[Nepomuk] Review Request 110794: FileIndexer: Do not attempt to index files multiple times

[Simeon Bird]

On Sat, 8 Jun 2013 04:13:35 PM Simeon Bird wrote:
>> FileIndexer: Do not attempt to index files multiple times
>>
>> If a file indexing job fails, set kext::indexingLevel to 4, so
>> that the next time the file queue fills, it is omitted.
>> This means we don't burn CPU trying to index things multiple times.
>
> What would happen if a file was in the middle of downloading and hence its 
> data is probably not valid, e.g a music file being downloaded by bittorrent.
>
> Would it be flagged as failed and never reindexed, causing its meta data to be 
> ignored?
>
> Relevant to me at the moment as I'm downloading my entire Google Music 
> collection, some 12 GB/2000 files of data.

It should be ok. Files that are modified are re-indexed anyway,
irrespective of their current indexing level. This is only for not-yet
indexed files.
And of course in an emergency you can always re-index a file yourself
manually by calling the dbus methods of the FileIndexer.

Simeon