From martin@nagmx1.nag.co.uk Mon Sep 11 16:36:02 2000 Received: from frisco.nag.co.uk (frisco.nag.co.uk [192.156.217.22]) by nagmx1.nag.co.uk (8.9.3/8.9.3) with SMTP id QAA03882 for ; Mon, 11 Sep 2000 16:36:01 +0100 (BST) From: Martin Dunstan Date: Mon, 11 Sep 2000 16:29:32 +0100 Message-Id: <200009111529.QAA23687@frisco.nag.co.uk> Received: by frisco.nag.co.uk (SMI-8.6) id QAA23687; Mon, 11 Sep 2000 16:29:32 +0100 To: ax-bugs@nag.co.uk@nagmx1.nag.co.uk Subject: fixbug By: mnd Fixed: bug1140.as --* From youssef@mailer.scri.fsu.edu Sun May 10 03:52:56 1998 --* Received: from nagmx1.nag.co.uk by red.nag.co.uk via SMTP (920330.SGI/920502.SGI) --* for /home/red5/axiom/support/recvbug id AA09831; Sun, 10 May 98 03:52:56 +0100 --* Received: from mailer.scri.fsu.edu (mailer.scri.fsu.edu [144.174.112.142]) --* by nagmx1.nag.co.uk (8.8.4/8.8.4) with ESMTP --* id DAA00986 for ; Sun, 10 May 1998 03:57:36 +0100 (BST) --* Received: from dirac.scri.fsu.edu (margit.scri.fsu.edu [144.174.128.45]) by mailer.scri.fsu.edu (8.8.7/8.7.5) with SMTP id WAA08866; Sat, 9 May 1998 22:54:51 -0400 (EDT) --* From: Saul Youssef --* Received: by dirac.scri.fsu.edu (5.67b) id AA26234; Sat, 9 May 1998 22:54:51 -0400 --* Date: Sat, 9 May 1998 22:54:51 -0400 --* Message-Id: <199805100254.AA26234@dirac.scri.fsu.edu> --* To: adk@mailer.scri.fsu.edu, ax-bugs@nag.co.uk, edwards@mailer.scri.fsu.edu, --* youssef@mailer.scri.fsu.edu --* Subject: [3] axiomxl has a segentation fault --@ Bug Number: bug1140.as --@ Fixed by: mnd --@ Tested by: none --@ Summary: seems to be working okay now -- Command line: none -- Version: 1.1.10c -- Original bug file name: bug.as --+ -- --+ -- Peter, --+ -- --+ -- I've managed to cause axiomxl to have a segmentation fault --+ -- for some reason and I can't think of any reason why this should --+ -- happen. If you would like to reproduce this, make the files --+ -- --+ -- ntuples.as and --+ -- data.as --+ -- --+ -- as indicated below. Then do --+ -- --+ -- % axiomxl -Fao ntuples.as --+ -- % axiomxl -Fao data.as --+ -- --+ -- you should get a segmentation fault on the second compilation --+ -- on an AIX system with aldor 1.1.10c --+ -- --+ --+ ===== put this in ntuples.as ============================================= --+ -- --+ -- SW n-Tuples --+ -- --+ --+ #include "axllib" --+ #pile --+ --+ S ==> SingleInteger --+ --+ define FloatCategory: Category == OrderedRing with --+ step: S -> (%,%) -> Generator % --+ exp: % -> % --+ float: Literal -> % --+ --+ SDoubleFloat: FloatCategory == DoubleFloat add --+ Rep ==> DoubleFloat --+ import from Rep, DoubleFloatElementaryFunctions --+ --+ exp(x:%):% == per exp rep x --+ float(l:Literal):% == per float l --+ --+ SFloat: FloatCategory == Float add --+ Rep ==> Float --+ --+ NTuple(F:FloatCategory,n:S): OrderedRing with --+ ntuple: Array F -> % -- create a new ntuple --+ ntuple: Tuple F -> % --+ val: % -> Array F -- return n-tuple as an array --+ apply: (%,S) -> F -- return n-tuple component --+ set!: (%,S,F) -> % --+ dim: % -> S -- dimension of the n-tuple --+ mag2: % -> F -- sum of squares of components --+ sum: % -> F -- sum of components --+ *: (F,%) -> % -- scalar muliplication --+ == add --+ Rep ==> Array F --+ import from S, Rep --+ -- --+ -- NTuple specific --+ -- --+ ntuple(a:Array F):% == --+ #a=n => per a --+ error "wrong array length for ntuple" --+ ntuple(t:Tuple F):% == --+ length t=n => per [t] --+ error "wrong number of arguments for ntuple" --+ val(t:%):Array F == rep t --+ dim(t:%):S == #rep t --+ apply(t:%,i:S):F == (rep t).i --+ set!