From noreply at github.com Thu Jul 1 13:53:17 2021 From: noreply at github.com (Jay Krell) Date: Thu, 01 Jul 2021 04:53:17 -0700 Subject: [M3commit] [modula3/cm3] b53298: m3core: WinTabCon and WinListView implementation a... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: b53298d0158a45e5f067aea2efd5cd298808888c https://github.com/modula3/cm3/commit/b53298d0158a45e5f067aea2efd5cd298808888c Author: Jay Krell Date: 2021-07-01 (Thu, 01 Jul 2021) Changed paths: M m3-libs/m3core/src/win32/m3makefile Log Message: ----------- m3core: WinTabCon and WinListView implementation are (#708) only provided for 32bit; make the interface also limited the same. This is a "temporary" hack, but has lived long (since 2013). see commit f75864f2ebe7a8479efa861afb2643639f5e3373 Auhor: Jay Michael Krell Date: Aug 31 2013 The point of this is so we can export what is declared. So we do not declare stuff that does not exist. From noreply at github.com Thu Jul 1 14:10:34 2021 From: noreply at github.com (Jay Krell) Date: Thu, 01 Jul 2021 05:10:34 -0700 Subject: [M3commit] [modula3/cm3] 40f31a: m3core: Fork Unix.i3 for WIN32 vs. POSIX. (#709) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 40f31a5a075fd71fb2a5be8b5546aae2d16ab5dc https://github.com/modula3/cm3/commit/40f31a5a075fd71fb2a5be8b5546aae2d16ab5dc Author: Jay Krell Date: 2021-07-01 (Thu, 01 Jul 2021) Changed paths: R m3-libs/m3core/src/unix/Common/Unix.i3 M m3-libs/m3core/src/unix/Common/m3makefile A m3-libs/m3core/src/unix/POSIX/Unix.i3 A m3-libs/m3core/src/unix/WIN32/Unix.i3 A m3-libs/m3core/src/unix/WIN32/m3makefile M m3-libs/m3core/src/unix/m3makefile Log Message: ----------- m3core: Fork Unix.i3 for WIN32 vs. POSIX. (#709) POSIX is unchanged. Win32 is somewhat reduced. This is so we do not declare things that are not implemented. From noreply at github.com Thu Jul 1 14:24:33 2021 From: noreply at github.com (Jay Krell) Date: Thu, 01 Jul 2021 05:24:33 -0700 Subject: [M3commit] [modula3/cm3] 2381fa: m3core: Fork Unix.i3 for WIN32 vs. POSIX. (#709) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 2381fac1ad8e409cc727a4cb091cb754c6cda02a https://github.com/modula3/cm3/commit/2381fac1ad8e409cc727a4cb091cb754c6cda02a Author: Jay Krell Date: 2021-07-01 (Thu, 01 Jul 2021) Changed paths: R m3-libs/m3core/src/unix/Common/Unix.i3 M m3-libs/m3core/src/unix/Common/m3makefile A m3-libs/m3core/src/unix/POSIX/Unix.i3 A m3-libs/m3core/src/unix/WIN32/Unix.i3 A m3-libs/m3core/src/unix/WIN32/m3makefile M m3-libs/m3core/src/unix/m3makefile Log Message: ----------- m3core: Fork Unix.i3 for WIN32 vs. POSIX. (#709) POSIX is unchanged. Win32 is somewhat reduced. This is so we do not declare things that are not implemented. From noreply at github.com Fri Jul 2 09:20:19 2021 From: noreply at github.com (Jay Krell) Date: Fri, 02 Jul 2021 00:20:19 -0700 Subject: [M3commit] [modula3/cm3] de6a37: m3core: Layout out IPv6 for OSF/1 compatibility. (... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: de6a370b076ab31287d3982985b65c43a5c59089 https://github.com/modula3/cm3/commit/de6a370b076ab31287d3982985b65c43a5c59089 Author: Jay Krell Date: 2021-07-02 (Fri, 02 Jul 2021) Changed paths: M m3-libs/m3core/src/unix/Common/Uin.i3 M m3-libs/m3core/src/unix/Common/Usocket.c Log Message: ----------- m3core: Layout out IPv6 for OSF/1 compatibility. (#710) In particular in6_addr has 8-alignment. Move unsigned int to before it, and add another for explicit padding for alignment. Otherwise OSF/1 fails static assert here. From noreply at github.com Fri Jul 2 14:50:48 2021 From: noreply at github.com (Jay Krell) Date: Fri, 02 Jul 2021 05:50:48 -0700 Subject: [M3commit] [modula3/cm3] bc0e29: m3core: Remove C99 math support for OSF/1. (#711) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: bc0e29d2d0ac32cb32f7cb1598f5f997bf98e412 https://github.com/modula3/cm3/commit/bc0e29d2d0ac32cb32f7cb1598f5f997bf98e412 Author: Jay Krell Date: 2021-07-02 (Fri, 02 Jul 2021) Changed paths: M m3-libs/m3core/src/float/m3makefile Log Message: ----------- m3core: Remove C99 math support for OSF/1. (#711) OSF/1 C99 math support is incomplete. We should be able to provide most/all of it on top of nonstandard names. See: read_rnd, write_rnd (and compiler flags) ieee_get_fp_control From noreply at github.com Fri Jul 2 14:52:17 2021 From: noreply at github.com (Jay Krell) Date: Fri, 02 Jul 2021 05:52:17 -0700 Subject: [M3commit] [modula3/cm3] aa8e72: m3c: Avoid duplicate typedefs of REAL, LONGREAL, E... