From lechner@cs.uml.edu Fri Sep 3 02:25:20 2004 From: Bob Lechner Subject: hominid test using your olc vs. Team1's To: ajlopez@cs.uml.edu (Alfredo Lopez) Date: Fri, 3 Sep 2004 02:25:19 -0400 (EDT) Hello Alfredo. You have worked long and hard at this, and we have sometimes been at cross-purposes. Since genv12 cannot produce a new usable pr_util fileset, lets agree on a fallback position: just use Team1's pr_util and application code and input data and your new olc version of the state model interpreter. The acceptance test is output data files from run phase that agree with Team1's test - that is , when run with input file Hominid.dat they produce hompathlog.txt and hompathlogDB1.dat files that bde can replay. I did this by cp -r ajlopez/lcp/test lechner then running LCPmain.exe with HominidDB.dat as input. That produced 2 new files in my copy of your lcp/test directory: ------------------ pwd /nfs/earth/misc/proj3/case/04s522/lechner/lcp/test mercury.cs.uml.edu(66)> lg hompath* lg hompath* -rw-r--r-- 1 lechner fac 8005 Sep 3 01:35 hompathlogDB1.dat -rw-r--r-- 1 lechner fac 421 Sep 3 01:35 hompathlog.txt ------------------- (hompathlog.txt is incomplete - compare to $CASE/02s522/Team1/hompathlog.txt saturn.cs.uml.edu(44)> lg $CASE/02s522/Team1/hompath* lg $CASE/02s522/Team1/hompath* ... 14808 May 9 2002 /usr/proj3/case/02s522/Team1/hompathlog.txt ... 9007 May 9 2002 /usr/proj3/case/02s522/Team1/hompathlogDB1.dat ------------------ diff hompathlogDB1.dat /usr/proj3/case/02s522/Team1/hompathlogDB1.dat | wc 143 678 9062 -------------------- Where are your versions of these two output files frm your test run? I reran your test and generated these missing files: hompathlog.txt and hompathlogDB1.dat (see below) They do NOT agree with Team1's results: See my report in $CASE/04s522/lechner/lcp/test/hompathlogDB1.vs.Team1.dif This should have been discovered weeks ago, and is why I insist on a complete build and test typescript. I don't have time to look in your other directories right now. R Lechner -------------------- pwd /nfs/earth/misc/proj3/case/02s522/Team1 saturn.cs.uml.edu(19)> lg *.dat *.txt lg *.dat *.txt -rw-r--r-- 1 lechner 91573-ta 13721 May 9 2002 Hominid.dat -rw-r--r-- 1 lechner 91573-ta 7453 May 9 2002 HominidDB.dat -rw-r--r-- 1 lechner 91573-ta 14808 May 9 2002 hompathlog.txt -rw-r--r-- 1 lechner 91573-ta 9007 May 9 2002 hompathlogDB1.dat -rw-r--r-- 1 lechner 91573-ta 7613 May 9 2002 logdata.txt -------------------- More comments are below. R Lechner > From ajlopez@cs.uml.edu Thu Sep 2 21:38:18 2004 > Date: Thu, 2 Sep 2004 21:38:18 -0400 (EDT) > From: Alfredo Lopez > To: Bob Lechner > Subject: Re: HominibBuild040829 - OK to use Team1's pr_util with your olc. > On Thu, 2 Sep 2004, Bob Lechner wrote: > > > > > From ajlopez@cs.uml.edu Thu Sep 2 17:37:15 2004 > > > Subject: Re: HominibBuild040829 - the problem is more complex - bde does > > > eplay... > > > Hello, Dr. Lechner! > > > > > > I know you told me it was going to be fair game to make it work. I fixed > > > all problems, but it just gets an infinity loop. So, I check > > > > Where is the typescript that shows this loop? > > > The loop is due that LCPmain.c calls pr_load function, and then pr_load > function calls pr_replay function, and pr_replay calls pr_load against. > I couldn't figure out how it works. And I'm not sure if the loop is due to > I was fixing compilation's errors, or something else. I could remake the > project and rerun the loop after executing script, but I though it was not > helpful at all. If you want I can do it. > Do it. I still don't know that it is a loop. pr_load of logfile.dat is SUPPOSED to discover "SR" command and call pr_replay. pr_replay is SUPPOSED to call pr_init and pr_load on checkpoint file logfileDB1.dat. You don't say what else happens. > > > > > Jaganathan's srcMain, and there's not any different. I gave up generating > > > > Not any different than what? My diffs show src and srcMain differ a lot. > > > I mean I have to fix exactly the same errors, and that part of source code > is the same. I checked how srcMain directory code was compiled/linked, and > the executable is ../executables/alpha/chgen12. I run that executable and > I got the same result. I checked it out in the file Again, I don't know if you just used the existing executable or if you recompiled genv12 twice and made two different executables. (It can't be from two source code versions at the same time.) The date on genv12Make011219.2100 doesn't seem to be yours? The link command uses .o files but I don't know if they were recompiled first? I don't know if the link succeeded or the old executable remains? Do ls -alst to find out which file is newest. > > /usr/proj3/case/gen/ver_12/sjaganat/srcMain/genv12Make011219.2100 > > gcc -ansi -g -o ../executables/alpha/chgen12 add_pc_cp_entries.o > change_file_exten.o chgen.o chgen_global_vars.o chgen_release_date.o > close_files.o gen_c_files.o gen_define$ >From srcMain/genv12Make011219.2100: [saturn.cs.uml.edu1](61)> pwd /usr/proj3/case/gen/ver_12/sjaganat/src [saturn.cs.uml.edu1](62)> lg ../exec*/alpha total 1200 -rwxrwxrwx 1 sjaganat bdelog 1219632 Dec 19 21:01 chgen12 This executable was compiled on 19 Dec 2001. Where is the one you compiled from src and from srcMain? I'd expect you to copy these two src directories into different test directories and rerun the genv12 Makefile in both (after renaming srcMain so same Makefile works) then compare results. $CASE/gen/ver_12/sjaganat/diffsrc.vs.srcMain.hlp - RJL 040831 explains pr_log.c diffs between genv12 outputs from src or srcMain. > [saturn.cs.uml.edu1](54)> vi ~/man/gcc.man > > > > > pr_util's files. Team1 pr_util's files are the one with that proyect. I just > > > renamed th.h to Hominid.h, and I changed include "th.h" by include > > > "Hominid.h". > > > > You can stop working on trying to upgrade Team1's pr_util to newer genv12. > > Just use their pr_util files since you did not change their schema. > > > > I did it. I just rename the.h file to Hominid.h, because you said that > should be the name. > > > > So now you have Team1 and Team1 pr_util, generated from Team1 .sch > > by their genv12 run, and their Hominid application code, > > that should still run with your new olc code. > > YOu changed olc but not the .sch content, so your new olc > > should work with other previous Team1 elements. > > > > Does it work? If not, why not? > > > > That's exactly what works from the beginning. Where is the typescript including logfile.txt and HominidDB1.dat from your run of the Hominid navigate.c executable with olc state model interpreter and Team1's pr_util? TO verify these files agree with Team1's, show the diff output in your typescript. > > > `Without knowing why your olc won't work with Hominid, > > I'm reluctant to leave Hominid and begin another applicaiton. > > I don't have the time since Fall is here. > > > > olc works well with Team1 application. What doesn't work at all is > generating new pr_util files. > Good. Then show me the COMPLETE typescript (start by compiling pr_util not running genv12), buiold and run both phases and include ls -alst of the directories that contain new code and output data files. Then finish your report based on this. > > > > Could you give setgame source code path, or microwave controller path to > > > work on the sub-problem. I'm sorry if you gave me it before -source > > > code path-, but I would like to work with the project you have been > > > executing the most. I mean you know exactly its output behaviors. > > > > I don't have any more experience with setGame than with Hominid. > > > > What I mean is a source code project that you know works well. A source > code you have tested several time, and you know how it behaves when it's > compiled, linked, and executed. It will be faster to add new olc code, > because I don't spend any time fixing code. If it works ok, adding new > code should be easier. > > > > > > > > Alfredo Lopez > > > > > > Really, I doubt that source code works correctly. It could work just for > > > one application, and fixing the output -source code. > > > > > I don't understand what you mean - please be more specific. > > Without rerunning genv12, Olc is the only thing you changed, > > as far as I know. > > I was talking about genv12. > > > > > > [cut]