On Fri, 17 Jan 2014 09:12:56 -0500, Dmitry Olshansky <dmitry.olsh@gmail.com> wrote:

> 17-Jan-2014 18:03, Steven Schveighoffer пишет:
>> On Fri, 17 Jan 2014 05:01:35 -0500, Dmitry Olshansky
>> <dmitry.olsh@gmail.com> wrote:
>>
>>> BTW I haven't thought of writing into the buffer, but it works exactly
>>> the same. It could be even read/write -"discarded" data is written to
>>> underlying stream, freshly loaded is read from stream. Now in case of
>>> input stream only writes are nop, for output-only reads are nops.
>>>
>>> With pinning it makes for cool multi-pass algorithms that actually
>>> output stuff into the file.
>>
>> In my experience, the code/process for a write buffer is significantly
>> different than the code for a read buffer. A read/write buffer is very
>> difficult to make, because you either need 2 file pointers, or need to
>> constantly seek the single file pointer to overwrite the existing data.
>>
>
> Agreed, read/write buffer is a bad idea. As for write-only buffer implemented in the same style as read-only, well I need to try it first.
>

Keep in mind that a read/write buffered stream *should* exist. It's just that you need to switch how you use the buffer.

I think in my code, what I do is if someone wants to read, and is currently writing (on a read/write stream), I flush the buffer, then switch it to be in read mode, and then seek to the appropriate place, to fill the buffer.

When switching from read to write, it's much easier, since the buffer starts out empty.

So I think a buffer type that can be used for read *OR* write is good, just not simultaneously.

However, simultaneous read/write is probably something that could be useful for a niche use case (e.g. some transformative process), as long as you don't mind having 2 file descriptors open :)

-Steve

17-Jan-2014 18:52, Steven Schveighoffer пишет:
> On Fri, 17 Jan 2014 09:12:56 -0500, Dmitry Olshansky
> <dmitry.olsh@gmail.com> wrote:
>
>> 17-Jan-2014 18:03, Steven Schveighoffer пишет:
>>> On Fri, 17 Jan 2014 05:01:35 -0500, Dmitry Olshansky
>>> <dmitry.olsh@gmail.com> wrote:
>>>
>>>> BTW I haven't thought of writing into the buffer, but it works exactly
>>>> the same. It could be even read/write -"discarded" data is written to
>>>> underlying stream, freshly loaded is read from stream. Now in case of
>>>> input stream only writes are nop, for output-only reads are nops.
>>>>
>>>> With pinning it makes for cool multi-pass algorithms that actually
>>>> output stuff into the file.
>>>
>>> In my experience, the code/process for a write buffer is significantly
>>> different than the code for a read buffer. A read/write buffer is very
>>> difficult to make, because you either need 2 file pointers, or need to
>>> constantly seek the single file pointer to overwrite the existing data.
>>>
>>
>> Agreed, read/write buffer is a bad idea. As for write-only buffer
>> implemented in the same style as read-only, well I need to try it first.
>>
>
> Keep in mind that a read/write buffered stream *should* exist. It's just
> that you need to switch how you use the buffer.

I'm doubting that actually. Just use the freaking MM-file directly you'd better off efficiency wise. And there would be no need to switch back and forth. Modern OS use paging subsystem for files anyway (unless opened with O_DIRECT or similar).

> I think in my code, what I do is if someone wants to read, and is
> currently writing (on a read/write stream), I flush the buffer, then
> switch it to be in read mode, and then seek to the appropriate place, to
> fill the buffer.
>
> When switching from read to write, it's much easier, since the buffer
> starts out empty.
>
> So I think a buffer type that can be used for read *OR* write is good,
> just not simultaneously.

See above. I don't think there is a need for these paces with flushing stuff. And if it comes to pipes and sockets then just having 2 buffers fits the bill better.

> However, simultaneous read/write is probably something that could be
> useful for a niche use case (e.g. some transformative process), as long
> as you don't mind having 2 file descriptors open :)
>


-- 
Dmitry Olshansky

On Fri, 17 Jan 2014 10:40:02 -0500, Dmitry Olshansky <dmitry.olsh@gmail.com> wrote:

> 17-Jan-2014 18:52, Steven Schveighoffer пишет:
>> Keep in mind that a read/write buffered stream *should* exist. It's just
>> that you need to switch how you use the buffer.
>
> I'm doubting that actually. Just use the freaking MM-file directly you'd better off efficiency wise. And there would be no need to switch back and forth. Modern OS use paging subsystem for files anyway (unless opened with O_DIRECT or similar).

Remember we have to support whatever stdio.File does, and I think it does support this.

-Steve