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: aa8e722ef717adec9947989d856c209f114dd05b https://github.com/modula3/cm3/commit/aa8e722ef717adec9947989d856c209f114dd05b Author: Jay Krell Date: 2021-07-02 (Fri, 02 Jul 2021) Changed paths: M m3-sys/m3back/src/M3C.m3 Log Message: ----------- m3c: Avoid duplicate typedefs of REAL, LONGREAL, EXTENDED. (#712) This in particular fails with native cc on OSF/1. From noreply at github.com Fri Jul 2 15:01:15 2021 From: noreply at github.com (Jay Krell) Date: Fri, 02 Jul 2021 06:01:15 -0700 Subject: [M3commit] [modula3/cm3] 3477da: m3core/m3c: If _LONGLONG is defined, use long long... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 3477dab7ebc4b64a95924da37a19e5a32b0e41a0 https://github.com/modula3/cm3/commit/3477dab7ebc4b64a95924da37a19e5a32b0e41a0 Author: Jay Krell Date: 2021-07-02 (Fri, 02 Jul 2021) Changed paths: M m3-libs/m3core/src/m3core.h M m3-sys/m3back/src/M3C.m3 Log Message: ----------- m3core/m3c: If _LONGLONG is defined, use long long and LL/ULL. (#713) This fixes OSF/1 which defines __DECC. From noreply at github.com Fri Jul 2 15:29:00 2021 From: noreply at github.com (Jay Krell) Date: Fri, 02 Jul 2021 06:29:00 -0700 Subject: [M3commit] [modula3/cm3] 287240: m3c: Skip some duplicate typedefs. (#714) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 28724078cf59326d9a0b4fbd6f1ff2fdd492a814 https://github.com/modula3/cm3/commit/28724078cf59326d9a0b4fbd6f1ff2fdd492a814 Author: Jay Krell Date: 2021-07-02 (Fri, 02 Jul 2021) Changed paths: M m3-sys/m3back/src/M3C.m3 Log Message: ----------- m3c: Skip some duplicate typedefs. (#714) gcc on OSF/1 in particular errors wrt WIDECHAR. Though g++ and native cc and cxx were ok. From noreply at github.com Fri Jul 2 15:45:40 2021 From: noreply at github.com (Jay Krell) Date: Fri, 02 Jul 2021 06:45:40 -0700 Subject: [M3commit] [modula3/cm3] c5ca10: mingw / juno: Allow to work with forward slashes (... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: c5ca10b335b3dcbad1e109521821ffb484e50b8c https://github.com/modula3/cm3/commit/c5ca10b335b3dcbad1e109521821ffb484e50b8c Author: Jay Krell Date: 2021-07-02 (Fri, 02 Jul 2021) Changed paths: M m3-ui/juno-2/juno-app/src/m3makefile Log Message: ----------- mingw / juno: Allow to work with forward slashes (#715) From noreply at github.com Sat Jul 3 00:01:54 2021 From: noreply at github.com (Jay Krell) Date: Fri, 02 Jul 2021 15:01:54 -0700 Subject: [M3commit] [modula3/cm3] b60fce: MinGW: More forward slashes. (#717) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: b60fcee8dd42f36b5dbc34a0c38a4eb871efe98c https://github.com/modula3/cm3/commit/b60fcee8dd42f36b5dbc34a0c38a4eb871efe98c Author: Jay Krell Date: 2021-07-02 (Fri, 02 Jul 2021) Changed paths: M m3-sys/m3quake/src/M3Path.i3 Log Message: ----------- MinGW: More forward slashes. (#717) Hopefully this doesn't break anything, forward slashes do usually work on Windows, except sometimes the paths of things to run, with cmd. From noreply at github.com Sat Jul 3 01:11:53 2021 From: noreply at github.com (Jay Krell) Date: Fri, 02 Jul 2021 16:11:53 -0700 Subject: [M3commit] [modula3/cm3] 2d4ea9: m3core: Add m3makefile, missed in: (#718) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 2d4ea9e4fe12aacda7af24f3c3aea2af410bcb29 https://github.com/modula3/cm3/commit/2d4ea9e4fe12aacda7af24f3c3aea2af410bcb29 Author: Jay Krell Date: 2021-07-02 (Fri, 02 Jul 2021) Changed paths: A m3-libs/m3core/src/unix/POSIX/m3makefile Log Message: ----------- m3core: Add m3makefile, missed in: (#718) commit 2381fac1ad8e409cc727a4cb091cb754c6cda02a m3core: Fork Unix.i3 for WIN32 vs. POSIX. (#709) From noreply at github.com Sat Jul 3 01:57:55 2021 From: noreply at github.com (Jay Krell) Date: Fri, 02 Jul 2021 16:57:55 -0700 Subject: [M3commit] [modula3/cm3] b2cd50: config: AlphaOsf: Pass -o object to C compiler so ... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: b2cd50a09aa68f3b7094ef6c4dfd3627dc14833f https://github.com/modula3/cm3/commit/b2cd50a09aa68f3b7094ef6c4dfd3627dc14833f Author: Jay Krell Date: 2021-07-02 (Fri, 02 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF Log Message: ----------- config: AlphaOsf: Pass -o object to C compiler so it (#719) honors naming convention, which we probably have too many of as well. Else ar commands fail. From noreply at github.com Sat Jul 3 03:28:17 2021 From: noreply at github.com (Jay Krell) Date: Fri, 02 Jul 2021 18:28:17 -0700 Subject: [M3commit] [modula3/cm3] 42276b: m3core: Default host to ALPHA_OSF, not ALPHA_OSF1.... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 42276bfa81370362bc3d2b81b0d2f461bfdc729f https://github.com/modula3/cm3/commit/42276bfa81370362bc3d2b81b0d2f461bfdc729f Author: Jay Krell Date: 2021-07-02 (Fri, 02 Jul 2021) Changed paths: M m3-libs/m3core/src/m3core.h M m3-sys/m3quake/src/MxConfigC.c Log Message: ----------- m3core: Default host to ALPHA_OSF, not ALPHA_OSF1. (#720) It is hard to see where the "1" comes from. It is probably from calling uname() in MxConfigC.c. Granted, uname does come so close, maybe we should let it be the decider, thereby getting platform names automatically. Or rather output directory names and no platforms per se, just Posixish. From noreply at github.com Sat Jul 3 21:48:04 2021 From: noreply at github.com (Jay Krell) Date: Sat, 03 Jul 2021 12:48:04 -0700 Subject: [M3commit] [modula3/cm3] ab0acc: M3overrides cleanup. (#721) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: ab0accebe764ab62a2382e7345ac5d83bcd016ff https://github.com/modula3/cm3/commit/ab0accebe764ab62a2382e7345ac5d83bcd016ff Author: Jay Krell Date: 2021-07-03 (Sat, 03 Jul 2021) Changed paths: M m3-db/db/src/m3overrides M m3-db/db/test/src/m3overrides A m3-db/mysql/src/m3overrides M m3-db/odbc/src/m3overrides M m3-db/odbc/test/src/m3overrides M m3-db/pgodbc/src/m3overrides M m3-db/postgres95/src/m3overrides M m3-db/postgres95/test/src/m3overrides M m3-db/smalldb/src/m3overrides A m3-db/stable/example/src/m3overrides M m3-db/stable/src/m3overrides A m3-db/stable/test/src/m3overrides M m3-db/stablegen/src/m3overrides Log Message: ----------- M3overrides cleanup. (#721) - add some missing - unify some old ones on the "new" way For now at least, every m3makefile at the top of a package should have an m3override file and every such m3override should be identical, except for the very top one they all delegate to. On one hand this is kinda dumb, since it is all the same, it should be automatic, but I think the idea is maybe to accomodate multiple roots in a forest, not just one tree. Probably that would be a reason to rename ROOT to CM3_ROOT, so there can be other "ROOT"s. This is an area that I don't think even other systems do great on. There is a combination of - configure -with-foo=dir - moving under one root, one build like here (gcc+binutils+gmp, etc.) - install as you go, and find everything in common prefix, which is comman and seems ok until you really think about it, or see it in a setting where it fails 'cause 1. I think the cross build methodology is make install DESTDIR=x and then later tar/untar, so there is no prefix to find during build and 2. You end up running just-built things (if not cross building) which can be untested and fail. For our part we have the one root approach, but we don't practise much the deferred install or proper cross builds, etc. From noreply at github.com Tue Jul 6 20:12:13 2021 From: noreply at github.com (VictorMiasnikov) Date: Tue, 06 Jul 2021 11:12:13 -0700 Subject: [M3commit] [modula3/cm3] d6df0a: AMD64_MINGW m3_link fix by calling subst_chars() (... