Re: removing f2c from base distribution
There is a [good?] g77 port built into egcs-1.1.1, although I have never used it. Why would gcc-2.8.x be the stock compiler versus the [better?] egcs-1.1.x? C++ comes along for free in all it's glory. Tom Veldhouse ve...@visi.com If it is decided that Fortran support will disappear from the base system and nobody else wants to maintain g77, I will gladly do it. However, I will only maintain a version that I am using so that means I will maintain a port once gcc 2.8 is officially brought in as the stock compiler. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
By bloat, I assume you mean disk space. There should not be any difference as far as code bloat goes (i.e. templates in C++ are known to be prone to code bloat). Executables are probably a bit smaller than the g77 port due to the newer optimizations in egcs. If you are worried about disk space, I believe that you can configure the egcs compilation to not build all the extras like g++ and the java compiler. Tom Veldhouse ve...@visi.com -Original Message- From: gjohn...@nola.srrc.usda.gov gjohn...@nola.srrc.usda.gov To: Satoshi Asami as...@freebsd.org Cc: s...@troutmask.apl.washington.edu s...@troutmask.apl.washington.edu; curr...@freebsd.org curr...@freebsd.org; obr...@nuxi.com obr...@nuxi.com Date: Friday, February 19, 1999 2:17 PM Subject: Re: removing f2c from base distribution On Wed, Jan 27, 1999 at 09:29:38AM -0800, Satoshi Asami wrote: * The g77-0.5.19(.1) is *extremely* out-of-date. It should be dropped from * the ports collection, and if someone wants to use g77, then they should * install egcs. * * The newer versions of g77 do not work with gcc-2.7.2.x. The author of * g77 states that you shouldn't even try to back port g77 to any version * of gcc earlier than gcc-2.8 Well, Glenn said he got it to work with our compiler. :) No, I said I had a port of g77 0.5.19 that built against gcc 2.7.2.1. This was a while back. I am currently using g77 0.5.24, which needs gcc 2.8. Getting g77 from egcs is the best option right now. However, it seems to me that this adds a lot of bloat (duplication of C, C++, etc.) to the system for someone wanting to use FreeBSD as a scientific workstation platform. In contrast, how much bloat is added to the base system by having a g77 binary and a couple of libraries? But anyway, I don't have any problem to delete the g77 port for now. Is that ok with everyone else? This is OK by me. -- Glenn Johnson Technician USDA, ARS, SRRC New Orleans, LA To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
If it is decided that Fortran support will disappear from the base system and nobody else wants to maintain g77, I will gladly do it. However, I will only maintain a version that I am using so that means I will maintain a port once gcc 2.8 is officially brought in as the stock compiler. Lets just crush the g77 port and tell people to install EGCS. I build g77 as part of that port. I welcome patches and feedback on the usablability of g77 w/in EGCS. -- -- David(obr...@nuxi.com -or- obr...@freebsd.org) To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Getting g77 from egcs is the best option right now. However, it seems to me that this adds a lot of bloat (duplication of C, C++, etc.) to the system for someone wanting to use FreeBSD as a scientific workstation platform. Then update the g77 port to fetch egcs-core-XYZ.tar.gz and egcs-g77-XYZ.tar.gz and build it that way. You will have no C, C++ or objC support in the result. -- -- David(obr...@nuxi.com -or- obr...@freebsd.org) To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Chuck Robey wrote: I'm not sure if this argument is worth pushing anymore, because FreeBSD's stability and usefulness has become much more well known, but it did contribute at some point, and I think that is the idea that Daniel was trying to convey. Right? Me? No... Maybe Garret... :-) -- Daniel C. Sobral(8-DCS) d...@newsguy.com If you sell your soul to the Devil and all you get is an MCSE from it, you haven't gotten market rate. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
I highly doubt that I'll ever use FORTRAN directly or indirectly. If it's not used by a vast majority, it should be optional... -Original Message- From: Mike Smith m...@smith.net.au To: Mark Murray m...@grondar.za Cc: obr...@nuxi.com obr...@nuxi.com; Steve Kargl s...@troutmask.apl.washington.edu; freebsd-current@FreeBSD.ORG freebsd-current@FreeBSD.ORG Date: Wednesday, January 27, 1999 2:40 AM Subject: Re: removing f2c from base distribution David O'Brien wrote: I've got a Bmaked contribified version of EGCS, but didn't do g77. So maybe a consensus should be made what to do about FORTRAN in the base system. If you are collecting votes, please add mine; I feel quite strongly (knowing the scientists that I do that use Fortran) that it _should_ be in the base system. Please include it. We've already established that it doesn't need to be in the base system. Your scientist friends are probably already going to be installing their favorite text editors; it's no harder to install the Fortran support. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Mark Murray wrote: David O'Brien wrote: I've got a Bmaked contribified version of EGCS, but didn't do g77. So maybe a consensus should be made what to do about FORTRAN in the base system. If you are collecting votes, please add mine; I feel quite strongly (knowing the scientists that I do that use Fortran) that it _should_ be in the base system. Please include it. A lot of people use a lot of things out of ports. Why should Fortran be different? -- Daniel C. Sobral(8-DCS) d...@newsguy.com If you sell your soul to the Devil and all you get is an MCSE from it, you haven't gotten market rate. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
