From dpelleg+ at cs.cmu.edu Tue Jul 2 17:36:11 2002 From: dpelleg+ at cs.cmu.edu (Dan Pelleg) Date: Tue, 2 Jul 2002 17:36:11 -0400 Subject: [auton-users] nothing here, move along Message-ID: <15650.7371.992688.490746@gs166.sp.cs.cmu.edu> We've upgraded a few servers in the lab during the last few days. There were short disruptions of service, in particular around about 3pm today. In any case, things seem to be back to normal now. Please report (on this list) any problems you observe. From awm at cs.cmu.edu Mon Jul 8 10:55:47 2002 From: awm at cs.cmu.edu (Andrew W. Moore) Date: Mon, 08 Jul 2002 10:55:47 -0400 Subject: [auton-users] Lab temperature Message-ID: <20020708145601.CBAA3B4@mail.autonlab.org> Jim, During the weekend (and continuing now) the Auton Lab temperature has been hitting 90...it's even hoter than the surrounding corridors. Paul, Dan: Should we turn off the rack mounted alphas and if so how do I do it and how do I alert the users? Thanks, Andrew From jmjoseph at andrew.cmu.edu Mon Jul 8 14:12:01 2002 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Mon, 08 Jul 2002 11:12:01 -0700 Subject: [auton-users] Lab temperature References: <20020708145601.CBAA3B4@mail.autonlab.org> Message-ID: <3D29D5F1.10001@andrew.cmu.edu> I just checked the temperatures of the UPS's (they are the best way to remotely check the temperature). The hottest internal temperature is running right around 28C, so I think we're very safe. I'll check later to see if it goes up later. If we do end up turning them off, this list is the place to let everyone know. On the technical side, I can shut down any of the machines to the bios, where they should produce minimal heat. They're easier to start back up that way. -Jacob Andrew W. Moore wrote: >Jim, > During the weekend (and continuing now) the Auton Lab temperature >has been hitting 90...it's even hoter than the surrounding >corridors. > > Paul, Dan: Should we turn off the rack mounted alphas and if >so how do I do it and how do I alert the users? > > Thanks, > > Andrew > > From dpelleg+ at cs.cmu.edu Mon Jul 8 14:14:03 2002 From: dpelleg+ at cs.cmu.edu (Dan Pelleg) Date: Mon, 8 Jul 2002 14:14:03 -0400 Subject: [auton-users] Lab temperature In-Reply-To: <20020708145601.CBAA3B4@mail.autonlab.org> References: <20020708145601.CBAA3B4@mail.autonlab.org> Message-ID: <15657.54891.312731.889343@gargle.gargle.HOWL> Andrew W. Moore writes: > Jim, > During the weekend (and continuing now) the Auton Lab temperature > has been hitting 90...it's even hoter than the surrounding > corridors. > > Paul, Dan: Should we turn off the rack mounted alphas and if > so how do I do it and how do I alert the users? > > Thanks, > > Andrew The remotely-accessible sensors we have don't indicate anything out of the ordinary. For the record, these are: 1) internal sensors on two of the UPS units 2) the window in front of me, through which I see no smoke coming up from NSH. I'll go to the lab soon to have another look and I'll decide if there's stuff we need to turn off. If I do, I won't bother with giving an advance warning. This list is (supposed) to be the right forum to announce these things. We should probably get a digital sensor and hook it up so we get email when the lab gets warm. Maybe even it'll send it to this list. Something along the lines of "this would be a good time to save, logout, and RUN! RUN! RUN!". From David.Cohn at acm.org Mon Jul 8 14:31:28 2002 From: David.Cohn at acm.org (David 'Pablo' Cohn) Date: Mon, 8 Jul 2002 14:31:28 -0400 Subject: [auton-users] Lab temperature In-Reply-To: <15657.54891.312731.889343@gargle.gargle.HOWL> Message-ID: The temperature in the lab has come down quite a bit in the last hour. It's now actually pleasantly cool here. This morning, however, it was much warmer in a432 than in the adjacent (warm) corridor. I assume someone somewhere has flipped