeline to proceed and show a "passed with warnings" + # status (orange) instead of "failed" (red) if this job exits with an error. + allow_failure: true script: # Analysis can generate huge amounts of output. For now just save # the results to the log file instead of also having it go to the @@ -366,7 +365,15 @@ markdown-link-check: # Only the linux runner has markdown-link-check installed tags: - linux - when: manual + rules: + # 1. If the pipeline is a scheduled run, run this job automatically + - if: $CI_PIPELINE_SOURCE == "schedule" + when: on_success + # 2. If the variable is set (manually via UI/API), run automatically. + - if: $RUN_ANALYSIS + when: on_success + # 3. Otherwise, the job remains a manual button in the pipeline. + - when: manual # It's ok if this fails; we don't want to declare the entire # pipeline as having failed. allow_failure: true View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/979ad5700d428e718273fcb… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/979ad5700d428e718273fcb… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl] Pushed new branch issue-478-run-analyzer-link-chcker-on-schedule
by Raymond Toy (＠rtoy) 01 Mar '26

01 Mar '26

Raymond Toy pushed new branch issue-478-run-analyzer-link-chcker-on-schedule at cmucl / cmucl -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/tree/issue-478-run-analyzer-li… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][master] 2 commits: Fix #479: Clean up macros in irrat.c
by Raymond Toy (＠rtoy) 28 Feb '26

28 Feb '26

Raymond Toy pushed to branch master at cmucl / cmucl Commits: a225bbb5 by Raymond Toy at 2026-02-28T09:36:55-08:00 Fix #479: Clean up macros in irrat.c - - - - - 1a961701 by Raymond Toy at 2026-02-28T09:36:55-08:00 Merge branch 'issue-479-clean-up-irrat-signal-macros' into 'master' Fix #479: Clean up macros in irrat.c Closes #479 See merge request cmucl/cmucl!359 - - - - - 1 changed file: - src/lisp/irrat.c Changes: ===================================== src/lisp/irrat.c ===================================== @@ -62,20 +62,26 @@ extern void cr_sincosf(float, float *, float *); */ #define MAYBE_SIGNAL_INVALID(test, val) \ - if ((test)) { \ - return fdlibm_setexception(val, FDLIBM_INVALID); \ - } + do { \ + if ((test)) { \ + return fdlibm_setexception(val, FDLIBM_INVALID); \ + } \ + } while (0) + #define MAYBE_SIGNAL_OVERFLOW(x) \ - if (isinf(x)) { \ - return fdlibm_setexception(x, FDLIBM_OVERFLOW); \ - } + do { \ + if (isinf(x)) { \ + return fdlibm_setexception(x, FDLIBM_OVERFLOW); \ + } \ + } while (0) + double lisp_sin(double x) { #ifdef FEATURE_CORE_MATH - MAYBE_SIGNAL_INVALID(isinf(x), x) + MAYBE_SIGNAL_INVALID(isinf(x), x); return cr_sin(x); #else @@ -87,7 +93,7 @@ double lisp_cos(double x) { #ifdef FEATURE_CORE_MATH - MAYBE_SIGNAL_INVALID(isinf(x), x) + MAYBE_SIGNAL_INVALID(isinf(x), x); return cr_cos(x); #else @@ -99,7 +105,7 @@ double lisp_tan(double x) { #ifdef FEATURE_CORE_MATH - MAYBE_SIGNAL_INVALID(isinf(x), x) + MAYBE_SIGNAL_INVALID(isinf(x), x); return cr_tan(x); #else @@ -151,7 +157,7 @@ double lisp_sinh(double x) { #ifdef FEATURE_CORE_MATH - MAYBE_SIGNAL_OVERFLOW(x) + MAYBE_SIGNAL_OVERFLOW(x); return cr_sinh(x); #else @@ -183,7 +189,7 @@ double lisp_asinh(double x) { #ifdef FEATURE_CORE_MATH - MAYBE_SIGNAL_OVERFLOW(x) + MAYBE_SIGNAL_OVERFLOW(x); return cr_asinh(x); #else @@ -195,9 +201,9 @@ double lisp_acosh(double x) { #ifdef FEATURE_CORE_MATH - MAYBE_SIGNAL_INVALID(x < 1, x) + MAYBE_SIGNAL_INVALID(x < 1, x); - MAYBE_SIGNAL_OVERFLOW(x) + MAYBE_SIGNAL_OVERFLOW(x); return cr_acosh(x); #else View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/compare/8034ae1f395b02974a33e8… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/compare/8034ae1f395b02974a33e8… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl] Pushed new branch issue-479-clean-up-irrat-signal-macros
by Raymond Toy (＠rtoy) 28 Feb '26