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: d6df0a2fb33da2af7ba69fac26f801694c5513c4 https://github.com/modula3/cm3/commit/d6df0a2fb33da2af7ba69fac26f801694c5513c4 Author: VictorMiasnikov <48795548+VictorMiasnikov at users.noreply.github.com> Date: 2021-07-06 (Tue, 06 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/AMD64_MINGW Log Message: ----------- AMD64_MINGW m3_link fix by calling subst_chars() (#722) Allow use boot2min.py and boot2.py with target boot2.py From noreply at github.com Wed Jul 7 02:43:05 2021 From: noreply at github.com (Jay Krell) Date: Tue, 06 Jul 2021 17:43:05 -0700 Subject: [M3commit] [modula3/cm3] 081285: OSF1: Fix alloca for Tru64/OSF1 with native cc. (#... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 0812856a8838ae2f4efd9036fbf15495273c9b05 https://github.com/modula3/cm3/commit/0812856a8838ae2f4efd9036fbf15495273c9b05 Author: Jay Krell Date: 2021-07-06 (Tue, 06 Jul 2021) Changed paths: M m3-libs/m3core/src/m3core.h M m3-sys/m3back/src/M3C.m3 Log Message: ----------- OSF1: Fix alloca for Tru64/OSF1 with native cc. (#723) Hypothetical alloca for OpenVMS was breaking it. Disable that "for now". It was already working for gcc. From noreply at github.com Wed Jul 7 02:45:46 2021 From: noreply at github.com (Jay Krell) Date: Tue, 06 Jul 2021 17:45:46 -0700 Subject: [M3commit] [modula3/cm3] 7ea13c: OSF1: Switch Tru64 / OSF1 to C backend "only". (#724) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 7ea13c713c256d07d4703ed889a49bcea8cba6e5 https://github.com/modula3/cm3/commit/7ea13c713c256d07d4703ed889a49bcea8cba6e5 Author: Jay Krell Date: 2021-07-06 (Tue, 06 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF M scripts/python/pylib.py Log Message: ----------- OSF1: Switch Tru64 / OSF1 to C backend "only". (#724) This is config and scripts. The gcc backend did work a long time ago, and very recently also. All of native cc, native cxx, gcc (3.4.6? 4.7?) and g++ should all work. With native cxx, our install is slightly broken and linking requires -oldcxx, just because one copy of one of the startup files is missing. >From a frontend point of view, old is likely Digital-authored and "new" is EDG. Fix path of C compiler in bootstrap. From noreply at github.com Fri Jul 9 00:56:38 2021 From: noreply at github.com (Jay Krell) Date: Thu, 08 Jul 2021 15:56:38 -0700 Subject: [M3commit] [modula3/cm3] 7bb8ea: m3override: add all of m3-db (#725) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 7bb8ea799eea8ff2d5ef965cf9842a6ced879d68 https://github.com/modula3/cm3/commit/7bb8ea799eea8ff2d5ef965cf9842a6ced879d68 Author: Jay Krell Date: 2021-07-08 (Thu, 08 Jul 2021) Changed paths: M m3overrides Log Message: ----------- m3override: add all of m3-db (#725) Followup to ab0accebe764ab62a2382e7345ac5d83bcd016ff From noreply at github.com Fri Jul 9 18:27:12 2021 From: noreply at github.com (Jay Krell) Date: Fri, 09 Jul 2021 09:27:12 -0700 Subject: [M3commit] [modula3/cm3] 90b078: scripts: Fix fatal typos in 7ea13c713c256d07d4703e... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 90b07893570f09640adbe827a7bb25a2b9bf3179 https://github.com/modula3/cm3/commit/90b07893570f09640adbe827a7bb25a2b9bf3179 Author: Jay Krell Date: 2021-07-09 (Fri, 09 Jul 2021) Changed paths: M scripts/python/pylib.py Log Message: ----------- scripts: Fix fatal typos in 7ea13c713c256d07d4703ed889a49bcea8cba6e5 (#726) From noreply at github.com Mon Jul 12 13:28:55 2021 From: noreply at github.com (Jay Krell) Date: Mon, 12 Jul 2021 04:28:55 -0700 Subject: [M3commit] [modula3/cm3] 7a6faa: config: Remove OSF 4 vs. 5 SYSTEM_CC configuration... