mORMot Open Source

Amir

Member

From: UAE

Registered: 2010-08-17

Posts: 36

Length of unicode string by pointer in new PosExStringPas...

Hi dear Arnaud,

As PChar is compiled with {$POINTERMATH ON}, direct pointer arithmatic operations should be applied with the multiplication of the pointer's base type-size with the value specified by operation.

Thus in for eg:

function PosExStringPas(pSub, p: PChar; Offset: PtrUInt): PtrInt;

We calculate the length of the "p" and "pSub" in Delphi version as

PInteger(p - 4)^; //and so for "pSub"

This actually is computed as "PByet(p) - 8", Where Char is two-byte sized (in Unicode Versions of Delphi).

Seems it should be edited for both "p" and "pSub" as for eg:

PInteger(PByte(p) - 4)^;

For always correct access to string's length field.

Please let me know if I am wrong...

Thanks.

Last edited by Amir (2019-12-04 11:32:04)

---

Amir

ab

Administrator

From: France

Registered: 2010-06-21

Posts: 14,206

Website

Re: Length of unicode string by pointer in new PosExStringPas...

Now, you are perfectly right: this was a bug of this new function.

It should be properly fixed by https://synopse.info/fossil/info/57e8496c5d
I also added some associated regression tests - as expected.