* From: Steve Kargl s...@troutmask.apl.washington.edu * g77 is a frontend to the FSF compiler backend, and thus it is bound * to specific versions. So, it could become a support nightmare to ensure * a g77 port is in sync with the egcs backend in the base distribution. I don't think it would be that much of a support nightmare. Compilers in the base system don't change that often. If the port maintainer takes care to synchronize it with the system compiler, we should be fine. * It might even be impractical to try to build a standalone g77 port. That I don't know. I believe the compiler driver (gcc) can be instructed to call frontends in places other than /usr/libexec though. Maybe it would be feasible if we leave in hooks for that. (That is, if people want to compile fortran programs with /usr/bin/gcc.) Satoshi To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
* The biggest problem has been that the port of g77 has not worked * properly for quite some time and in fact is currently marked as * broken. I would anticipate that this situation would not change much in That (and bug fix issues, as DavidO contends) all depends on the commitment of the maintainer (which there is none for the g77 port). Unless someone who uses g77 regularly steps up to maintain it, it will remain broken. This is the same for all ports, and I don't see why a fortran compiler should be an exception. Granted, if won't be blatantly broken if it's in the base distribution, but that's only because people will yell and scream if their make world doesn't work. If the amount of noise that generates is significantly different from what happens if it's a broken port, that's actually a pretty good argument *against* putting it in the base distribution, as it means we can keep g77 running only by annoying people who don't use it when it's broken. (1/2 :) This port has been marked broken since July last year. Sorry, but I just don't have a whole lot of sympathy for something that can stay broken that long without anyone fixing it. ;) Satoshi To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
On Tue, 26 Jan 1999, RT wrote: I highly doubt that I'll ever use FORTRAN directly or indirectly. If it's not used by a vast majority, it should be optional... So the problem seems to be that 'included in the system' is a problem because the system gets unwieldy in terms of junk a lot of people don't use, but 'included as a port' or 'included as a package' means it might be too detached from the system. There's quite a list of things on this list, including UUCP which the majority do not use. There's also that list of things which almost everyone use except some people who find it inconvenient that it's included, such as sendmail. What might be really nice is to see all user-land files broken out into whatever the new package format will be, in the style of RedHat packages for the base system. At install, needless to say, you have a default install that looks just like today's (it installs the packages that map directly to the current system), but you also have other installs, and the option to flag things in and out of the install, in the style of existing packages, with dependencies, etc. One sad side-effect of this would, of course, be managing the dependencies (both in source and in binary form)and the screwing up of the existing build tree if the build tree was to be restructured to match the packages. But the RedHat arrangement does have appeal: I understand that even / is part of a package :). And I certainly don't have time (and probably not the understanding) to figure out how to make all this work. :) Robert N Watson rob...@fledge.watson.org http://www.watson.org/~robert/ PGP key fingerprint: 03 01 DD 8E 15 67 48 73 25 6D 10 FC EC 68 C1 1C Carnegie Mellon Universityhttp://www.cmu.edu/ TIS Labs at Network Associates, Inc. http://www.tis.com/ SafePort Network Services http://www.safeport.com/ To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
On Wed, Jan 27, 1999 at 05:14:33AM -0800, Satoshi Asami wrote: * The biggest problem has been that the port of g77 has not worked * properly for quite some time and in fact is currently marked as * broken. I would anticipate that this situation would not change much in That (and bug fix issues, as DavidO contends) all depends on the commitment of the maintainer (which there is none for the g77 port). Unless someone who uses g77 regularly steps up to maintain it, it will remain broken. This is the same for all ports, and I don't see why a Fortran compiler should be an exception. Granted, if won't be blatantly broken if it's in the base distribution, but that's only because people will yell and scream if their make world doesn't work. If the amount of noise that generates is significantly different from what happens if it's a broken port, that's actually a pretty good argument *against* putting it in the base distribution, as it means we can keep g77 running only by annoying people who don't use it when it's broken. (1/2 :) This port has been marked broken since July last year. Sorry, but I just don't have a whole lot of sympathy for something that can stay broken that long without anyone fixing it. ;) Your points are well taken. I had a local port of g77 that built against our current gcc. I never submitted it however for a couple of reasons: 1. The port I had was for 0.5.19. This will build against our current gcc, but g77 has advanced significantly since then. Unfortunately, the newer versions need gcc 2.8. It was simply easier for me to use the newer versions of g77 with gcc 2.8 or egcs release versions. Note that I said easier for me; some colleagues of mine could/would not want to have to maintain a compiler on their own. Yes, I was told this on a couple of occasions. I can not see a point in me becoming the g77 0.5.19 port maintainer when I am using newer versions of g77. 2. In light of the above, it seemed that f77 (f2c/gcc) was good enough for most cases. The g77 port was not essential because there was fairly good Fortran support in the base system. Apparently this will no longer be the case and therefore the g77 (or f2c) port will become essential. That is to say, essential for those needing Fortran. If it is decided that Fortran support will disappear from the base system and nobody else wants to maintain g77, I will gladly do it. However, I will only maintain a version that I am using so that means I will maintain a port once gcc 2.8 is officially brought in as the stock compiler. -- Glenn Johnson Technician USDA, ARS, SRRC New Orleans, LA To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