a switch. -David -----Original Message----- From: users-admin at autonlab.org [mailto:users-admin at autonlab.org]On Behalf Of Dan Pelleg Sent: Monday, July 08, 2002 2:14 PM To: users at autonlab.org Subject: [auton-users] Lab temperature Andrew W. Moore writes: > Jim, > During the weekend (and continuing now) the Auton Lab temperature > has been hitting 90...it's even hoter than the surrounding > corridors. > > Paul, Dan: Should we turn off the rack mounted alphas and if > so how do I do it and how do I alert the users? > > Thanks, > > Andrew The remotely-accessible sensors we have don't indicate anything out of the ordinary. For the record, these are: 1) internal sensors on two of the UPS units 2) the window in front of me, through which I see no smoke coming up from NSH. I'll go to the lab soon to have another look and I'll decide if there's stuff we need to turn off. If I do, I won't bother with giving an advance warning. This list is (supposed) to be the right forum to announce these things. We should probably get a digital sensor and hook it up so we get email when the lab gets warm. Maybe even it'll send it to this list. Something along the lines of "this would be a good time to save, logout, and RUN! RUN! RUN!". From dpelleg+ at cs.cmu.edu Mon Jul 8 14:37:52 2002 From: dpelleg+ at cs.cmu.edu (Dan Pelleg) Date: Mon, 8 Jul 2002 14:37:52 -0400 Subject: [auton-users] Lab temperature In-Reply-To: <3D29D5F1.10001@andrew.cmu.edu> References: <20020708145601.CBAA3B4@mail.autonlab.org> <3D29D5F1.10001@andrew.cmu.edu> Message-ID: <15657.56320.58777.834857@gargle.gargle.HOWL> Just back from the lab: the BigPapa disk array recorded a max of 50C (CPU) and 56C (motherboard). That's dangerously high. The BigMomma array doesn't record max temps so we can't tell how warm it got. At any rate, they're both much cooler now. David "asbestos skin" Cohn, who is actually in the lab, reports things have chilled down significantly since Andrew's original mail. But it's important to know that the sensors we have don't give us the whole picture. From skees at cs.cmu.edu Sun Jul 14 20:10:58 2002 From: skees at cs.cmu.edu (Jim Skees) Date: Sun, 14 Jul 2002 20:10:58 -0400 Subject: [auton-users] Re: Lab temperature Message-ID: <5.1.0.14.2.20020714201051.023fcb60@ux5.sp.cs.cmu.edu> Has this problem been fixed? -- Jim At 10:55 AM 7/8/02 -0400, Andrew W. Moore wrote: >Jim, > During the weekend (and continuing now) the Auton Lab temperature >has been hitting 90...it's even hoter than the surrounding >corridors. > > Paul, Dan: Should we turn off the rack mounted alphas and if >so how do I do it and how do I alert the users? > > Thanks, > > Andrew -------------- next part -------------- An HTML attachment was scrubbed... URL: From komarek at andrew.cmu.edu Tue Jul 16 09:43:39 2002 From: komarek at andrew.cmu.edu (Paul Komarek) Date: Tue, 16 Jul 2002 09:43:39 -0400 (EDT) Subject: [auton-users] Re: Lab temperature In-Reply-To: <5.1.0.14.2.20020714201051.023fcb60@ux5.sp.cs.cmu.edu> Message-ID: It has. Thanks! -Paul Komarek On Sun, 14 Jul 2002, Jim Skees wrote: > Has this problem been fixed? -- Jim > > At 10:55 AM 7/8/02 -0400, Andrew W. Moore wrote: > >Jim, > > During the weekend (and continuing now) the Auton Lab temperature > >has been hitting 90...it's even hoter than the surrounding > >corridors. > > > > Paul, Dan: Should we turn off the rack mounted alphas and if > >so how do I do it and how do I alert the users? > > > > Thanks, > > > > Andrew > From dpelleg+ at cs.cmu.edu Sun Jul 21 08:00:24 2002 From: dpelleg+ at cs.cmu.edu (Dan Pelleg) Date: Sun, 21 Jul 2002 08:00:24 -0400 Subject: [auton-users] limey In-Reply-To: References: Message-ID: <15674.41560.296580.271172@gargle.gargle.HOWL> nichol writes: > > > Hi Dan & Joseph, > > Is limey down? Yes. > Is it because of the storms? Don't think so. It was brought back up after