28 Feb '26

Raymond Toy pushed new branch issue-479-clean-up-irrat-signal-macros at cmucl / cmucl -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/tree/issue-479-clean-up-irrat-… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-435-add-core-math-lisp-support] For darwin, explicitly signal overflow in lisp_expf
by Raymond Toy (＠rtoy) 28 Feb '26

28 Feb '26

Raymond Toy pushed to branch issue-435-add-core-math-lisp-support at cmucl / cmucl Commits: 2ce06664 by Raymond Toy at 2026-02-28T07:42:43-08:00 For darwin, explicitly signal overflow in lisp_expf On darwin, lisp_expf isn't signaling overflow when the arg is too large. To fix this, save the value from __ieee754_exp. Then check to see it's too large to fit in a single-float and signal an overflow if so. Otherwise, just convert the value to a float and return it. This should fix ansi-test failures for EXP.ERROR.4 and EXP.ERROR.5 on darwin. We'll probably need to do the same thing for the other single-float functions that can overflow on Darwin. - - - - - 1 changed file: - src/lisp/irrat.c Changes: ===================================== src/lisp/irrat.c ===================================== @@ -73,6 +73,16 @@ extern float cr_log2f(float); return fdlibm_setexception(x, FDLIBM_OVERFLOW); \ } +#ifdef DARWIN +#define DARWIN_SINGLE_FLOAT_OVERFLOW(y) \ + do { \ + if (fabs(y) >= 0x1.0p128) { \ + return fdlibm_setexception(y, FDLIBM_OVERFLOW); \ + } \ + } while (0) + +#endif + double lisp_sin(double x) { @@ -483,9 +493,17 @@ lisp_expf(float x) { #ifdef FEATURE_CORE_MATH return cr_expf(x); -#else +#else +#if defined(DARWIN) + double y = __ieee754_exp((double) x); + + DARWIN_SINGLE_FLOAT_OVERFLOW(y); + + return (float) y; +#else return (float) __ieee754_exp((double) x); #endif +#endif } float View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/2ce06664397eeaa5de65570… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/2ce06664397eeaa5de65570… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-435-add-core-math-lisp-support] Revert "Use the same floating-point-modes for all x86."
by Raymond Toy (＠rtoy) 28 Feb '26

28 Feb '26

Raymond Toy pushed to branch issue-435-add-core-math-lisp-support at cmucl / cmucl Commits: b2c03d5d by Raymond Toy at 2026-02-27T19:16:48-08:00 Revert "Use the same floating-point-modes for all x86." This reverts commit 4e342cc5337cbed20601b8aca7e3d37fe86f9e7d. - - - - - 1 changed file: - src/code/float-trap.lisp Changes: ===================================== src/code/float-trap.lisp ===================================== @@ -92,7 +92,7 @@ (setf (x87-floating-point-modes) x87-modes))) ) -#+(and sse2) +#+(and sse2 (not darwin)) (progn (defun floating-point-modes () ;; Combine the modes from the FPU and SSE2 units. Since the sse @@ -133,7 +133,7 @@ (setf (vm::x87-floating-point-modes) (ldb (byte 30 0) x87-modes))) new-mode)) -#+(and nil sse2 darwin) +#+(and sse2 darwin) (progn (defun floating-point-modes () ;; Get just the SSE2 mode bits. View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/b2c03d5dcb8ef100b932dda… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/b2c03d5dcb8ef100b932dda… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0

[Git][cmucl/cmucl][issue-435-add-core-math-lisp-support] Use the same floating-point-modes for all x86.
by Raymond Toy (＠rtoy) 28 Feb '26

28 Feb '26

Raymond Toy pushed to branch issue-435-add-core-math-lisp-support at cmucl / cmucl Commits: 4e342cc5 by Raymond Toy at 2026-02-27T17:18:42-08:00 Use the same floating-point-modes for all x86. Previously, Darwin x86 had special float-point-modes functions. However, with lisp_expf, we weren't signaling overflows because x87 instructions weren't causing traps. Use the same routines for all x86. - - - - - 1 changed file: - src/code/float-trap.lisp Changes: ===================================== src/code/float-trap.lisp ===================================== @@ -92,7 +92,7 @@ (setf (x87-floating-point-modes) x87-modes))) ) -#+(and sse2 (not darwin)) +#+(and sse2) (progn (defun floating-point-modes () ;; Combine the modes from the FPU and SSE2 units. Since the sse @@ -133,7 +133,7 @@ (setf (vm::x87-floating-point-modes) (ldb (byte 30 0) x87-modes))) new-mode)) -#+(and sse2 darwin) +#+(and nil sse2 darwin) (progn (defun floating-point-modes () ;; Get just the SSE2 mode bits. View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/4e342cc5337cbed20601b8a… -- View it on GitLab: https://gitlab.common-lisp.net/cmucl/cmucl/-/commit/4e342cc5337cbed20601b8a… You're receiving this email because of your account on gitlab.common-lisp.net.

1 0