code:

---

$iBytesRead=ReadArray($fdFileDescriptor,$aiDataRead,$iBytesToRead)

---

code:

---

$iBytesWritten=WriteArray($fdFileDescriptor,$aiWriteData,$iBytesToWrite

---

code:

---

$aiDataRead=ReadArray($fdFileDescriptor,$iBytesToRead)

---

@ERROR can be checked for end-of-file and error conditions. The number of bytes read is the size of the array.
I prefer the first method as there is a small problem with this way of doing reads in that it is tricky to identify when some data was read before an error occurred.

The other change required is that the Open() function will probably need a flag to ensure that the file is opened in binary or "raw" mode. This is because it will probably be required by the underlying file open API.

Mixing binary and non-binary reads and writes may be possible, but the results will be hard to predict and should be avoided.

Performing binary reads and writes on a file not opened in binary mode (and vice versa) will also have unexpected results.

Some more thoughts (shouldn't think and type at the same time )

If binary file I/O is implemented it should be able to support random file I/O, i.e.:

• A file can be opened for both read and write.
• Seek() function to move the file cursor to an arbitrary location in the file (1=start, -1=end)
• Truncate() function to reduce the size of the file (when you've deleted data from the end) or to initialize a file which has data in it already that you want to discard

Quote:

---

Hi Richard, do you have a link to that post?

---