On Wed, 27 Jan 1999 20:44:33 +0900, Daniel C. Sobral d...@newsguy.com said: A lot of people use a lot of things out of ports. Why should Fortran be different? Because Berkeley Unix has /always/ included a FORTRAN compiler. -GAWollman -- Garrett A. Wollman | O Siem / We are all family / O Siem / We're all the same woll...@lcs.mit.edu | O Siem / The fires of freedom Opinions not those of| Dance in the burning flame MIT, LCS, CRS, or NSA| - Susan Aglukark and Chad Irschick To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Um, I'm still alive but can someone explain me why this can't be a regular port? Being useful to some but not the majority, no other parts of the system depending on it, this looks like a model citizen in the ideal ports world. :) Because we loose control over it. There is a move to push some things out of the base system and into ports. Fine, but then how does bug fixes happen? I think this is a moot point. As Steven pointed out months ago, the version of f2c in the base system had rotted due to lack of maintainer. Keeping it in the base system isn't going to keep it updated unless *someone* wants to update it. I believe the committed that fixed the f2c problems on the Alpha really liked having the abilility to do so. And that person would have been able to do it in the ports if they were so motivated just as easily. Nate To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
* From: Garrett Wollman woll...@khavrinen.lcs.mit.edu * A lot of people use a lot of things out of ports. Why should Fortran * be different? * * Because Berkeley Unix has /always/ included a FORTRAN compiler. Maybe that's because Berkeley Unix never had (until recently, anyway) a ports system? :) Satoshi To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
A lot of people use a lot of things out of ports. Why should Fortran be different? Because Berkeley Unix has /always/ included a FORTRAN compiler. And they have /always/ included games. Next issue. Nate To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
* From: Glenn Johnson gjohn...@nola.srrc.usda.gov * Your points are well taken. I had a local port of g77 that built * against our current gcc. I never submitted it however for a couple * of reasons: * * 1. The port I had was for 0.5.19. This will build against our current *gcc, but g77 has advanced significantly since then. Unfortunately, the *newer versions need gcc 2.8. It was simply easier for me to use the *newer versions of g77 with gcc 2.8 or egcs release versions. Note that I *said easier for me; some colleagues of mine could/would not want to have *to maintain a compiler on their own. Yes, I was told this on a couple *of occasions. I can not see a point in me becoming the g77 0.5.19 port *maintainer when I am using newer versions of g77. I wasn't blaming you (or anyone in particular, for that matter). You don't even have to become a maintainer. But if you have a fix to the build problems, and you think it is an important enough piece of software for you and other people, please send in a patch so it can be built. Satoshi To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Garrett Wollman wrote: On Wed, 27 Jan 1999 20:44:33 +0900, Daniel C. Sobral d...@newsguy.com said: A lot of people use a lot of things out of ports. Why should Fortran be different? Because Berkeley Unix has /always/ included a FORTRAN compiler. Somehow I feared you might have said that... :-) All things considered, though, it just doesn't mean anything to me. So, this is were we agree to disagree, I guess. :-) -- Daniel C. Sobral(8-DCS) d...@newsguy.com If you sell your soul to the Devil and all you get is an MCSE from it, you haven't gotten market rate. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Satoshi Asami wrote: * From: Glenn Johnson gjohn...@nola.srrc.usda.gov * Your points are well taken. I had a local port of g77 that built * against our current gcc. I never submitted it however for a couple * of reasons: * * 1. The port I had was for 0.5.19. This will build against our current *gcc, but g77 has advanced significantly since then. Unfortunately, the *newer versions need gcc 2.8. It was simply easier for me to use the *newer versions of g77 with gcc 2.8 or egcs release versions. Note that I *said easier for me; some colleagues of mine could/would not want to have *to maintain a compiler on their own. Yes, I was told this on a couple *of occasions. I can not see a point in me becoming the g77 0.5.19 port *maintainer when I am using newer versions of g77. I wasn't blaming you (or anyone in particular, for that matter). You don't even have to become a maintainer. But if you have a fix to the build problems, and you think it is an important enough piece of software for you and other people, please send in a patch so it can be built. The g77-0.5.19(.1) is *extremely* out-of-date. It should be dropped from the ports collection, and if someone wants to use g77, then they should install egcs. The newer versions of g77 do not work with gcc-2.7.2.x. The author of g77 states that you shouldn't even try to back port g77 to any version of gcc earlier than gcc-2.8 -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Garrett Wollman wrote: On Wed, 27 Jan 1999 20:44:33 +0900, Daniel C. Sobral d...@newsguy.com said: A lot of people use a lot of things out of ports. Why should Fortran be different? Because Berkeley Unix has /always/ included a FORTRAN compiler. Didn't Berkeley Unix also include a Pascal compiler? -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