Friday's 2-hour power outage. But there can still be either resulting damage to hardware or improperly initialized software services that'll require a reboot. > When will it be back? The earlier of: * As soon as someone manages to boot it remotely (I tried and failed) * When I get to campus Monday morning and BRS it > > Cheers > > Bob > > From komarek at andrew.cmu.edu Mon Jul 22 17:09:06 2002 From: komarek at andrew.cmu.edu (Paul Komarek) Date: Mon, 22 Jul 2002 17:09:06 -0400 (EDT) Subject: [auton-users] New gcc problems Message-ID: Hello everyone, SCS GNU/Linux is (somewhat strangely) using gcc 3.1 with their RedHat-derived GNU/Linux distrobution. This causes subtle problems. gcc 3.1 supports C99 (a new standard for C) 'features' that are not supported by C89 or g89 (GNU extensions to C89). Among these C99 features is one which allows mixed declarations (of variables) and code. For example, the following is legal in C99 but not C89 or g89: int main(void) { printf("Hello, world\n"); int foo; return 0; } Any code like this will cause problems for gcc 2.96 and probably other C compilers. I feel it would be better if such code did not compile. Forcing gcc 3.1 to throw an error for this code requires the switch "-pedantic-errors". I would like to add this flag to our Make.common in the gmake-magic project. It would only be used when compiling with gcc, and would not affect g++ or any other compilers. However, this flag requires that we clean up other problems in our sources, such as source files that are empty after preprocessing (e.g. draw/fgraph.c). I cannot clean up the sources myself, and would need cooperation. That is why I am soliciting lab opinion before starting this project. If you oppose the addition of -pedantic-errors to our gcc compile lines, please let me know. Please provide a better reason than "C sux0rs and C++ r0Xors, so mixed-code-and-declarations should be forced down people's throats" (translation: "C should be like C++, because C++ is cool"). Anyone who does not object is implicitly volunteering 100% cooperation when I need help fixing their sources. =-) If you would like to see the effects of -pedantic-errors now, try compiling with gmake user_flags="-pedantic-errors" -Paul Komarek From komarek at andrew.cmu.edu Mon Jul 22 17:27:35 2002 From: komarek at andrew.cmu.edu (Paul Komarek) Date: Mon, 22 Jul 2002 17:27:35 -0400 (EDT) Subject: [auton-users] New gcc problems In-Reply-To: <200207222121.g6MLLjvc003806@mx2.andrew.cmu.edu> Message-ID: On Mon, 22 Jul 2002, Andrew W. Moore wrote: > I fully support your suggestion, except it's a shame to lose // > comments (and you might find complaints from others about that). > If it was possible to tell the current version of gcc to behave like > it did 6 months ago that would be even better. > It may be possible to re-enable certain features. I am not a gcc or C expert, nor have I researched how to do this. If such things are important to anyone, I would recommend complaining to SCS since they evidently have a compiler-expert who made the decision to replace Red Hat's default C compiler. -Paul Komarek From dpelleg+ at cs.cmu.edu Mon Jul 22 17:36:55 2002 From: dpelleg+ at cs.cmu.edu (Dan Pelleg) Date: Mon, 22 Jul 2002 17:36:55 -0400 Subject: [auton-users] New gcc problems In-Reply-To: References: <200207222121.g6MLLjvc003806@mx2.andrew.cmu.edu> Message-ID: <15676.31479.783382.737107@gs-eval1.fac.cs.cmu.edu> I don't mind losing // comments. In fact, I rather not have them at all. In the past I went through a lot of effort to clear the code from them, as well as other gcc-isms, so that it compiles under Compaq's compiler. I don't find them that useful or time-saving to begin with. For easy typing, we can set up an editor macro. Paul Komarek writes: > > On Mon, 22 Jul 2002, Andrew W. Moore wrote: > > > I fully support your suggestion, except it's