Re: Expat loading the entire file in memory?

XML_MIN_SIZE only affects the size of the library. It is not meant for
decreasing memory usage.

Also, Expat should not use 500MB on a 1GB file, unless that file contains a
500MB long element name or such things.

Maybe you should post your code.

On Mon, Apr 4, 2011 at 5:27 AM, Guillaume Drevon <
[email protected]> wrote:

> I have been able to solve my issue. In fact there is a macro that you need
> to define before building expat lib: XML_MIN_SIZE
>
> With this, expat will keep the memory usage very low. In my case, it was
> below 3MB, which is much much less than 500MB!
>
> Expat documentation warns that this option will make expat running slower
> so this is why it is not the default behavior. This makes sense as it will
> work well except for extreme files like mine.
>
> Note that expat was not loading the entire file in memory, as else it would
> have needed 1GB due to the unicode conversion.
>
>
>
>
>
> On Wed, Mar 30, 2011 at 12:08 PM, Guillaume Drevon
>
> <[email protected]> wrote:
>
> > I am trying to use Expat for parsing a 500MB file. I read the file by
> chunks (of 1000 byte) and send them to the parser. However, the memory used
> by my application goes up to 550 MB.
>
> > The memory usage is entirely done by expat as the XmlParserFree does
> remove almost all memory consumption. I see that the memory is mostly used
> by the m_buffer member of the parser.
>
> >
>
> > Am I doing something wrong? The purpose of this SAX-like parser is not
> memory usage?
>
> _______________________________________________
> Expat-discuss mailing list
> [email protected]
> http://mail.libexpat.org/mailman/listinfo/expat-discuss
>