[prev in list] [next in list] [prev in thread] [next in thread] 

List:       avro-dev
Subject:    Re: computing the size of an avro encoded message
From:       Martin Grigorov <mgrigorov () apache ! org>
Date:       2023-05-19 6:29:06
Message-ID: CAMomwMq+e+j4W0xKCkZ6rBurrAEGnZhW9omRzU3RO4vi3NqiZw () mail ! gmail ! com
[Download RAW message or body]


Hi Andrew,

On Thu, May 18, 2023 at 12:12 PM Andrew Marlow <marlow.agents@gmail.com>
wrote:

> Hello everyone,
>
> I have just run into the problem discussed on SO at
>
> https://stackoverflow.com/questions/43738118/size-of-encoded-avro-message-without-encoding-it
>
> I am using avro from C++ and currently have a payload buffer of fixed size
> into which I do the encode. I found that for some messages I can get an
> exception thrown with the error "EOF found". I worked around the problem by
> increasing the buffer size, experimenting with binary chop until I found
> something that worked without being too wasteful. But this is tiresome.
> There doesn't seem to be a way to find out how many bytes are required
> except by trying, catching the exception, and then retrying with a larger
> buffer. What can be done please?
>
> I note that in good ol' sprintf one can supply the null pointer for the
> target buffer and it will not try to write into it but will return the
> number of bytes it would have written. This allows one to allocate a buffer
> of exactly the right size. Is there any chance of adding such a facility to
> the C++ avro API please?
>

I think it is fine to add this kind of functionality to the C++ APIs!
Combined with the "exception catching" one will be able to expand the
buffer with a better precision!



> --
> Regards,
>
> Andrew Marlow
> http://www.andrewpetermarlow.co.uk
>


[prev in list] [next in list] [prev in thread] [next in thread]