a shame to lose // > > comments (and you might find complaints from others about that). > > > If it was possible to tell the current version of gcc to behave like > > it did 6 months ago that would be even better. > > > > It may be possible to re-enable certain features. I am not a gcc or C > expert, nor have I researched how to do this. If such things are > important to anyone, I would recommend complaining to SCS since they > evidently have a compiler-expert who made the decision to replace Red > Hat's default C compiler. > > -Paul Komarek > From komarek at andrew.cmu.edu Mon Jul 22 17:39:49 2002 From: komarek at andrew.cmu.edu (Paul Komarek) Date: Mon, 22 Jul 2002 17:39:49 -0400 (EDT) Subject: [auton-users] Another Make.common change Message-ID: For anyone bitten by the gcc on Alpha fp exception bug: I have added -mieee for all Alpha compiles with GNU compilers, and -ieee for all Compaq compilers. This *should* end the random fp exception reign of terror, and allow you to use the Alphas with confidence and in peace. According to the man page, it might also have a minor performance impact on your floating-point code. To get the new Make.common, cd to your equivalent of h/gmake-magic and do a "cvs update" or "cvs update Make.common". -Paul Komarek From awm at cs.cmu.edu Mon Jul 22 17:47:37 2002 From: awm at cs.cmu.edu (Andrew W. Moore) Date: Mon, 22 Jul 2002 17:47:37 -0400 Subject: [auton-users] New gcc problems Message-ID: <20020722214800.7AB34D4@mail.autonlab.org> Okay, I am convinced by Dan...we should also disallow // comment chars. So my vote is in favor of setting the default to pedantic mode. Does anyone have the ability to write an ingenious script to automatically turn all the current //'s into /* .. */'s? Andrew From dpelleg+ at cs.cmu.edu Mon Jul 22 18:06:22 2002 From: dpelleg+ at cs.cmu.edu (Dan Pelleg) Date: Mon, 22 Jul 2002 18:06:22 -0400 Subject: [auton-users] New gcc problems In-Reply-To: <20020722214800.7AB34D4@mail.autonlab.org> References: <20020722214800.7AB34D4@mail.autonlab.org> Message-ID: <15676.33246.699056.491135@gs-eval1.fac.cs.cmu.edu> Andrew W. Moore writes: > Okay, I am convinced by Dan...we should also disallow // comment chars. > So my vote is in favor of setting the default to pedantic mode. > > Does anyone have the ability to write an ingenious script to automatically > turn all the current //'s into /* .. */'s? > > Andrew I've actually done that for quite a bit of code. There are still files I wasn't using at the time, and so weren't modified. To see which ones they are, do this in your h/ directory: fgrep -l "//" */*.[ch] To fix a single file, in emacs: M-x query-replace-regexp //\(.*\) /* \1 */ // Note that it'll make comment blocks // that look like this // very ugly /* Note that it'll make comment blocks */ /* that look like this */ /* very ugly */ A script, well, this one here should do it (I didn't test it), but I would advise against using scripts on code. For example, there are several URLs in there, and we don't particulary want our users to see welcome messages of the form: please log on to http:/* www.autonlab.org */ ****************************** AN UNTESTED SCRIPT THAT'LL POTENTIALLY CHANGE ALL OF YOUR SOURCE FILES AND MAKE THEM UNCOMPILABLE: perl -i.bak -e 's#//(.*)#/* \1 */#' */*.