Sorry, there is no magic solution here.  The best suggestions I have are
1) Use SerializeToString() instead, which will always allocate a string
large enough to hold the message.

2) Build your message, then call ByteSize(), and if it is too large, remove
some fields and try again.  Once it is big enough, call
SerializeWithCachedSizesToArray() -- since you already called ByteSize(),
the sizes are already cached.  This way, if the message fits on the first
attempt, you perform just as well as you did before, because
SerializeToArray() calls ByteSize() followed by

On Wed, Jul 8, 2009 at 4:25 AM, edan <> wrote:

> Hi-
> We have a fixed-size buffer into which we are serializing a message (a
> pretty large and convoluted message, with many optional sub-messages and
> repeated fields, etc).
> We have found that when the message size gets bigger than the buffer size,
> SerializeToArray just fails to write anything - we lose the whole message.
> We'd like to be able to write as much message as we can, and just give up
> on some of the optional fields, or, more likely, to stop adding elements to
> repeated fields, etc.
> What is the best way to do this efficiently?  It seems that calling
> ByteSize() all the time will cost us in performance.  Is there a way to
> efficiently compute the space needing during our message-building phase?  Or
> is there a way to serialize as we go using some of the append methods, and
> then just stop appending when we reach the limit?  What I don't understand
> is how to avoid re-appending the same parts of the message that we already
> serialized?  Let's also assume that most of the places we are likely to "run
> out of space" is while we are adding more and more elements to repeated
> fields.
> Thanks for any help.
> --edan
> >

You received this message because you are subscribed to the Google Groups 
"Protocol Buffers" group.
To post to this group, send email to
To unsubscribe from this group, send email to
For more options, visit this group at

Reply via email to