* The g77-0.5.19(.1) is *extremely* out-of-date. It should be dropped from * the ports collection, and if someone wants to use g77, then they should * install egcs. * * The newer versions of g77 do not work with gcc-2.7.2.x. The author of * g77 states that you shouldn't even try to back port g77 to any version * of gcc earlier than gcc-2.8 Well, Glenn said he got it to work with our compiler. :) But anyway, I don't have any problem to delete the g77 port for now. Is that ok with everyone else? Satoshi To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Satoshi Asami wrote: * The g77-0.5.19(.1) is *extremely* out-of-date. It should be dropped from * the ports collection, and if someone wants to use g77, then they should * install egcs. * * The newer versions of g77 do not work with gcc-2.7.2.x. The author of * g77 states that you shouldn't even try to back port g77 to any version * of gcc earlier than gcc-2.8 Well, Glenn said he got it to work with our compiler. :) Maybe I misunderstood, but I thought Glenn got g77-0.5.19 to work with our gcc-2.7.x. g77 is now at version 0.5.24. Those micro numbers are significant changes, and these represent over a years work on g77. finger -l fort...@delysid.gnu.org |more But anyway, I don't have any problem to delete the g77 port for now. Is that ok with everyone else? Satoshi -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
* Maybe I misunderstood, but I thought Glenn got g77-0.5.19 to * work with our gcc-2.7.x. g77 is now at version 0.5.24. Those * micro numbers are significant changes, and these represent over * a years work on g77. No, I misunderstood. So Glenn got 0.5.19 to work, but it's very old. Anything newer than that (like the current 0.5.24) doesn't work with gcc 2.7.x and the author says don't bother trying. I got it now. Satoshi To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
In message 36af4948.d0f88...@newsguy.com, Daniel C. Sobral writes: Garrett Wollman wrote: On Wed, 27 Jan 1999 20:44:33 +0900, Daniel C. Sobral d...@newsguy.com said: A lot of people use a lot of things out of ports. Why should Fortran be different? Because Berkeley Unix has /always/ included a FORTRAN compiler. Somehow I feared you might have said that... :-) Yes, it was impolite to point out a mistake like that in public... Poul-Henning Just because Berkeley always did it that way doesn't mean it is the right way Kamp -- Poul-Henning Kamp FreeBSD coreteam member p...@freebsd.org Real hackers run -current on their laptop. FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
On Wed, Jan 27, 1999 at 09:29:38AM -0800, Satoshi Asami wrote: * The g77-0.5.19(.1) is *extremely* out-of-date. It should be dropped from * the ports collection, and if someone wants to use g77, then they should * install egcs. * * The newer versions of g77 do not work with gcc-2.7.2.x. The author of * g77 states that you shouldn't even try to back port g77 to any version * of gcc earlier than gcc-2.8 Well, Glenn said he got it to work with our compiler. :) No, I said I had a port of g77 0.5.19 that built against gcc 2.7.2.1. This was a while back. I am currently using g77 0.5.24, which needs gcc 2.8. Getting g77 from egcs is the best option right now. However, it seems to me that this adds a lot of bloat (duplication of C, C++, etc.) to the system for someone wanting to use FreeBSD as a scientific workstation platform. In contrast, how much bloat is added to the base system by having a g77 binary and a couple of libraries? But anyway, I don't have any problem to delete the g77 port for now. Is that ok with everyone else? This is OK by me. -- Glenn Johnson Technician USDA, ARS, SRRC New Orleans, LA To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Because Berkeley Unix has /always/ included a FORTRAN compiler. Maybe we should put Franz Lisp back in. bash-2.02$ uname -sr FreeBSD 3.0-RELEASE bash-2.02$ lisp Franz Lisp, Opus 38.92 - -- Richard To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
: A lot of people use a lot of things out of ports. Why should Fortran : be different? For g77, because it is integrated with the C compiler. The system has a lower maintenance cost if it is included. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
From: Daniel C. Sobral d...@newsguy.com A lot of people use a lot of things out of ports. Why should Fortran be different? Right on. If anything, I'd like to see the ports system continue its evolution to becoming able to build nearly any component of the system. (including patched kernel builds and release generation). One thing I'd like to understand is how to add CVSup and anonCVS fetching capability. This would be very useful for constructing 'ports' which are derived from FreeBSD sources (like an enhanced userland Ficl). This capability would also be useful for software like EGCS which has public anonCVS access. Cheers, Jerry Hicks wghi...@bellsouth.net To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Garrett Wollman writes: On Wed, 27 Jan 1999 20:44:33 +0900, Daniel C. Sobral d...