Rockbox mail archive

• Next message: Psyco Dedman: "RockBox graphic attempt"
• Previous message: alankorr: "Re: digrom"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]

Ok, new version to fix a very nasty bug :

  be_long_cluster =
(swabHI(le_short_fat[x+1])<<16)|swabHI(le_short_fat[x+0]);

the trouble is swabHI waits for a int parameter so we get an
undesirable signed extension.

to fix it use explicit unsigned extension :

  be_long_cluster =
(swabHI(le_short_fat[x+1])<<16)|((unsigned
short)swabHI(le_short_fat[x+0]));

I hope it is the only possible cause of problem of "invalid
FAT".

NOTE : le = little endian,
       be = big endian;
       swab = swap byte;
       SI = single integer (32-bit, gcc convention)
       HI = half integer (16-bit, gcc convention)
       QI = quarter integer (8-bit, gcc convention)

That's all, folks !

• application/x-gzip-compressed attachment: digrom2.tgz

• Next message: Psyco Dedman: "RockBox graphic attempt"
• Previous message: alankorr: "Re: digrom"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]