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 7a6faa376c86a94f4931fa50e10f453841c12e47 https://github.com/modula3/cm3/commit/7a6faa376c86a94f4931fa50e10f453841c12e47 Author: Jay Krell Date: 2021-07-12 (Mon, 12 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF Log Message: ----------- config: Remove OSF 4 vs. 5 SYSTEM_CC configuration. (#727) This is not to drop OSF 4 support per se, but the logic looks wrong and I suspect moot. That is: - I don't think OSF has the bad default ELF interposability inefficient semantics. - It does have something, but I think not as bad. - I don't think this syntax is related. - Even if it does, so many other people survive with it, we probably can do. And OSF4 vs. 5 was only about controlling that. I don't remember if I tested on OSF 4 but lately I am using OSF 5. From noreply at github.com Mon Jul 12 13:30:26 2021 From: noreply at github.com (Jay Krell) Date: Mon, 12 Jul 2021 04:30:26 -0700 Subject: [M3commit] [modula3/cm3] feb7a9: config: OSF: Remove configure_c_compiler. (#728) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: feb7a978200f3d5a6a268ceaf3a39eee195d4e2f https://github.com/modula3/cm3/commit/feb7a978200f3d5a6a268ceaf3a39eee195d4e2f Author: Jay Krell Date: 2021-07-12 (Mon, 12 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF Log Message: ----------- config: OSF: Remove configure_c_compiler. (#728) It is unconditional now. From noreply at github.com Mon Jul 12 13:35:27 2021 From: noreply at github.com (Jay Krell) Date: Mon, 12 Jul 2021 04:35:27 -0700 Subject: [M3commit] [modula3/cm3] 5f359b: config: OSF: Require C99. (#729) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 5f359b131ed1a3b8de96d512fd9bffb8d63850df https://github.com/modula3/cm3/commit/5f359b131ed1a3b8de96d512fd9bffb8d63850df Author: Jay Krell Date: 2021-07-12 (Mon, 12 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF Log Message: ----------- config: OSF: Require C99. (#729) From noreply at github.com Mon Jul 12 13:38:15 2021 From: noreply at github.com (Jay Krell) Date: Mon, 12 Jul 2021 04:38:15 -0700 Subject: [M3commit] [modula3/cm3] 3a0f4f: config: osf: Use dynamic floating point rounding. ... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 3a0f4fc84d1deba7107d4a1ac165d0c0d09f43c9 https://github.com/modula3/cm3/commit/3a0f4fc84d1deba7107d4a1ac165d0c0d09f43c9 Author: Jay Krell Date: 2021-07-12 (Mon, 12 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF Log Message: ----------- config: osf: Use dynamic floating point rounding. (#730) I believe this is what Modula-3 really wants. From noreply at github.com Mon Jul 12 13:46:54 2021 From: noreply at github.com (Jay Krell) Date: Mon, 12 Jul 2021 04:46:54 -0700 Subject: [M3commit] [modula3/cm3] 4b251f: config: OSF: Document gcc near equivalent. (#731) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 4b251f88eb067ad6f5370b9b7a3a1f52adb66afd https://github.com/modula3/cm3/commit/4b251f88eb067ad6f5370b9b7a3a1f52adb66afd Author: Jay Krell Date: 2021-07-12 (Mon, 12 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF Log Message: ----------- config: OSF: Document gcc near equivalent. (#731) From noreply at github.com Mon Jul 12 14:01:38 2021 From: noreply at github.com (Jay Krell) Date: Mon, 12 Jul 2021 12:01:38 +0000 (UTC) Subject: [M3commit] [modula3/cm3] 14a2e0: config: OSF: Switch to C++. (#732) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 14a2e049735690200a725e60afa9dfe15b62eba0 https://github.com/modula3/cm3/commit/14a2e049735690200a725e60afa9dfe15b62eba0 Author: Jay Krell Date: 2021-07-12 (Mon, 12 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF M scripts/python/pylib.py Log Message: ----------- config: OSF: Switch to C++. (#732) Use /usr/bin/cxx -x cxx should work. Link with -oldcxx because of an install problem. From noreply at github.com Wed Jul 14 07:14:56 2021 From: noreply at github.com (Jay Krell) Date: Tue, 13 Jul 2021 22:14:56 -0700 Subject: [M3commit] [modula3/cm3] 394732: Revert "Bootstrap: Remove the -o and -Fo./ flags a... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 3947327bc615a2e27adb15a33b3213d798918912 https://github.com/modula3/cm3/commit/3947327bc615a2e27adb15a33b3213d798918912 Author: Jay Krell Date: 2021-07-13 (Tue, 13 Jul 2021) Changed paths: M scripts/python/pylib.py Log Message: ----------- Revert "Bootstrap: Remove the -o and -Fo./ flags as we are" (#733) This reverts commit 42a53c5514ae5492048496123536b7df584b0762. This will be built upon shortly. The only reason was seemingly cosmetic, but now the opposite has a purpose. From noreply at github.com Wed Jul 14 08:51:56 2021 From: noreply at github.com (Jay Krell) Date: Tue, 13 Jul 2021 23:51:56 -0700 Subject: [M3commit] [modula3/cm3] a0cd68: scripts: Linking with -oldcxx breaks compiling Mai... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: a0cd68e52c901a655eb4cc098833ddb331e0cd8d https://github.com/modula3/cm3/commit/a0cd68e52c901a655eb4cc098833ddb331e0cd8d Author: Jay Krell Date: 2021-07-13 (Tue, 13 Jul 2021) Changed paths: M scripts/python/pylib.py Log Message: ----------- scripts: Linking with -oldcxx breaks compiling Main on OSF. (#734) It has no long long, or no int64, or something. Probably neither. Workaround by having .o as dependency. This depends on -o being restored. And add the equivalent -Fo to the other NT platforms. Add -o for OSF and Solaris. This begs the question of portability to older compilers. Really we should use autoconf and check sizef(long) == 8. long long is not really needed on most 64bit targets. Or maybe we should just require C99/stdint.h and be done with a lot of cruft. We can still link with -oldcxx. From noreply at github.com Wed Jul 14 09:45:17 2021 From: noreply at github.com (Jay Krell) Date: Wed, 14 Jul 2021 00:45:17 -0700 Subject: [M3commit] [modula3/cm3] d34231: config: OSF: display the short strings used to dis... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: d342316ed776a6b00bd0e72abf7015e794238cc5 https://github.com/modula3/cm3/commit/d342316ed776a6b00bd0e72abf7015e794238cc5 Author: Jay Krell Date: 2021-07-14 (Wed, 14 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF Log Message: ----------- config: OSF: display the short strings used to disable warnings (#735) From noreply at github.com Wed Jul 14 09:59:34 2021 From: noreply at github.com (Jay Krell) Date: Wed, 14 Jul 2021 00:59:34 -0700 Subject: [M3commit] [modula3/cm3] 743349: config: OSF: Suppress a few warnings. (#736) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 743349b3c34af56635b11c3593e6072cc4ac7af8 https://github.com/modula3/cm3/commit/743349b3c34af56635b11c3593e6072cc4ac7af8 Author: Jay Krell Date: 2021-07-14 (Wed, 14 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF Log Message: ----------- config: OSF: Suppress a few warnings. (#736) C backend produces these: codeunreachable: statement is unreachable badansialias1: This statement uses the type to reference a storage location. badansialias2: This statement uses the type to reference the same storage location. fprldonot: IEEE dynamic rounding mode is not supported for long doubles on this platform labelnotreach: code can never be executed at label missingreturn: missing return statement at end of non-void function underflow: underflow occurs in evaluating this expression Ideally we would work toward producing code that does not produce any of these warnings, but it can be a fair amount of work, for later/never. Regarding underflow, we should generate IEEE754 bytes, not floats. Including both endian and runtime selection. Regarding rounding: We do not even use long doubles. We get this warning because we are being even kind of advanced and anticipating putting in the C99 support to change rounding mode, which probably nobody calls. missing return is probably in functions with infinite loops. Augment comment about -oldcxx linking. From noreply at github.com Thu Jul 15 00:32:59 2021 From: noreply at github.com (Rodney M. Bates) Date: Wed, 14 Jul 2021 15:32:59 -0700 Subject: [M3commit] [modula3/cm3] c2ac41: Comment some confusing things regarding alignment ... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: c2ac41d6dd6cf68c05fa53f0fe8be1000b37fb7e https://github.com/modula3/cm3/commit/c2ac41d6dd6cf68c05fa53f0fe8be1000b37fb7e Author: Rodney Bates Date: 2021-07-14 (Wed, 14 Jul 2021) Changed paths: M m3-libs/m3core/src/runtime/common/RT0.i3 M m3-libs/m3core/src/runtime/common/RTAllocator.m3 M m3-libs/m3core/src/runtime/common/RTType.m3 Log Message: ----------- Comment some confusing things regarding alignment of heap objects. Commit: cc9120c816c1c29effe07734fc8a293cf6f76a58 https://github.com/modula3/cm3/commit/cc9120c816c1c29effe07734fc8a293cf6f76a58 Author: Rodney Bates Date: 2021-07-14 (Wed, 14 Jul 2021) Changed paths: M m3-sys/cminstall/src/config-no-install/ALPHA_OSF M scripts/python/pylib.py Log Message: ----------- Merge branch 'master' of https://github.com/modula3/cm3 Get recent updates before commit of local changes. Commit: fd4e37b369028381a4789f7629c4f87bbbda452b https://github.com/modula3/cm3/commit/fd4e37b369028381a4789f7629c4f87bbbda452b Author: Rodney Bates Date: 2021-07-14 (Wed, 14 Jul 2021) Changed paths: M m3-sys/m3front/src/builtinOps/New.m3 M m3-sys/m3front/src/exprs/Expr.m3 M m3-sys/m3front/src/exprs/QualifyExpr.m3 M m3-sys/m3front/src/exprs/TypeExpr.m3 M m3-sys/m3front/src/types/ObjectType.i3 M m3-sys/m3front/src/types/ObjectType.m3 M m3-sys/m3front/src/types/OpaqueType.m3 M m3-sys/m3front/src/types/RefType.m3 Log Message: ----------- More adaptation sparked by 64-bit alignments on 32-bit targets. Much of it response to the surprising (to me) discovery that not all traced allocations have the same alignment. Comments added to RT0.i3 summarize the confusing behavior. After circumventing some apparently leftover stuff supporting any of 8, 16, 32, or 64, all allocations end up aligned at least to a target word, but on a 32-bit target, some can be 64-bit aligned. Some comments and renames to save the next maintainer some of the difficult vetting this required. Also add some language section references to some compiler messages. Compare: https://github.com/modula3/cm3/compare/743349b3c34a...fd4e37b36902 From noreply at github.com Sat Jul 17 01:31:55 2021 From: noreply at github.com (Jay Krell) Date: Fri, 16 Jul 2021 16:31:55 -0700 Subject: [M3commit] [modula3/cm3] 2ab0eb: scripts: Python2 to Python3 conversion. (#737) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 2ab0eb8a6079f803bc6b6bdc3ab0e6a171396ccb https://github.com/modula3/cm3/commit/2ab0eb8a6079f803bc6b6bdc3ab0e6a171396ccb Author: Jay Krell Date: 2021-07-16 (Fri, 16 Jul 2021) Changed paths: M scripts/python/pylib.py Log Message: ----------- scripts: Python2 to Python3 conversion. (#737) This should work either way. From noreply at github.com Sun Jul 18 13:05:43 2021 From: noreply at github.com (Jay Krell) Date: Sun, 18 Jul 2021 04:05:43 -0700 Subject: [M3commit] [modula3/cm3] 77141c: scripts: More Python3 compatibility. (#738) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 77141c999981d6b725aec2e263a935d2b9cb41f8 https://github.com/modula3/cm3/commit/77141c999981d6b725aec2e263a935d2b9cb41f8 Author: Jay Krell Date: 2021-07-18 (Sun, 18 Jul 2021) Changed paths: M scripts/python/pylib.py Log Message: ----------- scripts: More Python3 compatibility. (#738) This makes the code Python2 and Python3 and have the same meaning. Previously it would run under Python3 but have much different meaning and fail early. From noreply at github.com Tue Jul 20 06:29:04 2021 From: noreply at github.com (peter mckinna) Date: Mon, 19 Jul 2021 21:29:04 -0700 Subject: [M3commit] [modula3/cm3] 6c46e4: Minor fixes Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 