httpd.py 文件源码

def chunkify(self, gen):
        bs = self.bs
        chunks = self.chunks()
        chunk = chunks.next()

        pos = 0
        for src in gen:
            srctype = type(src)
            src = memoryview(src) if srctype in (str, buffer, bytearray, memoryview) else memoryview(str(src))
            slen = len(src)
            try:
                # fast append
                chunk.payload[pos:pos + slen] = src
                pos += slen
            except ValueError:
                # oops - too big - slice & dice
                soff = bs - pos
                # pad buffer out to end using first n bytes from src
                chunk.payload[pos:bs] = src[0:soff]
                yield chunk
                chunk = chunks.next()
                pos = 0

                # then carve off full blocks directly from src
                while soff + bs <= slen:
                    chunk.payload[0:bs] = src[soff:soff+bs]
                    yield chunk
                    chunk = chunks.next()
                    soff += bs

                # and stash the remainder
                pos = slen - soff
                chunk.payload[0:pos] = src[soff:soff+pos]

        if pos:
            yield chunk(pos)


# because every multiprocessing.Process().start() very helpfully
# does a waitpid(WNOHANG) across all known children, and I want
# to use os.wait() to catch exiting children