Serialized instead of parallel path finding

[hendrikvanantwerpen]

This PR implements an idea I raised in #346: instead of processing all initial paths in parallel, process them in serial.

Note that I've made that change for find_minimal_partial_path_set_in_file and find_all_complete_partial_paths, not the stitcher itself. Especialyl the former is often called with initial paths for all clickable nodes in the graph, and the effect on max memory usage should be noticable, I hope.

[github-actions]

Performance Summary

Comparing base cd58abb with head 04aebfc on typescript_benchmark benchmark. For details see workflow artifacts. Note that performance is tested on the last commits with changes in stack-graphs, not on every commit.

Before

--------------------------------------------------------------------------------
Command:            base/target/release/tree-sitter-stack-graphs-typescript index -D base.sqlite --max-file-time=30 --hide-error-details -- base/data/typescript_benchmark
Massif arguments:   --massif-out-file=perf.out
ms_print arguments: --x=72 --y=12 base-perf-results/perf.out
--------------------------------------------------------------------------------


    GB
3.276^                                                                     :# 
     |                                                                     :# 
     |                                                                     :# 
     |                                                                     :#:
     |                                                                     :#:
     |                                                                     :#:
     |                                             :::@                : :@:#:
     |                                             :  @                :::@:#:
     |                                          @@@:  @               ::::@:#:
     |                                       :::@  :  @               ::::@:#:
     |                                 @:   ::  @  :  @        ::@:  @::::@:#:
     |              :  :@@     :    ::@@:: @::  @  :  @       :::@:::@::::@:#:
   0 +----------------------------------------------------------------------->Gi
     0                                                                   74.84

After

--------------------------------------------------------------------------------
Command:            head/target/release/tree-sitter-stack-graphs-typescript index -D head.sqlite --max-file-time=30 --hide-error-details -- head/data/typescript_benchmark
Massif arguments:   --massif-out-file=perf.out
ms_print arguments: --x=72 --y=12 head-perf-results/perf.out
--------------------------------------------------------------------------------


    MB
715.3^                                                           ##           
     |                                                           #            
     |                                                           #            
     |                                 @@:                     ::#            
     |                                 @ :           ::        ::#            
     |                                 @ :           ::        ::#            
     |                                 @ :::    :::::::        ::#            
     |              :     :            @ :::    ::: :::       @::# ::     ::::
     |              :::   :            @ :::  ::::: :::       @::# :    ::::::
     |              ::  ::: :         :@ :::@@::::: :::       @::# : :::::::::
     |              :: :: : ::  :::   :@ :::@ ::::: ::::  ::  @::# : :::::::::
     |             @:: :: : ::@@: ::  :@ :::@ ::::: ::::  : ::@::# : :::::::::
   0 +----------------------------------------------------------------------->Gi
     0                                                                   70.72

[hendrikvanantwerpen]

Hm, there seems some variation in those profiling runs, because the before graph is much higher than the after graph in #346 , even though they are supposed to be based on the same commit 😕.

Comparing the after graphs of this PR and #346 suggests maybe a slight decrease in memory usage? I'll do some more local experiments to see if it shows better for e.g. larger files.

[hendrikvanantwerpen]

Closing this old experiment.