6c46e48e1a31899dce768030e2c82b38ee7b20b3 https://github.com/modula3/cm3/commit/6c46e48e1a31899dce768030e2c82b38ee7b20b3 Author: peter mckinna Date: 2021-07-20 (Tue, 20 Jul 2021) Changed paths: M m3-mail/postcard/src/MiscUtils.m3 M m3-mail/postcard/src/OSUtils.m3 M m3-mail/postcard/src/m3makefile Log Message: ----------- Minor fixes Commit: 3afa6768e889ac3614329399fd7b67162bf21a54 https://github.com/modula3/cm3/commit/3afa6768e889ac3614329399fd7b67162bf21a54 Author: peter mckinna Date: 2021-07-20 (Tue, 20 Jul 2021) Changed paths: M m3-mail/postcard/src/MiscUtils.m3 Log Message: ----------- Fix compile probs Commit: ae2eba2ed017afad81ae151172dc2ebcc42f506b https://github.com/modula3/cm3/commit/ae2eba2ed017afad81ae151172dc2ebcc42f506b Author: peter mckinna Date: 2021-07-20 (Tue, 20 Jul 2021) Changed paths: M m3-sys/m3front/src/misc/Host.m3 Log Message: ----------- Ensure check-proc is disabled when compile with -X-NoChecks Commit: 8a842be10f0e5d1f1acf73597d13765aea305573 https://github.com/modula3/cm3/commit/8a842be10f0e5d1f1acf73597d13765aea305573 Author: peter mckinna Date: 2021-07-20 (Tue, 20 Jul 2021) Changed paths: M m3-sys/m3front/src/exprs/QualifyExpr.m3 Log Message: ----------- Add donilcheck check before doing a nil check Compare: https://github.com/modula3/cm3/compare/77141c999981...8a842be10f0e From noreply at github.com Tue Jul 20 14:47:03 2021 From: noreply at github.com (Jay Krell) Date: Tue, 20 Jul 2021 05:47:03 -0700 Subject: [M3commit] [modula3/cm3] f2c0ec: m3-sys: Start adding .cpp support, in addition to ... Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: f2c0ec19e0ac5742ea680d8d244e22c48f7a6b17 https://github.com/modula3/cm3/commit/f2c0ec19e0ac5742ea680d8d244e22c48f7a6b17 Author: Jay Krell Date: 2021-07-20 (Tue, 20 Jul 2021) Changed paths: M m3-sys/cm3/src/Builder.m3 M m3-sys/cm3/src/M3Build.m3 M m3-sys/cm3/src/Makefile.m3 M m3-sys/cm3ide/src/nodes/Source.i3 M m3-sys/cm3ide/src/nodes/Source.m3 M m3-sys/m3quake/src/M3Path.i3 M m3-sys/m3quake/src/M3Path.m3 Log Message: ----------- m3-sys: Start adding .cpp support, in addition to .c support. (#739) There are 2-3 reasons for this. - It should let us get out of the business of knowing the flag to the C compiler makes it compile as C++. i.e. gcc -xc++, osf cxx -x cxx, msvc cl -TP - It works toward an "escape hatch" of having separate C and C++ compilation, which really I don't think is really needed though. Which again might not need to know the xc++/xcxx/TP switches. - In producing "GNU build system" (autoconf/automake/libtool), it leads to C++ compilation, but really this is the same as the first point, and we could rename at that point, if it is in scripts/python (maybe, maybe not) Generate _m3main.cpp instead of _m3main.c. Fix warning about unused import. Produce .cpp from C backend. And to be clear, the primary benefit of C++ to cm3 is theorized to be efficient portable exception handling. As opposed to the highly inefficient scheme we have now. Possibly even more portable, like to systems without threads or thread locals. For ease of reading, cpp comes after c; not after h, not at the end. cxx would also be ok; cpp can mean C preprocesor, as CPP vs. CXX, CPPFLAGS vs. CXXFLAGS, etc. From noreply at github.com Wed Jul 28 11:13:59 2021 From: noreply at github.com (VictorMiasnikov) Date: Wed, 28 Jul 2021 02:13:59 -0700 Subject: [M3commit] [modula3/cm3] 296949: Minor fixes in cm3ide nodes/Source.m3 (#740) Message-ID: Branch: refs/heads/master Home: https://github.com/modula3/cm3 Commit: 296949912237e53f6f6bc0daaa90c8b95088db45 https://github.com/modula3/cm3/commit/296949912237e53f6f6bc0daaa90c8b95088db45 Author: VictorMiasnikov <48795548+VictorMiasnikov at users.noreply.github.com> Date: 2021-07-28 (Wed, 28 Jul 2021) Changed paths: M m3-sys/cm3ide/src/nodes/Source.m3 Log Message: ----------- Minor fixes in cm3ide nodes/Source.m3 (#740) Minor fixes in cm3ide nodes/Source.m3