-
Yes, there are tons of those warnings when turning it on globally on MSVC too. Those patches were mainly so that you can turn on the warning on a per-engine basis and not get warnings from the common header files.
Thanks for accepting it!
2009-12-09 19:34:57 UTC in ScummVM
-
I'm now compiling my engine (lastexpress) using MSVC with C4244 ('conversion' conversion from 'type1' to 'type2', possible loss of data) turned on and I've noticed a few warnings in headers included from common code. I'm opening this entry to track those cases.
Patch1: Cursor struct constructor using an int parameter to initialize a byte field: targetScale is defined as byte, but the...
2009-12-09 14:01:07 UTC in ScummVM
-
Thanks for applying those fixes. We will need someone working on SCI to help for the last warning.
Sorry about the broken SCI patch.
- Did I miss something about the text alignment case? Is offset (or alignment) modified in another place and so needs to be set to 0 again?
- For the bAdd case, it was very wrong indeed!
2009-12-08 21:25:06 UTC in ScummVM
-
C4701 only appears 7 times (there is an overview of the warnings in MSVC warnings table.txt). I'd rather disable it per-engine if it's ok (or just silence them since there is so few of them).
And after re-reading my older messages it looks like I'm pushing for the use of the secure string functions, which isn't the case (it was disabled in the original patch). Sorry (and I'll stop talking...
2009-12-08 13:32:27 UTC in ScummVM
-
Those secure string functions have been proposed to ISO/IEC JTC1/SC22/WG14 and they are probably going to end up in the next revision of the standard (see TR 24731-1 and http://www.open-std.org/jtc1/sc22/wg14/www/docs/n1225.pdf). Since it's indeed not in the standard yet (and has other false positives), I've put 4996 back in the globally disabled warnings in the updated patch.
The updated...
2009-12-08 07:49:49 UTC in ScummVM
-
4996: Yes, I'm fine with either disabling it globally or adding _CRT_SECURE_NO_WARNINGS to the global defines and a pair of MSVC-specific defines. In any case, the less MSVC-specific code / global changes to accommodate it, the better.
4702: Oh, I missed the exit() part being for every platform :( There is indeed no crash. But all the existing cleanup statements or other code after error(...
2009-12-07 17:11:24 UTC in ScummVM
-
Adding a dummy assignment operator for every class with const data fixes the warnings, but there are quite a few of them!
I added _CRT_SECURE_NO_WARNINGS to the global defines and that removes quite a lot of warnings.
The _stricmp case is trivially fixed as there is a MSVC specific define for it (behavior with VS2005 needs to checked though). For _strdup, you'll have to rename all the...
2009-12-07 14:58:32 UTC in ScummVM
-
Kirben: 4701 only appears 7 times (most of them trivial changes), so the current code is already "clean" and maybe it's better to have it to highlight cases where there might be issues.
fingolfin: I did another run with the warnings you mention turned on and updated the build log (6581k log file with all 12355 warnings :P).
4512: there are a lot of those (10k+...), mostly coming from the...
2009-12-07 14:29:08 UTC in ScummVM
-
There was some discussion a few days ago on the IRC channel about disabled warnings in Visual Studio. After removing all the default warnings and looking at VS throwing 10k+ warnings, I ended up with a list of warnings that can be disabled globally, a list of warnings that are never thrown and a list of warnings which might be useful.
This patch removes some of the default warnings and adds...
2009-12-07 12:28:08 UTC in ScummVM
-
Project files have language extensions turned on by default. This allows things like enumeration forward declarations to compile fine on MSVC and fail on GCC.
This patch disables language extensions for all engines via the global property file (and turns them back on in the existing special case for the scummvm project file as some seem to used in Windows include files).
After applying the...
2009-12-07 05:36:58 UTC in ScummVM
