From i.normann at iu-bremen.de Tue Aug 1 14:58:36 2006 From: i.normann at iu-bremen.de (i.normann at iu-bremen.de) Date: Tue, 1 Aug 2006 14:58:36 +0200 (CEST) Subject: [Hets-devel] compile problems with release 0.6 Message-ID: <54005.84.161.237.217.1154437116.squirrel@mail2web.iu-bremen.de> Dear hets developers, currently I am trying - without success - to compile hets release 0.6. More precisely I downloaded from http://www.informatik.uni-bremen.de/ agbkb/forschung/formal_methods/CoFI/hets/src-distribution/versions/ these tarballs: Hets-src-0.60.tgz uni-ghc-6.4.x-2006-02-01.tgz programatica-2006-05-31.tgz uDrawGraph-3.1.1-0-linux-i386.tar.gz I extracted all tarballs into one directory ~/hets; i.e. ~/hets> ls HetCATS programatica uDrawGraph-3.1 uni I followed the install instructions: uDrawGraph-3.1: export UDG_HOME="${HOME}/hets/uDrawGraph-3.1" uni: ./configure --enable-Het make packages HetCATS: make The last failed with: Could not find module `PropParser': use -v to see a list of the files searched for (imported from ./Haskell/HatParser.hs) make: *** [hets] Error 1 I tried to fix this by adjusting in the Makefile the path for programmatica; i.e. adjusting PFE_DIRS. In fact the many files couln't be found in the PFE_DIRS as it is defined in the Makefile which comes with Hets-src-0.60.tgz - at least it doesn't fit to programatica-2006-05-31.tgz. So is there any version of programmatica that is compatible with Hets-src-0.60.tgz? After all I don't what version of uni/uDrawGraph/programmatica/HetCATS I have as long as it is reasonable uptodate and - much more important - compilable! Is there any reliably compilable combination of these packages? Best, Immanuel From maeder at tzi.de Tue Aug 1 16:34:16 2006 From: maeder at tzi.de (Christian Maeder) Date: Tue, 01 Aug 2006 16:34:16 +0200 Subject: [Hets-devel] compile problems with release 0.6 In-Reply-To: <54005.84.161.237.217.1154437116.squirrel@mail2web.iu-bremen.de> References: <54005.84.161.237.217.1154437116.squirrel@mail2web.iu-bremen.de> Message-ID: <44CF6668.40905@tzi.de> i.normann at iu-bremen.de schrieb: > Could not find module `PropParser': > use -v to see a list of the files searched for > (imported from ./Haskell/HatParser.hs) > make: *** [hets] Error 1 The file PropParser.hs should be in the directory: programmatic/tools/property/parse2/Parser/ There is only a file PropParser.y that needs to be transformed by "happy" to PropParser.hs. (I hope you have happy, otherwise I could also send you the generated source file.) > happy PropParser.y unused terminals: 1 shift/reduce conflicts: 2 or by: make -f Makefile.orig \ ../programmatic/tools/property/parse2/Parser/PropParser.hs After doing this I just realized that also the file HsLex.hs is missing and that base/parse2/NumberNames.hs needs to be patched. You can fix this by: make -f Makefile.orig ../programatica/tools/base/parse2/Lexer/HsLex.hs make -f Makefile.orig patch > I tried to fix this by adjusting in the Makefile the path for > programmatica; i.e. adjusting PFE_DIRS. > In fact the many files couln't be found in the PFE_DIRS as it is defined > in the Makefile which comes with Hets-src-0.60.tgz - at least it doesn't > fit to > programatica-2006-05-31.tgz. PFE_DIRS should be correct and programatica-2006-05-31.tgz works with Hets-src-0.60.tgz (at least for me). Why is Programatica important for you? (it could be easily omitted by moving away the directory programatica.) I apologize for the above inconveniences. Maybe I should include the generated and patched files in programatica in the future (currently it is only a snapshot of the original cvs sources at that time.) Cheers Christian From maeder at tzi.de Tue Aug 1 17:05:18 2006 From: maeder at tzi.de (Christian Maeder) Date: Tue, 01 Aug 2006 17:05:18 +0200 Subject: [Hets-devel] compile problems with release 0.6 In-Reply-To: <44CF6668.40905@tzi.de> References: <54005.84.161.237.217.1154437116.squirrel@mail2web.iu-bremen.de> <44CF6668.40905@tzi.de> Message-ID: <44CF6DAE.6@tzi.de> >> Could