Bug 8986 – The PVS-Studio analyzer has something for you

Perhaps there are other errors. I have studied the report unthoughtful. The report can be viewed by downloading the PVS-Studio trial version of the analyzer. I think trial version will be enough. --------------------- V522 Dereferencing of the null pointer 'v' might take place. interpret.c 1711 Expression *getVarExp(Loc loc, InterState *istate, Declaration *d, CtfeGoal goal) { .... VarDeclaration *v = d->isVarDeclaration(); if (v) { .... } else if (s) { if (s->dsym->toInitializer() == s->sym) .... else error(loc, "cannot interpret symbol %s at compile time", v->toChars()); } .... } --------------------- V519 The 'c2' variable is assigned values twice successively. Perhaps this is a mistake. Check lines: 1860, 1861. cod2.c 1861 code *cdcond(elem *e,regm_t *pretregs) { code *cc,*c,*c1,*cnop1,*c2,*cnop2; .... c2 = codelem(e22,&retregs,FALSE); c2 = cat(c1, fixresult(e22,retregs,pretregs)); .... } Most likely this is what should be written here: c2 = cat(c2, ....); --------------------- V612 An unconditional 'return' within a loop. cgxmm.c 825 code *cdvector(elem *e, regm_t *pretregs) { .... code *c = CNIL; for (int i = 0; i < n; i++) { c = cat(c, codelem(params[i], pretregs, FALSE)); *pretregs = 0; // in case they got set return c; } .... }

(In reply to comment #0) > Perhaps there are other errors. I have studied the report unthoughtful. The > report can be viewed by downloading the PVS-Studio trial version of the > analyzer. I think trial version will be enough. Thank you for your valuable work Andrey, we read often your articles linked and discussed on Reddit. PVS-Studio seems a interesting tool. D language itself (and not its compilers) is designed to help D programmers avoid or remove many of the bugs found by PVS-Studio, see as example: http://d.puremagic.com/issues/show_bug.cgi?id=5409 So in a sense D makes tools like PVS-Studio less necessary. > V519 The 'c2' variable is assigned values twice successively. Perhaps this is a > mistake. Check lines: 1860, 1861. cod2.c 1861 > > code *cdcond(elem *e,regm_t *pretregs) > { > code *cc,*c,*c1,*cnop1,*c2,*cnop2; > .... > c2 = codelem(e22,&retregs,FALSE); > c2 = cat(c1, fixresult(e22,retregs,pretregs)); > .... > } > > Most likely this is what should be written here: c2 = cat(c2, ....); That's a bug in the C++ code. Maybe it's possible to add a warning inside the D compiler to guard against similar bugs in D code.

Yes, I know what you are studying common problems in C + + to improve D. Nice to know that my examples are useful.

Commit pushed to master at https://github.com/D-Programming-Language/dmd https://github.com/D-Programming-Language/dmd/commit/b72f726f84ee58d14b26cc7fdb49a6665932ec80 fix Issue 8986 - The PVS-Studio analyzer has something for you

Commit pushed to dmd-1.x at https://github.com/D-Programming-Language/dmd https://github.com/D-Programming-Language/dmd/commit/8dcabd3debb7b891fa5471da98034264caf90545 fix Issue 8986 - The PVS-Studio analyzer has something for you