(t:%,i:S,x:F):% == {z:=rep t; z.i := x; per z} --+ mag2(t:%):F == --+ x:F := 1 --+ for y in rep t repeat x := x * y --+ x --+ sum(t:%):F == --+ x:F := 0 --+ for y in rep t repeat x := x + y --+ x --+ (s:F)*(t:%):% == ntuple [s*rep(t).i for i:S in 1..dim t] --+ -- --+ -- OrderedRing signatures --+ -- --+ 0:% == ntuple new (n,0) --+ 1:% == ntuple new (n,1) --+ (t1:%)+(t2:%):% == ntuple ([rep(t1).i+rep(t2).i for i:S in 1..n]) --+ (t1:%)*(t2:%):% == ntuple ([rep(t1).i*rep(t2).i for i:S in 1..n]) --+ (t1:%)-(t2:%):% == ntuple ([rep(t1).i-rep(t2).i for i:S in 1..n]) --+ -(t:%):% == ntuple ([-rep(t).i for i:S in 1..n]) --+ (t:%)^(p:Integer):% == --+ x:% := 1 --+ for i:S in 1..n repeat --+ x := x * t --+ x --+ (t1:%)>(t2:%):Boolean == --+ q := true --+ for i:S in 1..n repeat --+ q := q and rep(t1).i>rep(t2).i --+ q --+ coerce(m:Integer):% == --+ t:% := 0 --+ for i:Integer in 1..m repeat --+ t := t + 1 --+ t --+ coerce(m:S):% == --+ t:% := 0 --+ for i:S in 1..m repeat --+ t := t + 1 --+ t --+ -- --+ -- BasicType etc. --+ -- --+ (t1:%)=(t2:%):Boolean == --+ q := true --+ for i:S in 1..n repeat --+ q := q and rep(t1).i=rep(t2).i --+ q --+ <<(t:TextWriter,nt:%):TextWriter == t << rep nt --+ --+ NTupleProjection(F:FloatCategory,n1:S,n2:S): with --+ project: (NTuple(F,n1),Array S) -> NTuple(F,n2) --+ == add --+ project(t1:NTuple(F,n1),ind:Array S):NTuple(F,n2) == --+ #ind = n2 => --+ a:Array F := [val(t1).(ind.k) for k in 1..n2] --+ ntuple a --+ error "wrong number of indices for projection in project" --+ --+ #if TEST --+ --+ import from S, SFloat, NTuple(SFloat,3), NTupleProjection(SFloat,6,3) --+ ones:Array SFloat := [1.0,1.0,1.0] --+ --+ low: NTuple(SFloat,3) := ntuple ones --+ high: NTuple(SFloat,3) := ntuple (5.0,5.0,5.0) --+ long: NTuple(SFloat,6) := ntuple (10.0,20.0,30.0,40.0,50.0,60.0) --+ ind: Array S := [2,4,6] --+ --+ print << low << newline --+ print << high << newline --+ print << low+high << newline --+ print << low*high << newline --+ print << high*(high+low) << newline --+ print << long << newline --+ print << project(long,ind) << newline --+ print << mag2 high << newline --+ print << 0.5*high << newline --+ print << exp(1.0) << newline --+ --+ #endif --+ --+ =================== end of ntuples.as ==================================== --+ --+ ================= put the following in data.as =========================== --+ -- --+ #include "axllib" --+ #include "ntuples" --+ #pile --+ --+ S ==> SingleInteger --+ import from S --+ --+ Lepton == 'electron, muon' --+ --+ DataJet(F:FloatCategory): BasicType with --+ datajet: (F,F,F)->(F,F,F) -> % --+ datajet: (NTuple(F,3),NTuple(F,3)) -> % --+ distance2: (%,NTuple(F,3)) -> F --+ apply: (%,'px') -> F --+ apply: (%,'py') -> F --+ apply: (%,'pz') -> F --+ apply: (%,'sigx') -> F --+ apply: (%,'sigy') -> F --+ apply: (%,'sigz') -> F --+ apply: (%,S) -> (F,F) --+ tag?: % -> Boolean --+ mutag?: % -> Boolean --+ etag?: % -> Boolean --+ == add --+ Rep ==> Record(p:NTuple(F,3),sigp:NTuple(F,3),tag:Boolean,leptype:Lepton) --+ import from Lepton, Rep --+ --+ datajet(px:F,py:F,pz:F)(sx:F,sy:F,sz:F):% == --+ sx>=0.0 and sy>=0.0 and sz>= 0.0 => --+ per [ntuple(px,py,pz), ntuple(sx,sy,sz),false,electron] --+ error "negative sigmas in datajet" --+ datajet(t1:NTuple(F,3),t2:NTuple(F,3)):% == per [t1,t2,false,electron] --+ --+ =(a:%,b:%):Boolean == rep a = rep b --+ sample: % == per [sample, sample, sample, sample ] --+ apply(j:%,x:'px'):F == rep(j).p.1 --+ apply(j:%,x:'py'):F == rep(j).p.2 --+ apply(j:%,x:'pz'):F == rep(j).p.3 --+ apply(j:%,x:'sigx'):F == rep(j).sigp.1 --+ apply(j:%,x:'sigy'):F == rep(j).sigp.2 --+ apply(j:%,x:'sigz'):F == rep(j).sigp.3 --+ apply(j:%,i:S):(F,F) == --+ 1<=i and i<=3 => (rep(j).p.i, rep(j).sigp.i) --+ error "DataJet index out of range" --+ tag?(j:%):Boolean == rep(j).tag --+ mutag?(j:%):Boolean == rep(j).tag and rep(j).leptype=muon --+ etag?(j:%):Boolean == rep(j).tag and rep(j).leptype=electron --+ <<(t:TextWriter,j:%):TextWriter == --+ for i in 1..3 repeat --+ t << "Jet:" --+ pr j.i where --+ pr(x:F,y:F):() == t <<" "<< x <<" +- " << y << newline --+ t --+ distance2(j:%,v:NTuple(F,3)):F == 1.0 --+ -- distance2(j:%,v:NTuple(F,3)):F == --+ -- temp:NTuple(F,3) := ((rep(j).p - v) * rep(j).sigp) --+ -- temp.1*temp.1 + temp.2*temp.2 + temp.3*temp.3 --+ --+ --+ --+ --+ --+ -- -- Peter, -- -- I've managed to cause axiomxl to have a segmentation fault -- for some reason and I can't think of any reason why this should -- happen. If you would like to reproduce this, make the files -- -- ntuples.as and -- data.as -- -- as indicated below. Then do -- -- % axiomxl -Fao ntuples.as -- % axiomxl -Fao data.as -- -- you should get a segmentation fault on the second compilation -- on an AIX system with aldor 1.1.10c -- ===== put this in ntuples.as ============================================= -- -- SW n-Tuples -- #include "axllib" #pile S ==> SingleInteger define FloatCategory: Category == OrderedRing with step: S -> (%,%) -> Generator % exp: % -> % float: Literal -> % SDoubleFloat: FloatCategory == DoubleFloat add Rep ==> DoubleFloat import from Rep, DoubleFloatElementaryFunctions exp(x:%):% == per exp rep x float(l:Literal):% == per float l SFloat: FloatCategory == Float add Rep ==> Float NTuple(F:FloatCategory,n:S): OrderedRing with ntuple: Array F -> % -- create a new ntuple ntuple: Tuple F -> % val: % -> Array F -- return n-tuple as an array apply: (%,S) -> F -- return n-tuple component set!: (%,S,F) -> % dim: % -> S -- dimension of the n-tuple mag2: % -> F -- sum of squares of components sum: % -> F -- sum of components *: (F,%) -> % -- scalar muliplication == add Rep ==> Array F import from S, Rep -- -- NTuple specific -- ntuple(a:Array F):% == #a=n => per a error "wrong array length for ntuple" ntuple(t:Tuple F):% == length t=n => per [t] error "wrong number of arguments for ntuple" val(t:%):Array F == rep t dim(t:%):S == #rep t apply(t:%,i:S):F == (rep t).i set!