not find module `PropParser': I've rebundled a current snapshot of programatica as programatica-2006-08-01.tgz that includes our manipulations and which happens to work with Hets-src-0.60.tgz http://www.informatik.uni-bremen.de/agbkb/forschung/formal_methods/CoFI/hets/src-distribution/versions What ghc version are you using on what machine? ghc-6.4.1 or ghc-6.4.2 should be fine. Aren't our binary distributions of any help? If you have a mac, the binary may become to large (for linking) if you include programatica. Cheers Christian From i.normann at iu-bremen.de Tue Aug 1 18:00:04 2006 From: i.normann at iu-bremen.de (i.normann at iu-bremen.de) Date: Tue, 1 Aug 2006 18:00:04 +0200 (CEST) Subject: [Hets-devel] compile problems with release 0.6 In-Reply-To: <44CF6668.40905@tzi.de> References: <54005.84.161.237.217.1154437116.squirrel@mail2web.iu-bremen.de> <44CF6668.40905@tzi.de> Message-ID: <58385.84.161.237.217.1154448004.squirrel@mail2web.iu-bremen.de> > PFE_DIRS should be correct and programatica-2006-05-31.tgz works with > Hets-src-0.60.tgz (at least for me). Why is Programatica important for > you? (it could be easily omitted by moving away the directory > programatica.) this is a good point :-) Actually I didn't know at all what this programatica is good for. I just raed the INSTALL file form HetCATS which says: In order to build "hets" from sources ... "HetCATS" and "uni" must reside in the same directory. Additionally a "programatica" directory may be there. (We use slightly modified files, so ask us.) This information is a bit fuzzy. My simple heuristic was: The more packages mentioned in the INSALL or README file I provide the less will be missing during the compilation. So it would be a good idea to give some more precise information about the role that programmatica playz for HetCATS. In particular that 'make' automatically includes programmatica in HetCATS if it exists in HetCATS/.. , but also compiles without programmatica. Probably then I dropped programmatica beforehand. Moreover for the programmatica interested user the comment "We use slightly modified files" is also rather irritating than helpful. Anyway, as HetCATS seems to compile successfully without programmatica I am happy. Thanks for the help. Finally I would propose to add to the INSTALL instructions (at least for release versions) not just the dependencies from other packages but more specific which versions of them. Cheers, Immanuel From i.normann at iu-bremen.de Tue Aug 1 18:03:37 2006 From: i.normann at iu-bremen.de (i.normann at iu-bremen.de) Date: Tue, 1 Aug 2006 18:03:37 +0200 (CEST) Subject: [Hets-devel] compile problems with release 0.6 In-Reply-To: <44CF6DAE.6@tzi.de> References: <54005.84.161.237.217.1154437116.squirrel@mail2web.iu-bremen.de> <44CF6668.40905@tzi.de> <44CF6DAE.6@tzi.de> Message-ID: <58386.84.161.237.217.1154448217.squirrel@mail2web.iu-bremen.de> >>> Could not find module `PropParser': > > I've rebundled a current snapshot of programatica as > programatica-2006-08-01.tgz that includes our manipulations and which > happens to work with Hets-src-0.60.tgz Nice! >http://www.informatik.uni-bremen.de/agbkb/forschung/formal_methods/CoFI/hets/src-distribution/versions > > What ghc version are you using on what machine? ghc-6.4.1 or ghc-6.4.2 > should be fine. I am using ghc-6.4.2 (Linux). Is it normal that it consumes 400MB memory for compilation of HetCATS? I only have 512MB on my machine :-( Immanuel From maeder at tzi.de Tue Aug 1 18:22:50 2006 From: maeder at tzi.de (Christian Maeder) Date: Tue, 01 Aug 2006 18:22:50 +0200 Subject: [Hets-devel] compile problems with release 0.6 In-Reply-To: <58386.84.161.237.217.1154448217.squirrel@mail2web.iu-bremen.de> References: <54005.84.161.237.217.1154437116.squirrel@mail2web.iu-bremen.de> <44CF6668.40905@tzi.de> <44CF6DAE.6@tzi.de> <58386.84.161.237.217.1154448217.squirrel@mail2web.iu-bremen.de> Message-ID: <44CF7FDA.2030409@tzi.de> i.normann at iu-bremen.de schrieb: > Is it normal that it consumes 400MB