gbadev.org forum archive

I noticed that when I use malloc() and similar functions in DS apps, I always have to cast their return types (unless I'm actually using void*) or else it warns me about the conversion. (I use arm-elf-g++ to compile.) Can this warning be disabled? I know it's bad form, but I found that LAME generates several dozen of these warnings. (A lot of compilers don't seem to care about not casting void*.) I've tried fooling with the warning options but I only managed to make it create more warnings for trivial things like assigning an unsigned variable to a signed one.

[edit] Bah, the preview shows my full thread title but it gets cut off when I submit. >_>
_________________
I'm a PSP hacker now, but I still <3 DS.

I'm using the following:

Code:

arm-elf-g++ --no-warnings

melw wrote:

I'm using the following:

Code:

arm-elf-g++ --no-warnings

If your car stops working and you tape the gas meter to full, will that help? No.

I also get a lot of warnings but I can't figure out how to fix it, I don't care about --no-warnings option, I just see what written there anyways
_________________
Important: Please send messages about FWNITRO to the public forum, not privately to me.

Last edited by zzo38computer on Tue May 23, 2006 3:10 pm; edited 1 time in total

melw wrote:

I'm using the following:

Code:

arm-elf-g++ --no-warnings

Your.. a genius! give that man a nobel prize!
_________________
MrShlee.com
Projects
Dev-Scene
MyTechpedia

zzo38computer wrote:

If your car stops working and you tape the gas meter to full, will that help? No.

The man asked how to disable warnings. A simple answer to a simple question, that's all. :)

I want to disable one specific warning, not all of them. Most of the time they're quite useful (many failures are easily resolved by warning messages telling me I forgot some character here or there), but in this case they're just an artifact of someone else's not-quite-correct-but-still-working code. Hell if I'm going to go through and fix them all, there's like 200 of them and I already another fixed 50 or so. >_>
_________________
I'm a PSP hacker now, but I still <3 DS.

I'd also much appreciate a proper warning reference. The gcc docs never seem to tell how to disable the pointless ones. Like, when building Chishm's FAT lib on devkitARM r18, I get lots of "attribute packed has no effect" or something like that, because some u8 variables are set to be packed. Yes, I'm aware that it has no effect, but isn't it safer to specify it than to do it selectively according to the variable type?
_________________
___________
The best optimization is to do nothing at all.
Therefore a fully optimized program doesn't exist.
-Deku

DekuTree64 wrote:

I'd also much appreciate a proper warning reference. The gcc docs never seem to tell how to disable the pointless ones. Like, when building Chishm's FAT lib on devkitARM r18, I get lots of "attribute packed has no effect" or something like that, because some u8 variables are set to be packed. Yes, I'm aware that it has no effect, but isn't it safer to specify it than to do it selectively according to the variable type?

The struct should be packed, not individual elements i.e.

Code:

// Boot Sector - must be packed
typedef struct
{
u8 jmpBoot[3];
u8 OEMName[8];
// BIOS Parameter Block
u16 bytesPerSector;
u8 sectorsPerCluster;
u16 reservedSectors;
u8 numFATs;
u16 rootEntries;
u16 numSectorsSmall;
u8 mediaDesc;
u16 sectorsPerFAT;
u16 sectorsPerTrk;
u16 numHeads;
u32 numHiddenSectors;
u32 numSectors;
union // Different types of extended BIOS Parameter Block for FAT16 and FAT32
{
   struct
   {
      // Ext BIOS Parameter Block for FAT16
      u8 driveNumber;
      u8 reserved1;
      u8 extBootSig;
      u32 volumeID;
      u8 volumeLabel[11];
      u8 fileSysType[8];
      // Bootcode
      u8 bootCode[448];
   } __PACKED fat16;
   struct
   {
      // FAT32 extended block
      u32 sectorsPerFAT32;
      u16 extFlags;
      u16 fsVer;
      u32 rootClus;
      u16 fsInfo;
      u16 bkBootSec;
      u8 reserved[12];
      // Ext BIOS Parameter Block for FAT16
      u8 driveNumber;
      u8 reserved1;
      u8 extBootSig;
      u32 volumeID;
      u8 volumeLabel[11];
      u8 fileSysType[8];
      // Bootcode
      u8 bootCode[420];
   } __PACKED fat32;
} __PACKED extBlock;

u16 bootSig;

} __PACKED BOOT_SEC;

// Directory entry - must be packed
typedef struct
{
u8 name[8];
u8 ext[3];
u8 attrib;
u8 reserved;
u8 cTime_ms;
u16 cTime;
u16 cDate;
u16 aDate;
u16 startClusterHigh;
u16 mTime;
u16 mDate;
u16 startCluster;
u32 fileSize;
} __PACKED DIR_ENT;

// Long file name directory entry - must be packed
typedef struct
{
u8 ordinal; // Position within LFN
u16 char0;
u16 char1;
u16 char2;
u16 char3;
u16 char4;
u8 flag; // Should be equal to ATTRIB_LFN
u8 reserved1; // Always 0x00
u8 checkSum; // Checksum of short file name (alias)
u16 char5;
u16 char6;
u16 char7;
u16 char8;
u16 char9;
u16 char10;
u16 reserved2; // Always 0x0000
u16 char11;
u16 char12;
} __PACKED DIR_ENT_LFN;

_________________
devkitPro - professional toolchains at amateur prices
devkitPro IRC support
Personal Blog

wintermute wrote:

The struct should be packed, not individual elements i.e.

You'd think so, and so did I. However, I found it didn't work that way. I decided to stop with the compiler dependent packing, and extract the values using an enum of offsets.
_________________
http://chishm.drunkencoders.com
http://dldi.drunkencoders.com

gbadev.org forum archive

DS Misc > Invalid conversion from void* to foo*; disable this warning?

#84548 - HyperHacker - Tue May 23, 2006 7:40 am

#84579 - melw - Tue May 23, 2006 2:34 pm

#84585 - zzo38computer - Tue May 23, 2006 3:05 pm

#84587 - Mrshlee - Tue May 23, 2006 3:09 pm

#84614 - melw - Tue May 23, 2006 6:52 pm

#84623 - HyperHacker - Tue May 23, 2006 7:36 pm

#84629 - DekuTree64 - Tue May 23, 2006 8:07 pm

#84647 - wintermute - Tue May 23, 2006 10:34 pm

#84716 - chishm - Wed May 24, 2006 11:14 am