From roart at nvg.ntnu.no Tue May 5 07:53:57 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Tue, 5 May 2015 07:53:57 +0200 Subject: [FreeVMS] More questions In-Reply-To: References: Message-ID: <20150505055357.GB18627@sabre-wulf.nvg.ntnu.no> On Wed, Mar 25, 2015 at 07:15:18AM +0100, Guido de Jong wrote: > > Some more questions: And one from me. Debian 8 is out. Have you started to do anything with building under that? I have barely started, though. -- Regards, Roar From roart at nvg.ntnu.no Tue May 5 07:53:57 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Tue, 5 May 2015 07:53:57 +0200 Subject: [FreeVMS] More questions In-Reply-To: References: Message-ID: <20150505055357.GB18627@sabre-wulf.nvg.ntnu.no> On Wed, Mar 25, 2015 at 07:15:18AM +0100, Guido de Jong wrote: > > Some more questions: And one from me. Debian 8 is out. Have you started to do anything with building under that? I have barely started, though. -- Regards, Roar From guidoj2269 at gmail.com Tue May 5 23:11:43 2015 From: guidoj2269 at gmail.com (Guido) Date: Tue, 05 May 2015 23:11:43 +0200 Subject: [FreeVMS] More questions In-Reply-To: <20150505055357.GB18627@sabre-wulf.nvg.ntnu.no> References: <20150505055357.GB18627@sabre-wulf.nvg.ntnu.no> Message-ID: <5549320F.90601@gmail.com> On 05/05/2015 07:53 AM, Roar Thronæs wrote: > On Wed, Mar 25, 2015 at 07:15:18AM +0100, Guido de Jong wrote: >> Some more questions: > And one from me. > > Debian 8 is out. > Have you started to do anything with building under that? > I have barely started, though. > No, I'm still on Wheezy. One of the reasons I brought up cross compiling was to be less dependent on linux distribution/version/etc. Gave up on that for now, too many easier tasks. At present I'm expanding my system services test. I intend to use it as a regression test for all future changes. It can be loaded by grub and executed in stead of dcl. You have to edit the boot menu entry each time you boot into freevms though, because I have no idea how to add a menu entry on the boot floppy. Anyway, I ran into the fact that the getsyi and getjpi system services do not return the length of the data in the item list. So I fixed that locally and I will try to commit it later this week. Maybe I'll try to upgrade to Jessie next ... Regards, Guido From guidoj2269 at gmail.com Sun May 10 23:33:21 2015 From: guidoj2269 at gmail.com (Guido de Jong) Date: Sun, 10 May 2015 23:33:21 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: References: Message-ID: Hi, The short answer to whether freevms can be build on Debian 8 is: no. My branch fails on make bzImage when headers from the host Linux system are included. Of course such headers should never have been included in the first place. The master branch fails even earlier at make dep and which is something I fixed in my branch some time ago. I have not put any effort into fixing the build on debian 8. Regards, Guido -------------- next part -------------- An HTML attachment was scrubbed... URL: From roart at nvg.ntnu.no Mon May 11 07:54:06 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Mon, 11 May 2015 07:54:06 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: References: Message-ID: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> On Sun, May 10, 2015 at 11:33:21PM +0200, Guido de Jong wrote: > > The short answer to whether freevms can be build on Debian 8 is: no. My > branch fails on make bzImage when headers from the host Linux system are > included. Of course such headers should never have been included in the > first place. The master branch fails even earlier at make dep and which is > something I fixed in my branch some time ago. I have not put any effort > into fixing the build on debian 8. Noticed that from when I last tried a couple of weeks ago. But is the problems the files in /usr/include? I only got problems related to that the compiler is pickyer with regards to redefines, typing, which is a good thing. -- Regards, Roar Thronæ From roart at nvg.ntnu.no Mon May 11 07:54:06 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Mon, 11 May 2015 07:54:06 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: References: Message-ID: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> On Sun, May 10, 2015 at 11:33:21PM +0200, Guido de Jong wrote: > > The short answer to whether freevms can be build on Debian 8 is: no. My > branch fails on make bzImage when headers from the host Linux system are > included. Of course such headers should never have been included in the > first place. The master branch fails even earlier at make dep and which is > something I fixed in my branch some time ago. I have not put any effort > into fixing the build on debian 8. Noticed that from when I last tried a couple of weeks ago. But is the problems the files in /usr/include? I only got problems related to that the compiler is pickyer with regards to redefines, typing, which is a good thing. -- Regards, Roar Thronæ From guidoj2269 at gmail.com Mon May 11 11:56:20 2015 From: guidoj2269 at gmail.com (Guido) Date: Mon, 11 May 2015 11:56:20 