René J.V. Bertin
rjvbertin at gmail.com
Sat Apr 7 09:48:48 UTC 2018
On Saturday April 07 2018 11:14:20 Sven Brauch wrote:
>I think Milian investigated this once and concluded that no database on
>the market can do the lookups we need to do in an efficient-enough
>From what I see, topducontextdynamicdata.cpp has had a minor change in 2016, a bit more by Milian in 2014, and the rest is unchanged for 10 years. It's almost inevitable that new database solutions have seen the light since, esp. lean and mean ones.
Also: was this tested with projects or sessions large enough that the file IO starts being affected by the number of files in the directory, on file systems (and/or -fragmented- HDDs) where this can become an issue?
>Also, this is a huge amount of work to change.
That I believe (but limited to topducontextdynamicdata.cpp it seems?)
>That said, if there is a solution with the same or superior performance
>and somebody is willing to do the porting, I'm all for it ...
I'm not unwilling here, but utterly unprepared. I've done a bit of tinkering though, and from the looks of it the files are written only after all queued files finished being parsed, so that part at least will unlikely be a bottleneck.
It also seems that the code is designed to cope with write failure. Exactly would you miss/lose when these topcontext files are not saved?
More information about the KDevelop-devel