Hi,
Is there any support (future processors in next webharvest version or plugin(s) available) for parallel execution of processors in WebHarvest ?
Parallel execution would be really helpful when fetching data from multiples urls for instance.
I draw my thoughts here :
<config>
<threadpool count="5">
<body>
// Here i would call some processors for workng on the urls
// The processors here are run in parallel
</body>
<on-success>
// Optionnal tag containing processors to be executed each time one thread has successfully finished
</on-success>
<on-error>
// Optionnal tag containing processors to be executed each time one thread has finished abruptly
</on-error>
</threadpool>
</config>
Stephan
Nice question, Stephan.
Currently there is no concurrency support in WH. neither in v.2.0 or 2.1.
I also think it would be a good feature. I have raised this idea earlier, but nobody seemed to be excited about it.
Before the concurrency support can be added the WH processors needs to be refactored first to ensure their immutability.
For me the biggest issue now is time. If anybody could help with that, that would be perfect.
If BaseProcessor is immutable, is it enough ?
Log in to post a comment.
Hi,
Is there any support (future processors in next webharvest version or
plugin(s) available) for parallel execution of processors in WebHarvest ?
Parallel execution would be really helpful when fetching data from multiples
urls for instance.
I draw my thoughts here :
<config>
<threadpool count="5">
<body>
// Here i would call some processors for workng on the urls
// The processors here are run in parallel
</body>
<on-success>
// Optionnal tag containing processors to be executed each time one thread has
successfully finished
</on-success>
<on-error>
// Optionnal tag containing processors to be executed each time one thread has
finished abruptly
</on-error>
</threadpool>
</config>
Stephan
Nice question, Stephan.
Currently there is no concurrency support in WH. neither in v.2.0 or 2.1.
I also think it would be a good feature. I have raised this idea earlier, but
nobody seemed to be excited about it.
Before the concurrency support can be added the WH processors needs to be
refactored first to ensure their immutability.
For me the biggest issue now is time. If anybody could help with that, that
would be perfect.
If BaseProcessor is immutable, is it enough ?