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> Message-ID: <55507CC4.3000403@gmail.com> On 05/11/2015 07:54 AM, Roar Thronæs wrote: > On Sun, May 10, 2015 at 11:33:21PM +0200, Guido de Jong wrote: >> The short answer to whether freevms can be build on Debian 8 is: no. My >> branch fails on make bzImage when headers from the host Linux system are >> included. Of course such headers should never have been included in the >> first place. The master branch fails even earlier at make dep and which is >> something I fixed in my branch some time ago. I have not put any effort >> into fixing the build on debian 8. > Noticed that from when I last tried a couple of weeks ago. > > But is the problems the files in /usr/include? > > I only got problems related to that the compiler is pickyer with regards > to redefines, typing, which is a good thing. > Yes, you get loads of warnings about redefinitions, missing return types, unused variables, casting, etc. I've been trying to fix some of them with each commit, but there are still many left. I would even like to enable more warnings, because I have seen other issues (shadowing variables for instance), but that would not help now. The errors I ran into originate here: arch/x86_64/ia32/sys_ia32.c:#include arch/x86_64/ia32/ipc32.c:#include arch/x86_64/kernel/sys_x86_64.c:#include arch/i386/kernel/sys_i386.c:#include fs/binfmt_elf.c:#include ipc/shm.c:#include ipc/util.c:#include The file ./linux/include/linux/shm.h is missing form the repository (both branches), so the version from the host system in /usr/include/ is used in stead. This no longer works with the newer linux kernel version that comes with debian 8. As a quick fix, a linux 2.4 version of this file cpould be added to the repository. One could also try to remove the #include lines and see what happens. What worries me is that there might be more cases where headers are missing. I think I'm going to write a little script to find those first (I already did something similar to find unneeded header files). Regards, Guido From roart at nvg.ntnu.no Mon May 11 12:14:44 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Mon, 11 May 2015 12:14:44 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <55507CC4.3000403@gmail.com> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> Message-ID: <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> On Mon, May 11, 2015 at 11:56:20AM +0200, Guido wrote: > > The errors I ran into originate here: > > arch/x86_64/ia32/sys_ia32.c:#include > arch/x86_64/ia32/ipc32.c:#include > arch/x86_64/kernel/sys_x86_64.c:#include > arch/i386/kernel/sys_i386.c:#include > fs/binfmt_elf.c:#include > ipc/shm.c:#include > ipc/util.c:#include > > The file ./linux/include/linux/shm.h is missing form the repository > (both branches), so the version from the host system in /usr/include/ is > used in stead. This no longer works with the newer linux kernel version > that comes with debian 8. > > As a quick fix, a linux 2.4 version of this file cpould be added to the > repository. One could also try to remove the #include lines and see what > happens. The shm.h file was removed in a previous cleanup, it seems. And in this case you may remove all shm.h plus related code, because the shm support has been removed from the kernel. (Just leave empty system calls at most.) -- Regards, Roar Thronæs From roart at nvg.ntnu.no Mon May 11 12:14:44 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Mon, 11 May 2015 12:14:44 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <55507CC4.3000403@gmail.com> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> Message-ID: <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> On Mon, May 11, 2015 at 11:56:20AM +0200, Guido wrote: > > The errors I ran into originate here: > > arch/x86_64/ia32/sys_ia32.c:#include > arch/x86_64/ia32/ipc32.c:#include > arch/x86_64/kernel/sys_x86_64.c:#include > arch/i386/kernel/sys_i386.c:#include > fs/binfmt_elf.c:#include > ipc/shm.c:#include > ipc/util.c:#include > > The file ./linux/include/linux/shm.h is missing form the repository > (both branches), so the version from the host system in /usr/include/ is > used in stead. This no longer works with the newer linux kernel version > that comes with debian 8. > > As a quick fix, a linux 2.4 version of this file cpould be added to the > repository. One could also try to remove the #include lines and see what > happens. The shm.h file was removed in a previous cleanup, it seems. And in this case you may remove all shm.h plus related code, because the shm support has been removed from the kernel. (Just leave empty system calls at most.) -- Regards, Roar Thronæs From guidoj2269 at gmail.com Tue May 12 10:41:54 2015 From: guidoj2269 at gmail.com (Guido) Date: Tue, 12 May 2015 10:41:54 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> Message-ID: <5551BCD2.3010801@gmail.com> On 05/11/2015 12:14 PM, Roar Thronæs wrote: > On Mon, May 11, 2015 at 11:56:20AM +0200, Guido wrote: >> The errors I ran into originate here: >> >> arch/x86_64/ia32/sys_ia32.c:#include >> arch/x86_64/ia32/ipc32.c:#include >> arch/x86_64/kernel/sys_x86_64.c:#include >> arch/i386/kernel/sys_i386.c:#include >> fs/binfmt_elf.c:#include >> ipc/shm.c:#include >> ipc/util.c:#include >> >> The file ./linux/include/linux/shm.h is missing form the repository >> (both branches), so the version from the host system in /usr/include/ is >> used in stead. This no longer works with the newer linux kernel version >> that comes with debian 8. >> >> As a quick fix, a linux 2.4 version of this file cpould be added to the >> repository. One could also try to remove the #include lines and see what >> happens. > The shm.h file was removed in a previous cleanup, it seems. > And in this case you may remove all shm.h plus related code, > because the shm support has been removed from the kernel. > (Just leave empty system calls at most.) > Removed all references to shm and now make bzImage succeeds. Next problem is the function typedefs in hashtab.h (in both cdu and msgfil, seems an exact copy ...), which causes make vmsall to fail. I don't understand the typedefs myself yet and I wonder how it could have worked before. Guido From guidoj2269 at gmail.com Tue May 12 12:03:57 2015 From: guidoj2269 at gmail.com (Guido) Date: Tue, 12 May 2015 12:03:57 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <5551BCD2.3010801@gmail.com> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> <5551BCD2.3010801@gmail.com> Message-ID: <5551D00D.9050907@gmail.com> On 05/12/2015 10:41 AM, Guido wrote: > Removed all references to shm and now make bzImage succeeds. Next > problem is the function typedefs in hashtab.h (in both cdu and msgfil, > seems an exact copy ...), which causes make vmsall to fail. I don't > understand the typedefs myself yet and I wonder how it could have > worked before. > OK, that was easy enough. ansidecl.h has changed so that PARAMS is no longer defined. Since it is rather useless I removed it from hashtab.h too. Now I'm running into a slightly more difficult issue: something happened with the NOKERNEL and __KERNEL__ defines (mount and cmuip require a rather schrozifrenic definition of those). Guido From roart at nvg.ntnu.no Tue May 12 18:13:47 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Tue, 12 May 2015 18:13:47 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <5551BCD2.3010801@gmail.com> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> <5551BCD2.3010801@gmail.com> Message-ID: <20150512161347.GE18627@sabre-wulf.nvg.ntnu.no> On Tue, May 12, 2015 at 10:41:54AM +0200, Guido wrote: > > > Removed all references to shm and now make bzImage succeeds. Next > problem is the function typedefs in hashtab.h (in both cdu and msgfil, > seems an exact copy ...), which causes make vmsall to fail. I don't > understand the typedefs myself yet and I wonder how it could have worked > before. I remember problems with these, I think they are getting too old. They were taken (together wish hashtable.c) from gcc, hopefully it will be easy to replace them with recent versions? -- Regards, Roar Thronæs From roart at nvg.ntnu.no Tue May 12 18:13:47 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Tue, 12 May 2015 18:13:47 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <5551BCD2.3010801@gmail.com> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> <5551BCD2.3010801@gmail.com> Message-ID: <20150512161347.GE18627@sabre-wulf.nvg.ntnu.no> On Tue, May 12, 2015 at 10:41:54AM +0200, Guido wrote: > > > Removed all references to shm and now make bzImage succeeds. Next > problem is the function typedefs in hashtab.h (in both cdu and msgfil, > seems an exact copy ...), which causes make vmsall to fail. I don't > understand the typedefs myself yet and I wonder how it could have worked > before. I remember problems with these, I think they are getting too old. They were taken (together wish hashtable.c) from gcc, hopefully it will be easy to replace them with recent versions? -- Regards, Roar Thronæs From roart at nvg.ntnu.no Tue May 12 18:17:34 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Tue, 12 May 2015 18:17:34 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <5551D00D.9050907@gmail.com> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> <5551BCD2.3010801@gmail.com> <5551D00D.9050907@gmail.com> Message-ID: <20150512161734.GF18627@sabre-wulf.nvg.ntnu.no> On Tue, May 12, 2015 at 12:03:57PM +0200, Guido wrote: > On 05/12/2015 10:41 AM, Guido wrote: > >Removed all references to shm and now make bzImage succeeds. Next Good. Thanks. > OK, that was easy enough. ansidecl.h has changed so that PARAMS is no > longer defined. Since it is rather useless I removed it from hashtab.h Good. (Never mind what I wrote about renewing.) > too. Now I'm running into a slightly more difficult issue: something > happened with the NOKERNEL and __KERNEL__ defines (mount and cmuip > require a rather schrozifrenic definition of those). Sorry, the NOKERNEL is my creation. First, I had the CMUIP IPACP running as a kernel thread, before I realized that it belonged and would perform more correctly as a user level process. (As it did under VMS itself.) -- Regards, Roar