@newsguy.com said: A lot of people use a lot of things out of ports. Why should Fortran be different? Because Berkeley Unix has /always/ included a FORTRAN compiler. So FreeBSD v12.4, released in 2026, had better include a FORTRAN compiler, because Berkely Unix has /always/ included a FORTRAN compiler? I'm sure there are a fair number of ways FreeBSD has diverged from the way Berkeley Unix has always done things (for example, to conform to POSIX), is that such a bad thing? If it's a port, and sysinstall gives the user an option to install a FORTRAN compiler, is that so radically different from Berkeley Unix /always/ including a FORTRAN compiler? Is it wrong to move things that most people installing FreeBSD don't use out of the core and into ports? I've never used the FreeBSD FORTRAN compiler, but I do use something that a lot of other people single out as being in this category (uucp), but if uucp were to move to the ports, I'd still use it and FreeBSD. Are there any programs in the base sources for FreeBSD that are written in FORTRAN? To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Nate Williams wrote: A lot of people use a lot of things out of ports. Why should Fortran be different? Because Berkeley Unix has /always/ included a FORTRAN compiler. And they have /always/ included games. Next issue. Mmmm... can I get a VAX port going with this argument? :-) -- Daniel C. Sobral(8-DCS) d...@newsguy.com If you sell your soul to the Devil and all you get is an MCSE from it, you haven't gotten market rate. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
On Thu, 28 Jan 1999, Daniel C. Sobral wrote: Nate Williams wrote: A lot of people use a lot of things out of ports. Why should Fortran be different? Because Berkeley Unix has /always/ included a FORTRAN compiler. And they have /always/ included games. Next issue. Well, there's a little more to this argument. I'm not saying I'm necessarily in favor *myself* of programming in fortran (I like going to the dentist a whole lot more) but to many people, having a fortran compiler that works in FreeBSD is something of a badge of authenticity. Most of these folks are heavy research types, but it's certainly true that having the compiler at least *did* increase FreeBSD's reputation to some degree. I'm not sure if this argument is worth pushing anymore, because FreeBSD's stability and usefulness has become much more well known, but it did contribute at some point, and I think that is the idea that Daniel was trying to convey. Right? +--- Chuck Robey | Interests include any kind of voice or data chu...@glue.umd.edu | communications topic, C programming, and Unix. 213 Lakeside Drive Apt T-1 | Greenbelt, MD 20770 | I run Journey2 and picnic (FreeBSD-current) (301) 220-2114 | and jaunt (NetBSD). +--- To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Ladies and Gents, I have completed the portification of f2c and its support library. Who is going to pick this up? Last time I volunteered, Jordan balked at the idea. Nate To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Ladies and Gents, I have completed the portification of f2c and its support library. Who is going to pick this up? Last time I volunteered, Jordan balked at the idea. I was among the people that asked Steve to do the work; I'd have no trouble with you doing the integration/extraction. You might just want to check that the recent alpha-related changes that were submitted for f2c are covered in the portified version. -- \\ Sometimes you're ahead, \\ Mike Smith \\ sometimes you're behind. \\ m...@smith.net.au \\ The race is long, and in the \\ msm...@freebsd.org \\ end it's only with yourself. \\ msm...@cdrom.com To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Mike Smith wrote: Ladies and Gents, I have completed the portification of f2c and its support library. Who is going to pick this up? Last time I volunteered, Jordan balked at the idea. I was among the people that asked Steve to do the work; I'd have no trouble with you doing the integration/extraction. You might just want to check that the recent alpha-related changes that were submitted for f2c are covered in the portified version. Yes, I have included the recent patch to f2c.h in the port. In the Makefile file in lang/f2c-freebsd, you'll find: .if (${ARCH} == alpha) pre-configure: @cp files/f2c.h.alpha ${WRKSRC}/f2c/f2c.h .endif Someday, I'll install FreeBSD on my alphastation, and then I'll be able to test i386 and alpha architectures. -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Ladies and Gents, I have completed the portification of f2c and its support library. Who is going to pick this up? Last time I volunteered, Jordan balked at the idea. I was among the people that asked Steve to do the work; I'd have no trouble with you doing the integration/extraction. Agreed, I also thought this was a good idea. You might just want to check that the recent alpha-related changes that were submitted for f2c are covered in the portified version. Apparently so. I can't do it *right now* as I don't have a 3.0 box handy, but I should by this weekend. Nate To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
I was among the people that asked Steve to do the work; I'd have no trouble with you doing the integration/extraction. You might just want to check that the recent alpha-related changes that were submitted for f2c are covered in the portified version. This might be a good time to bring this up. Might we create src/opt/ where things we are used to maintaining can be put? Nightly the latest src/opt/foo can be checked out and tarred up. This way everyone can still work on and commit fixes to this software. Just as happened for the Alpha. With a tarball sitting outside of the FreeBSD world, it is hard to be as responsible for the code. Alternately, I guess we could just have the code live in /usr/ports/lang/f2c/src/, but I don't know if Satoshi wants /usr/ports to expand like that. -- -- David(obr...@nuxi.com -or- obr...@freebsd.org) To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
