debian build fix, +verbose "make check"

GNUmakefile

@@ -2,6 +2,9 @@ export PATH := .:$(PATH)
 $(shell mkdir -p config)
 export OL_HOME=libraries
 
+# don't spam "Entering/Leaving directory"
+MAKEFLAGS += --no-print-directory
+
 # default target
 all: release
 
@@ -25,9 +28,15 @@ MGCC32 ?= i686-w64-mingw32-gcc
 MGCC64?=x86_64-w64-mingw32-gcc
 
 # ansi colors
+ifeq ($(MAKE_TERMOUT),)
+red=
+green=
+done=
+else
 red=\033[1;31m
 green=\033[1;32m
 done=\033[0m
+endif
 
 # check submodules
 # ----------------

debian/Makefile

@@ -9,7 +9,11 @@ else
 ppa:
 	git checkout debian/changelog
 	sed 's/UNRELEASED/$(UNRELEASED)/g' -i debian/changelog
-
+
+	git checkout debian/control
+	# fix fuxking debhelper-compact for fuxking ubuntu build system
+	test "$(UNRELEASED)" = "xenial" && sed -i '/debhelper-compat/c\Build-Depends: debhelper-compat (= 9)' debian/control; true
+
 	rm -f debian/files
 	dpkg-buildpackage -us -uc -ui -S -sd \
 	   -rfakeroot
@@ -27,4 +31,12 @@ ppas:
 	done
 	git checkout debian/changelog
 
+xenial:
+	UNRELEASED=xenial $(MAKE) ppa
+	git checkout debian/changelog
+
+noble:
+	UNRELEASED=noble $(MAKE) ppa
+	git checkout debian/changelog
+
 endif

debian/changelog

@@ -1,4 +1,4 @@
-ol (2.6-UNRELEASED4) UNRELEASED; urgency=medium
+ol (2.6-UNRELEASED5) UNRELEASED; urgency=medium
 
   * 2.6
 

debian/rules

@@ -1,7 +1,9 @@
 #!/usr/bin/make -f
+MAKEFLAGS += --no-print-directory
 
 %:
 	dh $@ -O--no-parallel
 
 override_dh_auto_test:
+	echo | gcc -dM -E -
 	dh_auto_test -O--no-parallel

debian/source/options

@@ -1,5 +1,6 @@
 tar-ignore = ".git"
 tar-ignore = ".vscode"
+tar-ignore = ".github"
 tar-ignore = "examples"
 tar-ignore = "jni"
 tar-ignore = "samples"

doc/reference/bytevectors.md

@@ -19,7 +19,7 @@ Bytevectors are written using the notation `#u8(...)` and function `bytevector`.
                ===  #u8(0 1 2 127 128 255)
 
 ; all values wider than 8 bits will be truncated
-(bytevector 254 255 256 257 258 7778885551232190)
+(bytevector 254 255 256 257 258 53659306755069118)
                ==>  #u8(254 255 0 1 2 190)
 
 ```
@@ -86,8 +86,8 @@ Returns the length of bytevector *bv* in bytes as an exact integer.
 (bytevector-length #u8())  ==>  0
 (bytevector-length #u8(1 2 3))
                            ==>  3
-(bytevector-length (make-bytevector 1000000))
-                           ==>  1000000
+(bytevector-length (make-bytevector 100000))
+                           ==>  100000
 ```
 
 # bytevector-u8-ref

extensions/ffi.c

@@ -273,7 +273,7 @@ ret_t win64_call(word argv[], long argc, void* function, long type);
 	// edx - argc
 	// r8  - function
 	// r9d - type
-__ASM__("win64_call:_win64_call:",  // "int $3",
+__ASM__("win64_call:", "_win64_call:",  // "int $3",
 	"pushq %rbp",
 	"movq  %rsp, %rbp",
 
@@ -327,11 +327,14 @@ __ASM__("win64_call:_win64_call:",  // "int $3",
 	"movsd %xmm0, (%rsp)",
 	"pop   %rax", // можно попать, так как уже пушнули один r9 вверху (оптимизация)
 	"jmp   9b");
+# endif
 
-# else      // System V (unix, linux, osx)
-ret_t nix64_call(word argv[], double ad[], long i, long d, long mask, void* function, long type);
+# if __unix__ || __linux__ || __APPLE__ // System V (unix, linux, osx)
+static __attribute((__naked__))
+ret_t nix64_call(word argv[], double ad[], long i, long d, long mask, void* function, long type)
 	            //    rdi            rsi      edx     ecx       r8          r9           16(rbp)
-__ASM__("nix64_call:_nix64_call:", //"int $3",
+{
+__ASM__(//"int $3",
 	"pushq %rbp",
 	"movq  %rsp, %rbp",
 