Thronæs From roart at nvg.ntnu.no Tue May 12 18:17:34 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Tue, 12 May 2015 18:17:34 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <5551D00D.9050907@gmail.com> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> <5551BCD2.3010801@gmail.com> <5551D00D.9050907@gmail.com> Message-ID: <20150512161734.GF18627@sabre-wulf.nvg.ntnu.no> On Tue, May 12, 2015 at 12:03:57PM +0200, Guido wrote: > On 05/12/2015 10:41 AM, Guido wrote: > >Removed all references to shm and now make bzImage succeeds. Next Good. Thanks. > OK, that was easy enough. ansidecl.h has changed so that PARAMS is no > longer defined. Since it is rather useless I removed it from hashtab.h Good. (Never mind what I wrote about renewing.) > too. Now I'm running into a slightly more difficult issue: something > happened with the NOKERNEL and __KERNEL__ defines (mount and cmuip > require a rather schrozifrenic definition of those). Sorry, the NOKERNEL is my creation. First, I had the CMUIP IPACP running as a kernel thread, before I realized that it belonged and would perform more correctly as a user level process. (As it did under VMS itself.) -- Regards, Roar Thronæs From guidoj2269 at gmail.com Tue May 12 21:26:44 2015 From: guidoj2269 at gmail.com (Guido) Date: Tue, 12 May 2015 21:26:44 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <20150512161734.GF18627@sabre-wulf.nvg.ntnu.no> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> <5551BCD2.3010801@gmail.com> <5551D00D.9050907@gmail.com> <20150512161734.GF18627@sabre-wulf.nvg.ntnu.no> Message-ID: <555253F4.50502@gmail.com> On 05/12/2015 06:17 PM, Roar Thronæs wrote: > Sorry, the NOKERNEL is my creation. First, I had the CMUIP IPACP > running as a kernel thread, before I realized that it belonged and > would perform more correctly as a user level process. (As it did under > VMS itself.) Same thing for mount I guess. Compiling mountdsp.c fails (also on debian 7), but as mount is not installed it does not hurt too much for now. Anyway, I got the debian 8 build almost working. You need to install the libiberty-dev package, because libiberty.a no longer comes with binutils-dev. The only thing left is some obscure linker error on authorize. Something about truncating relocation on symbol memcmp ... Guido From guidoj2269 at gmail.com Wed May 13 11:16:36 2015 From: guidoj2269 at gmail.com (Guido de Jong) Date: Wed, 13 May 2015 11:16:36 +0200 Subject: [FreeVMS] Github Message-ID: Hi, Would it help if we start using some of the Github features like the issue tracker or the wiki? I considered registering tasks or work items in the issue tracker. Preferably a task should take only a limited amount of time (max 40 hours for instance), so even if you do not have much spare time you can still finish it within a reasonable time frame. Such tasks would not be too daunting for new developers either. Some examples: - get FreeVMS to build on debian 8 - add a test for sys$getdvi A bad example: - remove all Linux ' isms (too big, should be split up) Just an idea ... Regards, Guido -------------- next part -------------- An HTML attachment was scrubbed... URL: From roart at nvg.ntnu.no Wed May 13 16:16:50 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Wed, 13 May 2015 16:16:50 +0200 Subject: [FreeVMS] Github In-Reply-To: References: Message-ID: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> On Wed, May 13, 2015 at 11:16:36AM +0200, Guido de Jong wrote: > > Would it help if we start using some of the Github features like the issue > tracker or the wiki? Yes, issue tracking or wiki would be useful. -- Regards Roar Thronæs From roart at nvg.ntnu.no Wed May 13 16:16:50 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Wed, 13 May 2015 16:16:50 +0200 Subject: [FreeVMS] Github In-Reply-To: References: Message-ID: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> On Wed, May 13, 2015 at 11:16:36AM +0200, Guido de Jong wrote: > > Would it help if we start using some of the Github features like the issue > tracker or the wiki? Yes, issue tracking or wiki would be useful. -- Regards Roar Thronæs From joel.bertrand at systella.fr Wed May 13 18:17:14 2015 From: joel.bertrand at systella.fr (=?UTF-8?B?QkVSVFJBTkQgSm/Dq2w=?=) Date: Wed, 13 May 2015 18:17:14 +0200 Subject: [FreeVMS] Github In-Reply-To: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> References: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> Message-ID: <5553790A.7010707@systella.fr> Roar Thronæs a écrit : > Yes, issue tracking or wiki would be useful. There is a bugzilla interface at : https://www.systella.fr/bugzilla3 And you can obtain an account to write information on www.freevms.net if you want. Regards, JKB From guidoj2269 at gmail.com Thu May 14 09:13:40 2015 From: guidoj2269 at gmail.com (Guido) Date: Thu, 14 May 2015 09:13:40 +0200 Subject: [FreeVMS] Github In-Reply-To: <5553790A.7010707@systella.fr> References: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> <5553790A.7010707@systella.fr> Message-ID: <55544B24.6090105@gmail.com> On 05/13/2015 06:17 