(t:%,i:S,x:F):% == {z:=rep t; z.i := x; per z} mag2(t:%):F == x:F := 1 for y in rep t repeat x := x * y x sum(t:%):F == x:F := 0 for y in rep t repeat x := x + y x (s:F)*(t:%):% == ntuple [s*rep(t).i for i:S in 1..dim t] -- -- OrderedRing signatures -- 0:% == ntuple new (n,0) 1:% == ntuple new (n,1) (t1:%)+(t2:%):% == ntuple ([rep(t1).i+rep(t2).i for i:S in 1..n]) (t1:%)*(t2:%):% == ntuple ([rep(t1).i*rep(t2).i for i:S in 1..n]) (t1:%)-(t2:%):% == ntuple ([rep(t1).i-rep(t2).i for i:S in 1..n]) -(t:%):% == ntuple ([-rep(t).i for i:S in 1..n]) (t:%)^(p:Integer):% == x:% := 1 for i:S in 1..n repeat x := x * t x (t1:%)>(t2:%):Boolean == q := true for i:S in 1..n repeat q := q and rep(t1).i>rep(t2).i q coerce(m:Integer):% == t:% := 0 for i:Integer in 1..m repeat t := t + 1 t coerce(m:S):% == t:% := 0 for i:S in 1..m repeat t := t + 1 t -- -- BasicType etc. -- (t1:%)=(t2:%):Boolean == q := true for i:S in 1..n repeat q := q and rep(t1).i=rep(t2).i q <<(t:TextWriter,nt:%):TextWriter == t << rep nt NTupleProjection(F:FloatCategory,n1:S,n2:S): with project: (NTuple(F,n1),Array S) -> NTuple(F,n2) == add project(t1:NTuple(F,n1),ind:Array S):NTuple(F,n2) == #ind = n2 => a:Array F := [val(t1).(ind.k) for k in 1..n2] ntuple a error "wrong number of indices for projection in project" #if TEST import from S, SFloat, NTuple(SFloat,3), NTupleProjection(SFloat,6,3) ones:Array SFloat := [1.0,1.0,1.0] low: NTuple(SFloat,3) := ntuple ones high: NTuple(SFloat,3) := ntuple (5.0,5.0,5.0) long: NTuple(SFloat,6) := ntuple (10.0,20.0,30.0,40.0,50.0,60.0) ind: Array S := [2,4,6] print << low << newline print << high << newline print << low+high << newline print << low*high << newline print << high*(high+low) << newline print << long << newline print << project(long,ind) << newline print << mag2 high << newline print << 0.5*high << newline print << exp(1.0) << newline #endif =================== end of ntuples.as ==================================== ================= put the following in data.as =========================== -- #include "axllib" #include "ntuples" #pile S ==> SingleInteger import from S Lepton == 'electron, muon' DataJet(F:FloatCategory): BasicType with datajet: (F,F,F)->(F,F,F) -> % datajet: (NTuple(F,3),NTuple(F,3)) -> % distance2: (%,NTuple(F,3)) -> F apply: (%,'px') -> F apply: (%,'py') -> F apply: (%,'pz') -> F apply: (%,'sigx') -> F apply: (%,'sigy') -> F apply: (%,'sigz') -> F apply: (%,S) -> (F,F) tag?: % -> Boolean mutag?: % -> Boolean etag?: % -> Boolean == add Rep ==> Record(p:NTuple(F,3),sigp:NTuple(F,3),tag:Boolean,leptype:Lepton) import from Lepton, Rep datajet(px:F,py:F,pz:F)(sx:F,sy:F,sz:F):% == sx>=0.0 and sy>=0.0 and sz>= 0.0 => per [ntuple(px,py,pz), ntuple(sx,sy,sz),false,electron] error "negative sigmas in datajet" datajet(t1:NTuple(F,3),t2:NTuple(F,3)):% == per [t1,t2,false,electron] =(a:%,b:%):Boolean == rep a = rep b sample: % == per [sample, sample, sample, sample ] apply(j:%,x:'px'):F == rep(j).p.1 apply(j:%,x:'py'):F == rep(j).p.2 apply(j:%,x:'pz'):F == rep(j).p.3 apply(j:%,x:'sigx'):F == rep(j).sigp.1 apply(j:%,x:'sigy'):F == rep(j).sigp.2 apply(j:%,x:'sigz'):F == rep(j).sigp.3 apply(j:%,i:S):(F,F) == 1<=i and i<=3 => (rep(j).p.i, rep(j).sigp.i) error "DataJet index out of range" tag?(j:%):Boolean == rep(j).tag mutag?(j:%):Boolean == rep(j).tag and rep(j).leptype=muon etag?(j:%):Boolean == rep(j).tag and rep(j).leptype=electron <<(t:TextWriter,j:%):TextWriter == for i in 1..3 repeat t << "Jet:" pr j.i where pr(x:F,y:F):() == t <<" "<< x <<" +- " << y << newline t distance2(j:%,v:NTuple(F,3)):F == 1.0 -- distance2(j:%,v:NTuple(F,3)):F == -- temp:NTuple(F,3) := ((rep(j).p - v) * rep(j).sigp) -- temp.1*temp.1 + temp.2*temp.2 + temp.3*temp.3