[ch] From David.Cohn at acm.org Mon Jul 22 18:12:11 2002 From: David.Cohn at acm.org (David Cohn) Date: Mon, 22 Jul 2002 18:12:11 -0400 Subject: [auton-users] New gcc problems In-Reply-To: <20020722214800.7AB34D4@mail.autonlab.org> Message-ID: Arrggggggghh.... uh, I'm really attached to my //, if for no other reason that one doesn't get into those midn-boggling nested comment confusions when you can't comment out a block of code (using /* */) because the code you're commenting out has a */ in it. But I'm willing to adapt, especially if someone can write a clever "//"->/* */ script. -d -----Original Message----- From: users-admin at autonlab.org [mailto:users-admin at autonlab.org]On Behalf Of Andrew W. Moore Sent: Monday, July 22, 2002 5:48 PM To: users at autonlab.org Cc: awm at cs.cmu.edu Subject: Re: [auton-users] New gcc problems Okay, I am convinced by Dan...we should also disallow // comment chars. So my vote is in favor of setting the default to pedantic mode. Does anyone have the ability to write an ingenious script to automatically turn all the current //'s into /* .. */'s? Andrew From chong2 at andrew.cmu.edu Mon Jul 22 18:17:27 2002 From: chong2 at andrew.cmu.edu (Shang-Shan Chong) Date: Mon, 22 Jul 2002 18:17:27 -0400 Subject: [auton-users] New gcc problems In-Reply-To: <20020722214800.7AB34D4@mail.autonlab.org>; from awm@cs.cmu.edu on Mon, Jul 22, 2002 at 05:47:37PM -0400 References: <20020722214800.7AB34D4@mail.autonlab.org> Message-ID: <20020722181727.A4217@dipole> On Mon, Jul 22, 2002 at 05:47:37PM -0400, Andrew W. Moore wrote: > Okay, I am convinced by Dan...we should also disallow // comment chars. > So my vote is in favor of setting the default to pedantic mode. > > Does anyone have the ability to write an ingenious script to automatically > turn all the current //'s into /* .. */'s? > > Andrew Hi, I'm normally a silent observer here, but the following perl script should do the trick, since // are single line comments but /* */'s are multi-line. It's a code fragment from something I've done before. #!/usr/bin/perl while (<>) { s/\/\/(.*)$/\/\*\1\*\//g; print; } css. -- Shang-Shan CHONG Dept of Physics, Carnegie Mellon University office: Wean Hall 8419 tel: (412) 268-4206 email: chongss+ at cmu.edu From agray at cs.cmu.edu Mon Jul 22 18:48:45 2002 From: agray at cs.cmu.edu (agray at cs.cmu.edu) Date: Mon, 22 Jul 2002 18:48:45 -0400 Subject: [auton-users] New gcc problems In-Reply-To: Your message of "Mon, 22 Jul 2002 17:36:55 EDT." <15676.31479.783382.737107@gs-eval1.fac.cs.cmu.edu> Message-ID: <20020722224918.F3D2F7E@mail.autonlab.org> Hmm, I hate to be a thorn but... 1. I like mid-code declarations because they increase code clarity as well as locality, which can allow better code optimization (ie. SPEED). I believe that localizing variables is a good programming habit in general, up to the limit supported by the parser. 2. I like // comments because they remove extraneous characters and make my code more visually compressible in some cases - which I claim helps the readability of complex algorithms, at least for me. (Much less important than mid-code declarations though.) Of course if these departures from the classic K&R or pre-1999 ANSI C have a real cost for the general portability of Auton code, that would take priority I would say. (Note that I am not a C++-head, for portability and other pragmatic reasons.) Alex _______________________________________________________ Alexander G. Gray http://www.cs.cmu.edu/~agray CMU Computer Science office 412-268-3070 9-9pm EST From chong2 at andrew.cmu.edu Mon Jul 22 18:43:08 2002 From: chong2 at andrew.cmu.edu (Shang-Shan Chong) Date: Mon, 22 Jul 2002 18:43:08 -0400 Subject: [auton-users] New gcc problems In-Reply-To: <20020722181727.A4217@dipole>; from chong2@andrew.cmu.edu on Mon, Jul 22, 2002 at 06:17:27PM -0400 References: <20020722214800.7AB34D4@mail.autonlab.org> <20020722181727.A4217@dipole> Message-ID: <20020722184308.A4289@dipole> On Mon, Jul 22, 2002 at 06:17:27PM -0400, chong2 at andrew.cmu.edu wrote: > > On Mon, Jul 22, 2002 at 05:47:37PM -0400, Andrew W. Moore wrote: > > Okay, I am convinced by Dan...we should also disallow // comment chars. > > So my vote is in favor of setting the default to pedantic mode. > > > > Does anyone have the ability to write an ingenious script to automatically > > turn all the current //'s into /* .. */'s? > > > > Andrew > > Hi, > > I'm normally a silent observer here, but the following perl script > should do the trick, since // are single line comments but /* */'s are > multi-line. It's a code fragment from something I've done before. > > #!