David O'Brien wrote: I was among the people that asked Steve to do the work; I'd have no trouble with you doing the integration/extraction. You might just want to check that the recent alpha-related changes that were submitted for f2c are covered in the portified version. This might be a good time to bring this up. Might we create src/opt/ where things we are used to maintaining can be put? Nightly the latest src/opt/foo can be checked out and tarred up. I agree that there is a gray area between base distribution and ports. Sendmail, postfix, tcl, and others fall into this gray area. I think f2c falls into the ports catagory because it is not under active development (although it is actively maintained) and updating a port every few months seems reasonable. Alternately, I guess we could just have the code live in /usr/ports/lang/f2c/src/, but I don't know if Satoshi wants /usr/ports to expand like that. I think that this is a Bad Idea(tm) even if Satoshi does not. This would still require someone looking over the GNATS database for changes to the software, and then having someone with commit privelege willing to deal with the PR. Each Makefile under the ports systems contains a maintainer line. I do not think it unreasonable for someone to send patchs directly to the maintainer. -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Each Makefile under the ports systems contains a maintainer line. I do not think it unreasonable for someone to send patchs directly to the maintainer. Except that some maintainers dissapear, and maintainers w/o commit abilities still have to get someone to update the port for them. -- -- David(obr...@nuxi.com -or- obr...@freebsd.org) To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
David O'Brien wrote: Each Makefile under the ports systems contains a maintainer line. I do not think it unreasonable for someone to send patchs directly to the maintainer. Except that some maintainers dissapear, and maintainers w/o commit abilities still have to get someone to update the port for them. Yes, I recognize that this is problem. A partial solution might be anoncvs to a shadow tree of the master ports repository. Only those ports in the shadow tree which satisfy portlint and make; make install; make package would get committed to the master repository. Now, if Satoshi has read the above, could someone please revive him and make sure he takes his heart medication ;-) -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
* From: David O'Brien obr...@nuxi.com * Alternately, I guess we could just have the code live in * /usr/ports/lang/f2c/src/, but I don't know if Satoshi wants /usr/ports * to expand like that. Eek. I don't think people will appreciate the ports collection suddenly exploding in size with things like that. (Portlint and tcpblast are exceptions since they are so small.) Satoshi To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
* From: Steve Kargl s...@troutmask.apl.washington.edu * Yes, I recognize that this is problem. A partial solution might * be anoncvs to a shadow tree of the master ports repository. Only * those ports in the shadow tree which satisfy portlint and make; * make install; make package would get committed to the master * repository. * * Now, if Satoshi has read the above, could someone please revive him * and make sure he takes his heart medication ;-) Um, I'm still alive but can someone explain me why this can't be a regular port? Being useful to some but not the majority, no other parts of the system depending on it, this looks like a model citizen in the ideal ports world. :) Satoshi To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Satoshi Asami wrote: * From: Steve Kargl s...@troutmask.apl.washington.edu * Yes, I recognize that this is problem. A partial solution might * be anoncvs to a shadow tree of the master ports repository. Only * those ports in the shadow tree which satisfy portlint and make; * make install; make package would get committed to the master * repository. * * Now, if Satoshi has read the above, could someone please revive him * and make sure he takes his heart medication ;-) Um, I'm still alive but can someone explain me why this can't be a regular port? Being useful to some but not the majority, no other parts of the system depending on it, this looks like a model citizen in the ideal ports world. :) Well, actually I did f2c as a port, and it does indeed fit inside the ports paradigm. Please, see my original email in the thread. -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
I'm not currently balking at the idea of you picking it up - by all means, feel free! :) Just also remember that Peter will at some point be doing an egcs upgrade, so if that has issues for fortran they should be worked out at this time. - Jordan Ladies and Gents, I have completed the portification of f2c and its support library. Who is going to pick this up? Last time I volunteered, Jordan balked at the idea. Nate To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
egcs (was Re: removing f2c from base distribution)
On Tue, 26 Jan 1999, Jordan K. Hubbard wrote: Just also remember that Peter will at some point be doing an egcs upgrade, so if that has issues for fortran they should be worked out at this time. On this matter, I found out the other day that eg++-compiled binaries are not binary-compatible with those produced by gcc (e.g. libraries cannot apparently be linked to by gcc, etc). Has anyone considered the implications of this? Kris - (ASP) Microsoft Corporation (MSFT) announced today that the release of its productivity suite, Office 2000, will be delayed until the first quarter of 1901. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
* Well, actually I did f2c as a port, and it does indeed fit * inside the ports paradigm. Please, see my original email in * the thread. Yes, I know that. I was just wondering why people would want it otherwise. Satoshi To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
