On Wed, 1 Jul 1998, stefan kluth wrote:
>
> Hi,
>
> after installing 2.00.09 for Linux/egcs I find that rootcint generates
> code which does not compile. I created two dictionary code files, one with
> 2.00.08 and one with 2.00.09 and diff'ed it. The differences coincide with
> the lines where the compiler flags errors. Here is an example:
>
> dictionary code file made with rootcint 2.00.08 (compiles and works):
>
> ...
>
> static int G__AllJets_AllJets_2_0(G__value *result7,char *funcname,struct
> G__param *libp,int hash) {
> AllJets *p=NULL;
> switch(libp->paran) {
> case 3:
> p = new AllJets(
> *(HepConstAList<HepLorentzVector>*)libp->para[0].ref,*(AllJets::Algorithm*)
> libp->para[1].ref,*(int*)libp->para[2].ref);
>
> ...
>
>
> dictionary code file made with rootcint 2.00.09:
>
> ...
>
> static int G__AllJets_AllJets_2_0(G__value *result7,char *funcname,struct
> G__param *libp,int hash) {
> AllJets *p=NULL;
> switch(libp->paran) {
> case 3:
> p = new AllJets(
> *(HepConstAList<HepLorentzVector>*)libp->para[0].ref,
> libp->para[1].ref?*(AllJets::Algorithm*)libp->para[1].ref:
G__Mint(libp->para[1])
> ,libp->para[2].ref?*(int*)libp->para[2].ref:G__Mint(libp->para[2]));
>
> ...
>
>
> The egcs g++ compiler says:
> AllJetsCint.cc: In function `int G__AllJets_AllJets_2_0(struct G__value *,
> char *, struct G__param *, int)':
> AllJetsCint.cc:371: no matching function for call to `AllJets::AllJets
> (HepConstAList<HepLorentzVector> &, int, int &)'
>
>
> The difference between rootcint for 2.00.08 and 2.00.09 seems to be that
> rootcint inserts logical expressions " ... ? ... : ... " instead of just
> casting and dereferencing pointers. This leads to the wrong argument types
> for the call of the constructor.
And I forgot to say, that AllJets::Algorithm is a public enumeration of
class AllJets. The compiler only complains when the argument is expected
to be of type "AllJets::Algorithm", because in the " ... ? ... : ... "
construct the 2nd possibility "G__Mint(libp->para[1])" appearently has the
wrong type, i.e. int, and in general the conversion int->enumeration is
not allowed.
cheers, Stefan
---Stefan Kluth---------------Lynen Fellow----------------|\--|\-------
- LBNL, MS 50A 2160 - phone: +1 510 495 2376 - |/ |/ -
- 1 Cyclotron Rd. - fax: +1 510 495 2957 - |\/\|\/\|' -
---Berkeley, CA94720, USA-----e-mail: SKluth@lbl.gov------|/\/|/\/|----
This archive was generated by hypermail 2b29 : Tue Jan 04 2000 - 00:34:34 MET