You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@jena.apache.org by "Fell, Simon" <sf...@ebay.com.INVALID> on 2020/01/14 22:36:26 UTC
TypeMapper thread safety
Hi,
I was looking at org.apache.jena.datatypes.TypeMapper and noticed that it’s a shared instance, but it manages state in regular old HashMaps. Given that getSafeTypeByName can end up calling registerDatatype, I don’t think its safe to assume that these hashmaps are only written at construction/reset time. Seems like these should be swapped to ConcurentHashMaps.
Thoughts? I’d be happy to put together a patch for this.
Cheers
Simon
Re: TypeMapper thread safety
Posted by Andy Seaborne <an...@apache.org>.
Excellent!
On 22/01/2020 18:35, Fell, Simon wrote:
> PR Submitted https://github.com/apache/jena/pull/676
>
> Cheers
> Simon
>
> On 1/15/20, 2:11 AM, "Andy Seaborne" <an...@apache.org> wrote:
>
> Hi Simon,
>
> Makes sense to me to make that change - please do send in a patch.
>
> Andy
>
> On 14/01/2020 22:36, Fell, Simon wrote:
> > Hi,
> >
> > I was looking at org.apache.jena.datatypes.TypeMapper and noticed that it’s a shared instance, but it manages state in regular old HashMaps. Given that getSafeTypeByName can end up calling registerDatatype, I don’t think its safe to assume that these hashmaps are only written at construction/reset time. Seems like these should be swapped to ConcurentHashMaps.
> >
> > Thoughts? I’d be happy to put together a patch for this.
> >
> > Cheers
> > Simon
> >
>
>
Re: TypeMapper thread safety
Posted by "Fell, Simon" <sf...@ebay.com.INVALID>.
PR Submitted https://github.com/apache/jena/pull/676
Cheers
Simon
On 1/15/20, 2:11 AM, "Andy Seaborne" <an...@apache.org> wrote:
Hi Simon,
Makes sense to me to make that change - please do send in a patch.
Andy
On 14/01/2020 22:36, Fell, Simon wrote:
> Hi,
>
> I was looking at org.apache.jena.datatypes.TypeMapper and noticed that it’s a shared instance, but it manages state in regular old HashMaps. Given that getSafeTypeByName can end up calling registerDatatype, I don’t think its safe to assume that these hashmaps are only written at construction/reset time. Seems like these should be swapped to ConcurentHashMaps.
>
> Thoughts? I’d be happy to put together a patch for this.
>
> Cheers
> Simon
>
Re: TypeMapper thread safety
Posted by Andy Seaborne <an...@apache.org>.
Hi Simon,
Makes sense to me to make that change - please do send in a patch.
Andy
On 14/01/2020 22:36, Fell, Simon wrote:
> Hi,
>
> I was looking at org.apache.jena.datatypes.TypeMapper and noticed that it’s a shared instance, but it manages state in regular old HashMaps. Given that getSafeTypeByName can end up calling registerDatatype, I don’t think its safe to assume that these hashmaps are only written at construction/reset time. Seems like these should be swapped to ConcurentHashMaps.
>
> Thoughts? I’d be happy to put together a patch for this.
>
> Cheers
> Simon
>