Talk:Get localised font name

Hi,

Probably there is a bug in the code. The code first allocates a block of ${NSIS_MAX_STRLEN} bytes, but then copies it to $R0 as a WCHAR string of ${NSIS_MAX_STRLEN} characters, which are actually ${NSIS_MAX_STRLEN} * 2 bytes.

With the Unicode NSIS I have the code crashing from time to time on the line:

 System::Call *$R2(&w${NSIS_MAX_STRLEN}.R0)

I fixed the code by replacing the following lines:

 System::Call *(i${NSIS_MAX_STRLEN})i.R1
 System::Alloc ${NSIS_MAX_STRLEN}
 Pop $R2

with the new lines:

 IntOp $R1 ${NSIS_MAX_STRLEN} * 2 ; number of bytes in WCHAR string of NSIS_MAX_STRLEN length
 System::Call *(i$R1)i.R1
 System::StrAlloc ${NSIS_MAX_STRLEN}
 Pop $R2

I am not a NSIS expert, so the code may be not the most efficient. If my assumption is correct, please, update the code in case everyone else needs it. Thanks.

Alex Che Korotkin

---

System::Call *(&w${NSIS_MAX_STRLEN})i.R2 is probably a better alternative (You can verify this with System::Call kernel32::GlobalSize(iR2)i.r1 ).

NSIS_CHAR_SIZE is defined as 2 in the unicode build (There might be a NSIS_MAX_STRLEN as bytes define, I don't remember)

--Anders 17:40, 5 May 2011 (UTC)

---

Anders, you're right. My code won't work for non-Unicode version of NSIS, because System::StrAlloc in this case will allocate non-Unicode string, while the code explicitly uses the Unicode version of system function: GetFontResourceInfoW. I changed the lines to

 System::Call *(&w${NSIS_MAX_STRLEN})i.R2
 System::Call kernel32::GlobalSize(iR2)i.R1
 System::Call *(i$R1)i.R1

and it works OK in the Unicode NSIS. Thanks.

Alex Che Korotkin