/usr/bin/perl > while (<>) { > s/\/\/(.*)$/\/\*\1\*\//g; > print; > } After seeing Dan's comment, I realize that even http://... stuff would be changed, so the previous regexp is no good. This is the new regexp, which relies on the fact that stuff in comments should not be matched. #!/usr/bin/perl while (<>) { s/^(.+?)\/\/(.*)$/\1\/\*\2\*\//; print; } It still does not fix anything inside existing /* */'s, but should be fine for fixing code with only // as comments. css. -- Shang-Shan CHONG Dept of Physics, Carnegie Mellon University office: Wean Hall 8419 tel: (412) 268-4206 email: chongss+ at cmu.edu From komarek at andrew.cmu.edu Mon Jul 22 19:37:56 2002 From: komarek at andrew.cmu.edu (Paul Komarek) Date: Mon, 22 Jul 2002 19:37:56 -0400 (EDT) Subject: [auton-users] New gcc problems In-Reply-To: <20020722224918.F3D2F7E@mail.autonlab.org> Message-ID: On Mon, 22 Jul 2002 agray at cs.cmu.edu wrote: > Hmm, I hate to be a thorn but... > Of course if these departures from the classic K&R or pre-1999 ANSI C > have a real cost for the general portability of Auton code, that would > take priority I would say. (Note that I am not a C++-head, for > portability and other pragmatic reasons.) Portability is the issue I'm trying to address. I don't want to end up writing sed scripts off-site in front of some third party on a locked-down HPUX workstation, trying to make our code compile with their C89/G89 compiler, simultaneoulsly explaining why C99 is the way of the future. I'm interested in *any* solution that can prevent this. The easiest is probably to forbid C99, G89, and G99 extensions, but easy doesn't always mean best. -Paul From agray at cs.cmu.edu Mon Jul 22 20:42:35 2002 From: agray at cs.cmu.edu (agray at cs.cmu.edu) Date: Mon, 22 Jul 2002 20:42:35 -0400 Subject: [auton-users] New gcc problems In-Reply-To: Your message of "Mon, 22 Jul 2002 19:37:56 EDT." Message-ID: <20020723004320.670B97E@mail.autonlab.org> >Portability is the issue I'm trying to address. I don't want to end up >writing sed scripts off-site in front of some third party on a locked-down >HPUX workstation, trying to make our code compile with their C89/G89 >compiler, simultaneoulsly explaining why C99 is the way of the future. By all means, if you think this is likely to happen, it is better for us to live without the nice language extensions. The aim of Auton should be to spread unstoppably throughout the universe, feeding upon its fat life-giving datasets, crushing all inferior algorithms which dare to stand in its path. Alex From jmjoseph at andrew.cmu.edu Tue Jul 23 16:47:50 2002 From: jmjoseph at andrew.cmu.edu (Jacob M. Joseph) Date: Tue, 23 Jul 2002 16:47:50 -0400 (EDT) Subject: [auton-users] BigMomma-tmp In-Reply-To: <20020723004320.670B97E@mail.autonlab.org> Message-ID: Hi. I just wanted to alert everyone to a change of policies on /mnt/BigMomma-tmp. Since it was created, it's had very little use--a waste of 100GB. As it is writable by all, I originally wiped it clean at 5am every Sunday, but this has turned out to be far to frequent. It is now wiped the first Sunday of every month at 5am. Hopefully everyone will be more apt to use it. If the usage still remains low, I could wipe it even less frequently, or work towards reallocating part of it. -Jacob Joseph From jmjoseph at andrew.cmu.edu Thu Jul 25 06:32:44 2002 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Thu, 25 Jul 2002 03:32:44 -0700 Subject: [auton-users] Limey's back Message-ID: <3D3FD3CC.7020707@andrew.cmu.edu> I'm glad to report that Limey has reopened for business. It had some SSHD problems that took some time to debug. Let me know if you get anything odd at login. -Jacob