PM, BERTRAND Joël wrote: > Roar Thronæs a écrit : >> Yes, issue tracking or wiki would be useful. > > There is a bugzilla interface at : > https://www.systella.fr/bugzilla3 > > And you can obtain an account to write information on > www.freevms.net if you want. > I would like to take up the offer to use bugzilla. The github issue tracker provides only very basic functionality, where bugzilla has more features. The only downside is that bugzilla tends to call everything a bug, while in my opinion you can not have bugs until at least some functionality is considered done, but that is just being nitpicky. Only one question: will it not interfere with FreeVMS 0.4 development? Regards, Guido From joel.bertrand at systella.fr Thu May 14 10:13:24 2015 From: joel.bertrand at systella.fr (=?UTF-8?Q?BERTRAND_Jo=c3=abl?=) Date: Thu, 14 May 2015 10:13:24 +0200 Subject: [FreeVMS] Github In-Reply-To: <55544B24.6090105@gmail.com> References: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> <5553790A.7010707@systella.fr> <55544B24.6090105@gmail.com> Message-ID: <55545924.1030509@systella.fr> Guido a écrit : > On 05/13/2015 06:17 PM, BERTRAND Joël wrote: >> Roar Thronæs a écrit : >>> Yes, issue tracking or wiki would be useful. >> >> There is a bugzilla interface at : >> https://www.systella.fr/bugzilla3 >> >> And you can obtain an account to write information on >> www.freevms.net if you want. >> > I would like to take up the offer to use bugzilla. The github issue > tracker provides only very basic functionality, where bugzilla has more > features. The only downside is that bugzilla tends to call everything a > bug, while in my opinion you can not have bugs until at least some > functionality is considered done, but that is just being nitpicky. > > Only one question: will it not interfere with FreeVMS 0.4 development? No. It is possible to add sections for 0.3 and for 0.4. Today, 0.4 development is stopped as I don't have time enough to write a pager. Regards, JKB From guidoj2269 at gmail.com Thu May 14 23:57:24 2015 From: guidoj2269 at gmail.com (Guido) Date: Thu, 14 May 2015 23:57:24 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <555253F4.50502@gmail.com> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> <5551BCD2.3010801@gmail.com> <5551D00D.9050907@gmail.com> <20150512161734.GF18627@sabre-wulf.nvg.ntnu.no> <555253F4.50502@gmail.com> Message-ID: <55551A44.9080609@gmail.com> On 05/12/2015 09:26 PM, Guido wrote: > Anyway, I got the debian 8 build almost working. You need to install > the libiberty-dev package, because libiberty.a no longer comes with > binutils-dev. The only thing left is some obscure linker error on > authorize. Something about truncating relocation on symbol memcmp ... > It turns out that this linker error only occurs on x86_64, not on i386. Building FreeVMS on i386 Debian 8 works like a charm now, as a matter of fact even better than on Debian 7. Still working on the x86_64 build ... Guido From roart at nvg.ntnu.no Sun May 17 18:01:14 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Sun, 17 May 2015 18:01:14 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <55551A44.9080609@gmail.com> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> <5551BCD2.3010801@gmail.com> <5551D00D.9050907@gmail.com> <20150512161734.GF18627@sabre-wulf.nvg.ntnu.no> <555253F4.50502@gmail.com> <55551A44.9080609@gmail.com> Message-ID: <20150517160114.GI18627@sabre-wulf.nvg.ntnu.no> On Thu, May 14, 2015 at 11:57:24PM +0200, Guido wrote: > On 05/12/2015 09:26 PM, Guido wrote: > >Anyway, I got the debian 8 build almost working. You need to install > >the libiberty-dev package, because libiberty.a no longer comes with > >binutils-dev. The only thing left is some obscure linker error on > >authorize. Something about truncating relocation on symbol memcmp ... > > > It turns out that this linker error only occurs on x86_64, not on i386. > Building FreeVMS on i386 Debian 8 works like a charm now, as a matter of > fact even better than on Debian 7. Still working on the x86_64 build ... How is the truncation problem going (if the rest of the similar linkages works, this would be close, too)? How about mount/mountdsp (it is actually used), but the user part does not build until install time (for some reason)? For the Debian 6/7 based builds, have you noticed any cases where the virtual machine just suddenly reboots? There is at least one case of double fault (going immediately to triple fault and reset) that I am trying to reproduce, but forgot how it is triggered. -- Regards, Roar Thronæs From roart at nvg.ntnu.no Sun May 17 18:01:14 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Sun, 17 May 2015 18:01:14 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <55551A44.9080609@gmail.com> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> <5551BCD2.3010801@gmail.com> <5551D00D.9050907@gmail.com> <20150512161734.GF18627@sabre-wulf.nvg.ntnu.no> <555253F4.50502@gmail.com> <55551A44.9080609@gmail.com> Message-ID: <20150517160114.GI18627@sabre-wulf.nvg.ntnu.no> On Thu, May 14, 2015 at 11:57:24PM +0200, Guido wrote: > On 05/12/2015 09:26 PM, Guido wrote: > >Anyway, I got the debian 8 build almost working. You need to install > >the libiberty-dev package, because libiberty.a no longer comes with > >binutils-dev. The only thing left is some obscure linker error on > >authorize. Something about truncating relocation on symbol memcmp ... > > > It turns out that this linker error only occurs on x86_64, not on i386. > Building FreeVMS on i386 Debian 8 works like a charm now, as a matter of > fact even better than on Debian 7. Still working on the x86_64 build ... How is the truncation problem going (if the rest of the similar linkages works, this would be close, too)? How about mount/mountdsp (it is actually used), but the user part does not build until install time (for some reason)? For the Debian 6/7 based builds, have you noticed any cases where the virtual machine just suddenly reboots? There is at least one case of double fault (going immediately to triple fault and reset) that I am trying to reproduce, but forgot how it is triggered. -- Regards, Roar Thronæs From roart at nvg.ntnu.no Sun May 17 17:50:34 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Sun, 17 May 2015 17:50:34 +0200 Subject: [FreeVMS] Github In-Reply-To: <55544B24.6090105@gmail.com> References: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> <5553790A.7010707@systella.fr> <55544B24.6090105@gmail.com> Message-ID: <20150517155034.GH18627@sabre-wulf.nvg.ntnu.no> On Thu, May 14, 2015 at 09:13:40AM +0200, Guido wrote: > > I would like to take up the offer to use bugzilla. The github issue > tracker provides only very basic functionality, where bugzilla has more > features. The only downside is that bugzilla tends to call everything a > bug, while in my opinion you can not have bugs until at least some > functionality is considered done, but that is just being nitpicky. I have not tested the github issue tracker, but may it have advantages with having the same set of users and some connection to git itself? -- Regards, Roar Thronæs From roart at nvg.ntnu.no Sun May 17 17:50:34 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Sun, 17 May 2015 17:50:34 +0200 Subject: [FreeVMS] Github In-Reply-To: <55544B24.6090105@gmail.com> References: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> <5553790A.7010707@systella.fr> <55544B24.6090105@gmail.com> Message-ID: <20150517155034.GH18627@sabre-wulf.nvg.ntnu.no> On Thu, May 14, 2015 at 09:13:40AM +0200, Guido wrote: > > I would like to take up the offer to use bugzilla. The github issue > tracker provides only very basic functionality, where bugzilla has more > features. The only downside is that bugzilla tends to call everything a > bug, while in my opinion you can not have bugs until at least some > functionality is considered done, but that is just being nitpicky. I have not tested the github issue tracker, but may it have advantages with having the same set of users and some connection to git itself? -- Regards, Roar Thronæs From guidoj2269 at gmail.com Mon May 18 22:05:09 2015 From: guidoj2269 at gmail.com (Guido) Date: Mon, 18 May 2015 22:05:09 +0200 Subject: [FreeVMS] Github In-Reply-To: <20150517155034.GH18627@sabre-wulf.nvg.ntnu.no> References: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> <5553790A.7010707@systella.fr> <55544B24.6090105@gmail.com> <20150517155034.GH18627@sabre-wulf.nvg.ntnu.no> Message-ID: <555A45F5.5000501@gmail.com> On 05/17/2015 05:50 PM, Roar Thronæs wrote: > On Thu, May 14, 2015 at 09:13:40AM +0200, Guido wrote: >> I would like to take up the offer to use bugzilla. The github issue >> tracker provides only very basic functionality, where bugzilla has more >> features. The only downside is that bugzilla tends to call everything a >> bug, while in my opinion you can not have bugs until at least some >> functionality is considered done, but that is just being nitpicky. > I have not tested the github issue tracker, but may it have advantages > with having the same set of users and some connection to git itself? > The github issue tracker is really very simplistic. Basically it is just a list of issues that you can assign to a user, give a priority and label it. A bit like the sourceforge issue tracker, I guess. Bugzilla is much more sophisticated and it can be customized to fit your product, workflow, etc. One feature I really like is dependencies between issues, so that you do not have to worry about adding huge task as a single issue. Simply add it and make it depend on sub tasks. Having the same set of users by default is nice to have, but so far Joël has been very forthcoming, so I do not see that as a big advantage. Anyway, I started using bugzilla on freevms.net already. It is not much yet, but it will grow. Guido From guidoj2269 at gmail.com Mon May 18 22:25:12 2015 From: guidoj2269 at gmail.com (Guido) Date: Mon, 18 May 2015 22:25:12 +0200 Subject: [FreeVMS] Debian 8 In-Reply-To: <20150517160114.GI18627@sabre-wulf.nvg.ntnu.no> References: <20150511055406.GC18627@sabre-wulf.nvg.ntnu.no> <55507CC4.3000403@gmail.com> <20150511101444.GD18627@sabre-wulf.nvg.ntnu.no> <5551BCD2.3010801@gmail.com> <5551D00D.9050907@gmail.com> <20150512161734.GF18627@sabre-wulf.nvg.ntnu.no> <555253F4.50502@gmail.com> <55551A44.9080609@gmail.com> <20150517160114.GI18627@sabre-wulf.nvg.ntnu.no> Message-ID: <555A4AA8.3010103@gmail.com> On 05/17/2015 06:01 PM, Roar Thronæs wrote: > How is the truncation problem going (if the rest of the similar linkages > works, this would be close, too)? The root cause is that FreeVMS is build against two different glibc versions, the once from the linux host and the old debian version that you added for debian 7. That is also the reason his only occurs on x86_64. I guess the linker error could have happened on other images, because the glibc clash produces compiler warnings for almost all of them. > How about mount/mountdsp (it is actually used), but the user part does not > build until install time (for some reason)? Sorry, I haven't had the time to look into this yet. > For the Debian 6/7 based builds, have you noticed any cases where the > virtual machine just suddenly reboots? > There is at least one case of double fault (going immediately to triple > fault and reset) that I am trying to reproduce, but forgot how it is triggered. I have seen such a reboot, but that was when running one of my tests where I made a stupid mistake (I tried to dereference an integer value). So it is definitely not the same one you saw, though it might have the same cause given the compiler warnings about converting integers to pointers and vice versa. Once you know how to trigger it, use "log exec" in qemu (also make sure you specified a logfile), so you can pin point where it is triggered in the code. Guido From roart at nvg.ntnu.no Wed May 20 21:04:50 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Wed, 20 May 2015 21:04:50 +0200 Subject: [FreeVMS] Github In-Reply-To: <555A45F5.5000501@gmail.com> References: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> <5553790A.7010707@systella.fr> <55544B24.6090105@gmail.com> <20150517155034.GH18627@sabre-wulf.nvg.ntnu.no> <555A45F5.5000501@gmail.com> Message-ID: <20150520190450.GJ18627@sabre-wulf.nvg.ntnu.no> On Mon, May 18, 2015 at 10:05:09PM +0200, Guido wrote: > > Anyway, I started using bugzilla on freevms.net already. It is not much > yet, but it will grow. I have added myself to it. But: Bugzilla at https://www.systella.fr/bugzilla3/ was a bit hidden, it should be linked to from the main page. -- Regards, Roar Thronæs From roart at nvg.ntnu.no Wed May 20 21:04:50 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Wed, 20 May 2015 21:04:50 +0200 Subject: [FreeVMS] Github In-Reply-To: <555A45F5.5000501@gmail.com> References: <20150513141650.GG18627@sabre-wulf.nvg.ntnu.no> <5553790A.7010707@systella.fr> <55544B24.6090105@gmail.com> <20150517155034.GH18627@sabre-wulf.nvg.ntnu.no> <555A45F5.5000501@gmail.com> Message-ID: <20150520190450.GJ18627@sabre-wulf.nvg.ntnu.no> On Mon, May 18, 2015 at 10:05:09PM +0200, Guido wrote: > > Anyway, I started using bugzilla on freevms.net already. It is not much > yet, but it will grow. I have added myself to it. But: Bugzilla at https://www.systella.fr/bugzilla3/ was a bit hidden, it should be linked to from the main page. -- Regards, Roar Thronæs From roart at nvg.ntnu.no Thu May 21 13:12:53 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Thu, 21 May 2015 13:12:53 +0200 Subject: [FreeVMS] Bugzilla In-Reply-To: <51C32A1B.40104@systella.fr> References: <51C32A1B.40104@systella.fr> Message-ID: <20150521111253.GK18627@sabre-wulf.nvg.ntnu.no> On Thu, Jun 20, 2013 at 06:13:15PM +0200, BERTRAND Joël wrote: > > I have fixed FreeVMS bugzilla. It's now available at > https://www.systella.fr/bugzilla3 When accessing this, just ignore the warning about the certificate, and when getting the account creation mail, just add the token.cgi after https://www.systella.fr/bugzilla3/, getting https://www.systella.fr/bugzilla3/token.cgi... -- Regards, Roar Thronæs From roart at nvg.ntnu.no Thu May 21 13:12:53 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Thu, 21 May 2015 13:12:53 +0200 Subject: [FreeVMS] Bugzilla In-Reply-To: <51C32A1B.40104@systella.fr> References: <51C32A1B.40104@systella.fr> Message-ID: <20150521111253.GK18627@sabre-wulf.nvg.ntnu.no> On Thu, Jun 20, 2013 at 06:13:15PM +0200, BERTRAND Joël wrote: > > I have fixed FreeVMS bugzilla. It's now available at > https://www.systella.fr/bugzilla3 When accessing this, just ignore the warning about the certificate, and when getting the account creation mail, just add the token.cgi after https://www.systella.fr/bugzilla3/, getting https://www.systella.fr/bugzilla3/token.cgi... -- Regards, Roar Thronæs From guidoj2269 at gmail.com Fri May 29 23:55:08 2015 From: guidoj2269 at gmail.com (Guido) Date: Fri, 29 May 2015 23:55:08 +0200 Subject: [FreeVMS] x86_64 ia32 emulation Message-ID: <5568E03C.4040808@gmail.com> Hi, What is the reason that the FreeVMS linux kernel configuration has IA32 emulation support on x86_64 architecture? Regards, Guido From roart at nvg.ntnu.no Sat May 30 15:18:43 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Sat, 