Maybe buffering doesn't work for ReadByte/WriteByte functions? ... I'm doing some file conversion, and relying on what the helpfile said, I figured why do my own buffering? buffering "on top" of routines that are already buffered would just be extra overhead - so I'm reading/writing one byte at a time, like so:
Code: Select all
While ~Eof(#FileIn)
UseFile(#FileIn)
i = ReadByte()
UseFile(#FileOut)
WriteByte(i)
Wend
of course my routines do way more than that, but to see whether the 20 seconds of overhead were coming from the actual work or from reading/writing, I removed all the code ... the actual work doesn't make up for even 1% of the CPU usage! 8O ... in other words, if not for the incredibly slow read/write operations, my program would finish in about 0.2 seconds instead of 20!

so now I do have to do my own buffering.
which is incredibly tedious, because I was counting on the library to do what the helpfile promises, and I hate having to reinvent the wheel...
*sigh*