Posted 16 years ago
by Damir Bulic
![Avatar](https://secure.gravatar.com/avatar/c72584fe6d2d7a048ba64ffc57069de1.jpg?s=64&d=identicon&r=g)
I'd like to vote for the background parser service changes.
Now it's a queue and new requests are blocked until previous ones are handled. So I get a 5MB parse blocking 500 chars request.
I propose a pool of threads so that new small requests are done immeadiately if there are enough threads available. We could put a number of threads in the constructor. Also, ability to assign a priority to request would be additional nice touch.
I think it even shouldn't be a service, it should be a helper object providing background parsing. As it is now (not fitting to my requirements), I have to turn the service off in every application.
One more thing, two hashes passed to the service are a bit confusing to me. Why not simply add the new request and get an integer back? I think .NET auto increments internal counter each time an object is created, maybe we could use that to identify the Document parsed.
Now it's a queue and new requests are blocked until previous ones are handled. So I get a 5MB parse blocking 500 chars request.
I propose a pool of threads so that new small requests are done immeadiately if there are enough threads available. We could put a number of threads in the constructor. Also, ability to assign a priority to request would be additional nice touch.
I think it even shouldn't be a service, it should be a helper object providing background parsing. As it is now (not fitting to my requirements), I have to turn the service off in every application.
One more thing, two hashes passed to the service are a bit confusing to me. Why not simply add the new request and get an integer back? I think .NET auto increments internal counter each time an object is created, maybe we could use that to identify the Document parsed.