@@ -429,6 +432,7 @@ __ASM__("nix64_call:_nix64_call:", //"int $3",
 	"movsd %xmm0, (%rsp)",
 	"popq  %rax", // corresponded push not required (we already pushed %r9)
 	"jmp   9b");
+}
 
 // RDI, RSI, RDX, RCX (R10 in the Linux kernel interface[17]:124), R8, and R9
 // while XMM0, XMM1, XMM2, XMM3, XMM4, XMM5, XMM6 and XMM7 are used for floats
@@ -455,7 +459,7 @@ __ASM__("nix64_call:_nix64_call:", //"int $3",
 // 
 ret_t x86_call(word argv[], long i, void* function, long type);
 	// all variables are in stack
-__ASM__("x86_call:_x86_call:", //"int $3",
+__ASM__("x86_call:", "_x86_call:", //"int $3",
 	"pushl %ebp",
 	"movl  %esp, %ebp",
 
@@ -511,9 +515,12 @@ __ASM__("x86_call:_x86_call:", //"int $3",
 //   don't use r16, r17, r18
 //   don't forget "sudo DevToolsSecurity -enable" when debugging with lldb
 															 // mask is deprecated, todo: remove
+// "static __attribute((__naked__))"" is unsupported, so
 ret_t arm64_call(word argv[], double ad[], long i, long d, char* extra, void* function, long type, long e);
-				//    x0             x1        x2      x3       x4          x5             x6       x7
-__ASM__("arm64_call:", "_arm64_call:", // "brk #0",
+				//    x0             x1        x2      x3        x4           x5             x6        x7
+__ASM__(
+".global arm64_call");
+__ASM__("arm64_call:", "_arm64_call", // "brk #0",
 	"stp  x29, x30, [sp, -16]!", // fp + lr
 	"stp  x10, x26, [sp, -16]!",
 	"mov  x29, sp",
@@ -589,9 +596,7 @@ __ASM__("arm64_call:", "_arm64_call:", // "brk #0",
 	"b 0b",
 "7:", // double (8)
 	"str d0, [sp]",
-	"b 0b"
-);
-
+	"b 0b");
 
 // ------------------------------------------
 #elif __arm__
@@ -2776,20 +2781,17 @@ word* OLVM_ffi(olvm_t* this, word arguments)
 #if  __x86_64__
 # if _WIN64
 		got = win64_call(args, i, function, returntype & 0x3F);
-# else // (__unix__ || __APPLE__)
+# endif
+# if __unix__ || __linux__ || __APPLE__ // System V (unix, linux, osx)
 		got = nix64_call(args, ad, max(i, l), d, fpmask, function, returntype & 0x3F);
 # endif
 
 #elif __i386__
-# if _WIN32
 		// cdecl and stdcall in our case are same, so...
 		got = x86_call(args, i, function, returntype & 0x3F);
-# else // (__unix__ || __APPLE__)
-		got = x86_call(args, i, function, returntype & 0x3F);
-# endif
 
 #elif __aarch64__
-	got = arm64_call(args, ad, i, d, extra, function, returntype & 0x3F, WALIGN(e));
+		got = arm64_call(args, ad, i, d, extra, function, returntype & 0x3F, WALIGN(e));
 
 #elif __arm__
 	// arm calling abi http://infocenter.arm.com/help/topic/com.arm.doc.ihi0042f/IHI0042F_aapcs.pdf
@@ -2802,16 +2804,17 @@ word* OLVM_ffi(olvm_t* this, word arguments)
 #	endif
 
 #elif __mips__
-	got = mips32_call(args, i, function, returntype & 0x3F);
+		got = mips32_call(args, i, function, returntype & 0x3F);
 
 #elif __EMSCRIPTEN__
-	got = asmjs_call(args, fpmask, function, returntype & 0x3F);
+		got = asmjs_call(args, fpmask, function, returntype & 0x3F);
 
 #elif __sparc64__
-	got = call_x(args, i, function, returntype & 0x3F);
+		got = call_x(args, i, function, returntype & 0x3F);
 
 #else // ALL other
-	got = call_x(args, i, function, returntype & 0x3F);
+		got = call_x(args, i, function, returntype & 0x3F);
+
 /*	inline ret_t call_cdecl(word args[], int i, void* function, int type) {
 		CALL(__cdecl);
 	}
@@ -2836,7 +2839,7 @@ word* OLVM_ffi(olvm_t* this, word arguments)
 		break;
 	}*/
 #endif
-	else
+	else // just mirror argument
 		got = *(ret_t *)args;
 
 	// где гарантия, что C и B не поменялись?