Hi!
On Sun, Feb 14, 2010 at 02:15, Andre Poenitz
<andre.poen...@mathematik.tu-chemnitz.de> wrote:
>> 2. src/support/debug.h: DEBUG is defined as DEBUG = (1 << 31)
>> what is out of range for enum (32 bit) and I guess should be DEBUG
>> = (1 << 30) as there is no 30, yet.
>
> 1 << 31 is within the 32 bit range.
mmh, yeah, you are right. Why does the compiler complain then?
Error: lyx::Debug::Type is not within the range of a long or unsigned long.
I will ask on the Sun side...
>> 3. a forgotten static_cast in src/support/FileName.cpp
> - munmap(mm, info.st_size);
> + munmap(static_cast<char *>(mm), info.st_size);
>
> This seems to be Solaris-specific. So a configure check might be needed
> (unless we decide an #ifdef is fine...)
seems like it will be defined as extern int munmap ( caddr_t , size_t ) ;
when using C++ - at least here. Man page says int munmap(void *addr,
size_t len); but that maybe true for C, not C++.
Configure check may be ok, but configure tends to check with $CC
instead of $CXX, as I have seen...
>
> docstring indent;
> if (ind < 0) {
> - indent.insert(0, -ind, ' ');
> + indent.insert(static_cast<unsigned>(0), -ind, ' ');
>
> This looks strange. Why is that needed? Is the compiler confused by overloads?
> If so, '0u' might look a bit better than 'static_cast<unsigned>(0)'
"lstrings.cpp", line 953: Error: Overloading ambiguity between
"std::wstring::insert(wchar_t*, unsigned, wchar_t)" and
"std::wstring::insert(unsigned, unsigned, wchar_t)".
"lstrings.cpp", line 956: Error: Overloading ambiguity between
"std::wstring::insert(wchar_t*, unsigned, wchar_t)" and
"std::wstring::insert(unsigned, unsigned, wchar_t)".
But yes - 0u is the better way ;)
Bye...Frank