memory > for compilation of HetCATS? Yes, I'm afraid so. It's faster and less memory hungry if you omit optimization. (the flags "-O -optc-O2" in the Makefile) Aren't you able to execute our binaries hets-2006-08-01.bz2 or hets-0.60.bz2 (after bunzip2)? What does "ldd" say for a binary on your machine? http://www.informatik.uni-bremen.de/agbkb/forschung/formal_methods/CoFI/hets/linux/versions/ Christian From maeder at tzi.de Tue Aug 1 18:33:15 2006 From: maeder at tzi.de (Christian Maeder) Date: Tue, 01 Aug 2006 18:33:15 +0200 Subject: [Hets-devel] compile problems with release 0.6 In-Reply-To: <58385.84.161.237.217.1154448004.squirrel@mail2web.iu-bremen.de> References: <54005.84.161.237.217.1154437116.squirrel@mail2web.iu-bremen.de> <44CF6668.40905@tzi.de> <58385.84.161.237.217.1154448004.squirrel@mail2web.iu-bremen.de> Message-ID: <44CF824B.2060006@tzi.de> i.normann at iu-bremen.de schrieb: > Actually I didn't know at all what this > programatica is good for. Programatica is used to handle simple Haskell sources as in- and output of logic translations (as Logic "Haskell"), but is of limited use, currently. Parts of CASL or HasCASL can be translated to Haskell and (parts of) Haskell can be translated to Isabelle HOL and HOLCF. (More parts of HasCASL and CASL can be better translated to Isabelle HOL directly.) > Finally I would propose to add to the INSTALL instructions (at least for > release versions) not just the dependencies from other packages but more > specific which versions of them. Yes, INSTALL needs to be rephrased anyway. Christian From i.normann at iu-bremen.de Thu Aug 24 19:12:18 2006 From: i.normann at iu-bremen.de (Immanuel Normann) Date: Thu, 24 Aug 2006 19:12:18 +0200 Subject: [Hets-devel] Question about CASL.AS_Basic_CASL.Sort_gen_ax Message-ID: <44EDDDF2.5090408@iu-bremen.de> where can I find more about the intension of the FORMULA constructors "Sort_gen_ax [Constraint] Bool" and "Membership (TERM f) SORT Range" in CAS:/AS_Basic_CASL.hs? Perhaps a pointer to some page in the CASL Reference book or user manual would be helpful. What about the other strange construcors: - Mixfix_formula (TERM f) - Unparsed_formula String Range - ExtFORMULA f are they really in use? Immanuel -------------- next part -------------- A non-text attachment was scrubbed... Name: i.normann.vcf Type: text/x-vcard Size: 184 bytes Desc: not available URL: From maeder at tzi.de Fri Aug 25 10:45:26 2006 From: maeder at tzi.de (Christian Maeder) Date: Fri, 25 Aug 2006 10:45:26 +0200 Subject: [Hets-devel] Question about CASL.AS_Basic_CASL.Sort_gen_ax In-Reply-To: <44EDDDF2.5090408@iu-bremen.de> References: <44EDDDF2.5090408@iu-bremen.de> Message-ID: <44EEB8A6.6020408@tzi.de> Immanuel Normann schrieb: > where can I find more about the intension of the FORMULA constructors > "Sort_gen_ax [Constraint] Bool" and "Membership (TERM f) SORT Range" in > CAS:/AS_Basic_CASL.hs? Perhaps a pointer to some page in the CASL > Reference book or user manual would be helpful. See "sort-generation constraint" III:134-137 The actual set of sorts and functions and the signature morphism can be computed using the function: recover_Sort_gen_ax :: [Constraint] -> ([SORT],[OP_SYMB],[(SORT,SORT)]) The data type "Constraint" was chosen to better represent free and generated data types (with a signature morphism). > What about the other strange construcors: > - Mixfix_formula (TERM f) This is only an intermediate representation between parsing and mixfix resolution (i.e an unresolved mixfix-predicate application) > - Unparsed_formula String Range currently unused, but also an unresolved formula as string that used to be part of an early CASL tool written in SML. > - ExtFORMULA f This is used (as plugin) for several CASL extension (i.e. for modal formulas) > are they really in use? Yes, apart form the Unparsed_-bits also in TERM. In TERMs the constructor "Simple_id" is also only used for Kif2CASL. After mixfix resolution the constructors starting with Mixfix_ are eliminated. HTH Christian