I'm not currently balking at the idea of you picking it up - by all means, feel free! :) Roger Wilco, Ok-dokey, good deal. :) Just also remember that Peter will at some point be doing an egcs upgrade, so if that has issues for fortran they should be worked out at this time. Steven? Nate To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Nate Williams wrote: Just also remember that Peter will at some point be doing an egcs upgrade, so if that has issues for fortran they should be worked out at this time. Steven? egcs contains g77 and egcs can be configured to be built with or without g77. My port of f2c, libf2c, and my new f77(1) should not conflict with inclusion of g77. In fact, I think g77 which had used libf2c for its run time library has renamed the library to libg77, so no conflict should arise. The question is whether Peter wants to include g77, and whether people would see this as bloat. I know g77 outperforms f2c+gcc on my real-world benchmarks by a significant margin. If my port of f2c and libf2c and the new f77(1) are found to be adequate, then a committer should remove src/lib/lib{I77,F77,f2c}, src/usr.bin/f2c, and src/gnu/usr.bin/cc/f77. The removal of the last directory will delete our gcc's knowledge of Fortran. -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Um, I'm still alive but can someone explain me why this can't be a regular port? Being useful to some but not the majority, no other parts of the system depending on it, this looks like a model citizen in the ideal ports world. :) Because we loose control over it. There is a move to push some things out of the base system and into ports. Fine, but then how does bug fixes happen? I believe the committed that fixed the f2c problems on the Alpha really liked having the abilility to do so. For many parts of the system, I really think we need to make the optional and installable as ports, but the *maintance* of them should stay along our traditional means. -- -- David(obr...@nuxi.com -or- obr...@freebsd.org) To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Satoshi Asami wrote: * Well, actually I did f2c as a port, and it does indeed fit * inside the ports paradigm. Please, see my original email in * the thread. Yes, I know that. I was just wondering why people would want it otherwise. My original email provided an opportunity to revisit the sendmail versus postfix controversy. That is, some software seems to fall short of support for inclusion in the base distribution but it may be too important for ports. Maybe the package system that I seen discussed will obviate the controversy. -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
The question is whether Peter wants to include g77, and whether people would see this as bloat. I know g77 outperforms f2c+gcc on my real-world benchmarks by a significant margin. A good question, is how easy it is to download egcs-g77-1.1.1.tar.gz and build it into something workable assuming the EGCS C and C++ compilers are part of the system. I've got a Bmaked contribified version of EGCS, but didn't do g77. So maybe a consensus should be made what to do about FORTRAN in the base system. -- -- David(obr...@nuxi.com -or- obr...@freebsd.org) To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
David O'Brien wrote: Um, I'm still alive but can someone explain me why this can't be a regular port? Being useful to some but not the majority, no other parts of the system depending on it, this looks like a model citizen in the ideal ports world. :) Because we loose control over it. There is a move to push some things out of the base system and into ports. Fine, but then how does bug fixes happen? I believe the committed that fixed the f2c problems on the Alpha really liked having the abilility to do so. David, I'm the person who pointed out the problems with f2c on alpha, and my alphastation does not run FreeBSd, yet. Indeed, there were some changes on netlib that deal with 64bit issues and alignment on page boundaries that now exist in the port (and these change may not be in the base distribution). I sent Joerg a patch for current that was 35KB in size. -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
David O'Brien wrote: The question is whether Peter wants to include g77, and whether people would see this as bloat. I know g77 outperforms f2c+gcc on my real-world benchmarks by a significant margin. A good question, is how easy it is to download egcs-g77-1.1.1.tar.gz and build it into something workable assuming the EGCS C and C++ compilers are part of the system. I haven't read the egcs mailinglist in a few weeks, but my impression is that if you download egcs-g77-1.1.1.tar.gz after building the egcs C and C++ compilers, you then have to recompile at a minimum the C compiler. Things may have changed, but g77 is simply a frontend to the gcc backend. It is not a standalone compiler. -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
David O'Brien wrote: I've got a Bmaked contribified version of EGCS, but didn't do g77. So maybe a consensus should be made what to do about FORTRAN in the base system. If you are collecting votes, please add mine; I feel quite strongly (knowing the scientists that I do that use Fortran) that it _should_ be in the base system. Please include it. M -- Mark Murray Join the anti-SPAM movement: http://www.cauce.org To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
David O'Brien wrote: I've got a Bmaked contribified version of EGCS, but didn't do g77. So maybe a consensus should be made what to do about FORTRAN in the base system. If you are collecting votes, please add mine; I feel quite strongly (knowing the scientists that I do that use Fortran) that it _should_ be in the base system. Please include it. We've already established that it doesn't need to be in the base system. Your scientist friends are probably already going to be installing their favorite text editors; it's no harder to install the Fortran support. -- \\ Sometimes you're ahead, \\ Mike Smith \\ sometimes you're behind. \\ m...@smith.net.au \\ The race is long, and in the \\ msm...@freebsd.org \\ end it's only with yourself. \\ msm...@cdrom.com To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