30 May 2015 15:18:43 +0200 Subject: [FreeVMS] x86_64 ia32 emulation In-Reply-To: <5568E03C.4040808@gmail.com> References: <5568E03C.4040808@gmail.com> Message-ID: <20150530131843.GL18627@sabre-wulf.nvg.ntnu.no> On Fri, May 29, 2015 at 11:55:08PM +0200, Guido wrote: > > What is the reason that the FreeVMS linux kernel configuration has IA32 > emulation support on x86_64 architecture? It was inherited from the imported kernel. It might come in handy to run 32-bit images, but I have not used it yet. -- Regards, Roar Thronæs From roart at nvg.ntnu.no Sat May 30 15:18:43 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Sat, 30 May 2015 15:18:43 +0200 Subject: [FreeVMS] x86_64 ia32 emulation In-Reply-To: <5568E03C.4040808@gmail.com> References: <5568E03C.4040808@gmail.com> Message-ID: <20150530131843.GL18627@sabre-wulf.nvg.ntnu.no> On Fri, May 29, 2015 at 11:55:08PM +0200, Guido wrote: > > What is the reason that the FreeVMS linux kernel configuration has IA32 > emulation support on x86_64 architecture? It was inherited from the imported kernel. It might come in handy to run 32-bit images, but I have not used it yet. -- Regards, Roar Thronæs From guidoj2269 at gmail.com Sun May 31 16:46:21 2015 From: guidoj2269 at gmail.com (Guido) Date: Sun, 31 May 2015 16:46:21 +0200 Subject: [FreeVMS] x86_64 ia32 emulation In-Reply-To: <20150530131843.GL18627@sabre-wulf.nvg.ntnu.no> References: <5568E03C.4040808@gmail.com> <20150530131843.GL18627@sabre-wulf.nvg.ntnu.no> Message-ID: <556B1EBD.1040807@gmail.com> On 05/30/2015 03:18 PM, Roar Thronæs wrote: > On Fri, May 29, 2015 at 11:55:08PM +0200, Guido wrote: >> What is the reason that the FreeVMS linux kernel configuration has IA32 >> emulation support on x86_64 architecture? > It was inherited from the imported kernel. > It might come in handy to run 32-bit images, but I have not used it yet. > OK. I can not imagine what kind of 32 bit images that will be though, because IA32 emulation would only necessary for closed source x86 VMS binaries or maybe when wine (www.winehq.org) is ported to VMS. The problem with the current FreeVMS code is that it is very unstable. Changing anything in one place can easily break things in a seemingly unrelated other parts. I experienced this both with simple code changes, but also with small changes in the build environment. Reasons can be uninitialized variables, function parameters that are properly checked before use, using type "long" on both i386 (where long is 32 bits wide) and x86_64 (long is 64 bits wide), etc. I am trying to strip the code down to what is really needed and to add some automated tests to identify and fix these problems. The tests can also be used for regression testing. Anyway, this is quite lot of work, so do not expect new features any time soon. Regards, Guido From roart at nvg.ntnu.no Sun May 31 22:32:40 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Sun, 31 May 2015 22:32:40 +0200 Subject: [FreeVMS] x86_64 ia32 emulation In-Reply-To: <556B1EBD.1040807@gmail.com> References: <5568E03C.4040808@gmail.com> <20150530131843.GL18627@sabre-wulf.nvg.ntnu.no> <556B1EBD.1040807@gmail.com> Message-ID: <20150531203240.GA24151@sabre-wulf.nvg.ntnu.no> On Sun, May 31, 2015 at 04:46:21PM +0200, Guido wrote: > > > OK. I can not imagine what kind of 32 bit images that will be though, > because IA32 emulation would only necessary for closed source x86 VMS > binaries or maybe when wine (www.winehq.org) is ported to VMS. Maybe some 32-bit programs that are difficult to make 64-bit. (But I have already abandoned them.) > I am trying to strip the code down to what is really needed and to add > some automated tests to identify and fix these problems. The tests can > also be used for regression testing. Anyway, this is quite lot of work, > so do not expect new features any time soon. Yes, I think the IA32 emulation may be stripped away for now. Any other parts? -- Regards, Roar Thronæs From roart at nvg.ntnu.no Sun May 31 22:32:40 2015 From: roart at nvg.ntnu.no (Roar =?iso-8859-1?Q?Thron=C3=A6s?=) Date: Sun, 31 May 2015 22:32:40 +0200 Subject: [FreeVMS] x86_64 ia32 emulation In-Reply-To: <556B1EBD.1040807@gmail.com> References: <5568E03C.4040808@gmail.com> <20150530131843.GL18627@sabre-wulf.nvg.ntnu.no> <556B1EBD.1040807@gmail.com> Message-ID: <20150531203240.GA24151@sabre-wulf.nvg.ntnu.no> On Sun, May 31, 2015 at 04:46:21PM +0200, Guido wrote: > > > OK. I can not imagine what kind of 32 bit images that will be though, > because IA32 emulation would only necessary for closed source x86 VMS > binaries or maybe when wine (www.winehq.org) is ported to VMS. Maybe some 32-bit programs that are difficult to make 64-bit. (But I have already abandoned them.) > I am trying to strip the code down to what is really needed and to add > some automated tests to identify and fix these problems. The tests can > also be used for regression testing. Anyway, this is quite lot of work, > so do not expect new features any time soon. Yes, I think the IA32 emulation may be stripped away for now. Any other parts? -- Regards, Roar Thronæs