beneroth changed the topic of #picolisp to: PicoLisp language | The scalpel of software development | Channel Log: https://libera.irclog.whitequark.org/picolisp | Check www.picolisp.com for more information
aw- has joined #picolisp
aw- has quit [Quit: Leaving.]
aw- has joined #picolisp
tankf33der[m] has joined #picolisp
rob_w has joined #picolisp
aw- has quit [Ping timeout: 240 seconds]
<beneroth> abu[m], are you here?
<abu[m]> yeap!
<beneroth> yeah
<beneroth> asking again.. probably I forgot earlier answer from you...
<beneroth> how to decide block sizes for database index - primarily on index counter ?
<beneroth> and ofc string/fold/aux indexes are to be considered bigger than e.g. (+Key +Number)
<abu[m]> Primarily on the key size. Best is to look how many entries fit into a block
<beneroth> good way to determine key size?
<beneroth> using (step) to build a statistic? :-/
<abu[m]> (size ("abc" {A0} . {X}))
<abu[m]> or (vi {1}) then click to the nodes
<abu[m]> look at the lists in each block
<beneroth> ok.. so handpicking some of them (full calculation with step I consider to costly / not worth much)
<abu[m]> (length (val ...))
<abu[m]> yes, take the worst case
<abu[m]> longest
<beneroth> trees are never loaded fully, but as required, right?
<abu[m]> and then at least 20 or 30 entries should be in a node
<abu[m]> yes, only those nodes needed
<beneroth> so even for big key size you want smaller blocksize if usage is always only a few parts vs. cases where likely majority tree is accessed a lot?
<abu[m]> a path through the tree
<beneroth> T
<abu[m]> It would be bad if a key does not fit at all into a node
<beneroth> ah, yeah of course
<abu[m]> Then the node takes several blocks
<beneroth> yeah I see. would still work, but pretty bad
<abu[m]> right
<beneroth> ok thanks :)
rob_w has quit [Quit: Leaving]
<beneroth> [security] serious vulnerability found in Linux Kernel (also affecting Android): https://dirtypipe.cm4all.com/
aw- has joined #picolisp