Mike Smith wrote: David O'Brien wrote: I've got a Bmaked contribified version of EGCS, but didn't do g77. So maybe a consensus should be made what to do about FORTRAN in the base system. If you are collecting votes, please add mine; I feel quite strongly (knowing the scientists that I do that use Fortran) that it _should_ be in the base system. Please include it. We've already established that it doesn't need to be in the base system. This may be true of f2c because it is not tightly bound to the FSF compiler technology. Your scientist friends are probably already going to be installing their favorite text editors; it's no harder to install the Fortran support. g77 is a frontend to the FSF compiler backend, and thus it is bound to specific versions. So, it could become a support nightmare to ensure a g77 port is in sync with the egcs backend in the base distribution. It might even be impractical to try to build a standalone g77 port. -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: removing f2c from base distribution
On Tue, Jan 26, 1999 at 10:06:44PM -0800, David O'Brien wrote: The question is whether Peter wants to include g77, and whether people would see this as bloat. I know g77 outperforms f2c+gcc on my real-world benchmarks by a significant margin. A good question, is how easy it is to download egcs-g77-1.1.1.tar.gz and build it into something workable assuming the EGCS C and C++ compilers are part of the system. I've got a Bmaked contribified version of EGCS, but didn't do g77. So maybe a consensus should be made what to do about FORTRAN in the base system. I for one feel that Fortran should remain as part of the base system, either as f2c or g77. I would prefer g77 because of the performance advantage and compatability with Fortran code being ported from other systems. This is something I do quite a bit of. I have contributed a couple of ports that are written in Fortran and I plan on contributing more. I have been waiting to see what decisions were made in this area however before proceeding. The biggest problem has been that the port of g77 has not worked properly for quite some time and in fact is currently marked as broken. I would anticipate that this situation would not change much in the future if the base gcc (egcs?) is modified far enough away from a standard gcc distribution, as is currently the case with our gcc. As far as getting g77 from the egcs port, well, the release versions have been fine for g77 but the snapshots have been hit and miss. The ports system has not provided a reliable means of Fortran support, IMHO. However, if g77 were part of the base FreeBSD system, assuming f2c is ripped out, then Fortran support would be gaurenteed to be there when needed. I understand that most people using FreeBSD are using it for server tasks and C development. However, FreeBSD is also an excellent OS for a scientific workstation, and that means Fortran is essential. I run a farm of 6 dual CPU PPro/PII systems running quantum chemical calculations 24/7. I do my part to get my colleagues to try FreeBSD instead of NT and Linux. Removing Fortran support from the base system will make that a tougher job. Thanks. -- Glenn Johnson gljo...@bellsouth.net To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
removing f2c from base distribution
Ladies and Gents, I have completed the portification of f2c and its support library. In principle, src/usr.bin/f2c, src/lib/{libI77,libF77,libf2c}, and src/gnu/usr.bin/cc/f77 can be moved into the attic in -current (4.x). Appropriate adjustments to the Makefile files in src/usr.bin, src/lib, and src/gnu/usr.bin/cc need to be made. ftp://troutmask.apl.washington.edu/pub/f2c-freebsd.2.0.1.tar.gz ftp://troutmask.apl.washington.edu/pub/f2c-freebsd.tgz ftp://troutmask.apl.washington.edu/pub/f77-freebsd.0.3.tar.gz ftp://troutmask.apl.washington.edu/pub/f77-freebsd.tgz f2c-freebsd.2.0.1.tar.gz is a version of f2c and its library where I have merged the version in the FreeBSD source tree from Dec 1998 with the latest version of f2c and its library from www.netlib.org. The Makefile in f2c-freebsd.2.0.1/libf2c is setup to build only ELF libraries which is reasonable because this is as a replacement for functionality in a post-elf-transition source tree. f2c-freebsd.tgz is a gzipped tar file of the port. It should be placed in ports/lang. When unpacked it will produce a directory named f2c-freebsd, and it should be able to be built on both i386 and alpha axp architectures. f77-freebsd.0.3.tar.gz contains the source code for a new driver utility that is meant to replace the current f77(1) in src/gnu/usr.bin/cc/f77. By default, the new f77 will use Sun Microsystem's Fortran preprocessor (ports/devel/fpp), but it can be built to use GNU cpp. The new f77 recognizes all f2c and fpp (or cpp) options that make sense in the context of compilation. Any option not recognized as a valid f2c or fpp (or cpp) option is automatically passed to gcc except for gcc options that take space delimited arguments (these aren't supported, yet). f77-freebsd.tgz is a gzipped tar file of the port. It should be placed in ports/lang. When unpacked it will produce a directory named f77-freebsd. The new f77(1) should be architecture independent, but I don't have an alpha axp machine for testing. NOTE: Do *NOT* try to use the f2c port with the old f77(1) from src/gnu/usr.bin/cc/f77. The loader can't find the f2c.h header file or the new library locations. -- Steve finger ka...@troutmask.apl.washington.edu http://troutmask.apl.washington.edu/~clesceri/kargl.html To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message