When compare on two big folder(2G, but little diffs), exclude file is slow.

[zw963]

[Aaron]

Hello,

If you start a new session with no previously set filters and load these base folders, s this slow to select and exclude user.db? My hunch is that your existing filter string is very long and might be slowing things down a bit when processing and appending another filter to a long list.

[zw963]

Yes, you are right. in fact, my beyond compare start with a filter args.

When i do a compare on a git project, i check .gitignore, and convert both .gitignore content
into filter, e.g. for current case, the filter have 156 records, i will run command like this:

filter="
[FONT=monospace][COLOR=#000000]-*~;-.#*;-.*~;-05-11-2016.txt;-05-12-2016.txt;-agoda.txt;-airbnb_2015_2.csv;-airbnb_2015-3.csv;-Airhost-c61b4775d139.p12;-airhost-init-d.sh;-airhost_keys.json[/COLOR]
;-airhost_keys.p12;-airhost_local.db;-airhosts.db;-airhost-sidekiq.conf;-/app/assets/webpack/*;-app/javascript/mock.d.ts;-app/javascript/mock.d.ts\#*;-booking
_with_error2.log;-booking_with_error.log;-.bundle/;-/.bundle;-.byebug_history;-/cache;-/capybara*;-/capybara-*.html;-components/api/log/*.log;-components/api/
spec/dummy/log/*.log;-/config/application.yml;-/config/config.yml;-/config/initializers/braintree_debug.rb;-config/secrets.yml.key;-config/solargraph.rb;-conf
ig/symmetric-encryption.yml;-/coverage/;-/coverage.data;-/coverage-frontend/;-data.yml;-/db/dumps;-/db/*.javadb/;-db/schema.rb;-/db/*.sql;-/db/*.sqlite3;-db/*
.sqlite3;-/db/*.sqlite3-journal;-_deploy.xml;-*.diff;-/doc/api/;-/doc/app/;-/doc/features.html;-/doc/specs.html;-.DS_Store;-.elasticbeanstalk/;-*.elc;-.env;-.
env.*.local;-.env.local;-.env.test;-ery2.json;-ery.json;-features.psd;-/fixtures/req_cache/*.yml;-/gems;-homely.csv;-house.csv;-.idea;-*.iml;-insecure_private
_key;-iron.json;-jan.jpg;-latest.dump;-/lib/tasks/prof.rake;-locate.db;-/log/;-/log/*;-may.jpg;-migration.rb;-msg.json;-/nbproject/;-new_booking.sh;-new_inqui
ry.sh;-/node_modules;-node_modules;-npm-debug.log*;-**.orig;-*.orig;-*.patch;-pickle-email-*.html;-/pids;-.powenv;-/Procfile.dev;-Procfile.dev;-/Procfile.dev.
local;-/Procfile.prof;-procodile.log;-/profiles;-public/assets;-/public/cache;-/public/packs;-/public/packs-test;-/public/stylesheets/compiled;-/public/system
/*;-/public/uploads;-.rake_tasks~;-*.rbc;-*.rdb;-.redcar/;-rerun.txt;-.ruby-gemset;-.ruby-version;-.rvmrc;-*.sassc;-.sass-cache;-/scripts;-sidekiq.log;-sideki
q-spotfleet.json;-sorbet/rbi/hidden-definitions/errors.txt;-spec/examples.txt;-/specifications;-/spec/tmp/*;-*.sqlite3;-/ssl/;-storage;-*.sublime-project;-*.s
ublime-workspace;-**.swp;-sync.sh;-TAGS;-.tern-*;-test.jpg;-/test_shell/;-/test_shells/;-test_shellsagoda.html;-**.tmp;-/tmp/*;-tmp/;-*.tmproj;-todist.md;-.*.
~undo-tree~.gz;-unicorn.log;-unicorn.partial.log;-/vendor/assets/bower_components/;-/vendor/bundle/;-/vendor/ruby/;-/version_objects;-vivi.json;-.vscode;-/.vs
code;-**.war;-website1.json;-yarn-debug.log*;-/yarn-error.log;-yarn-error.log;-.yarn-integrity;-.zeus.sock;-#*#;-*~;-.#*.*;-.git/;-.gitignore[/FONT]
"

bcompare -filters="$filter" "$1" "$2"

I still consider this filter is not too large anyway, so, can we do some improvement on this?

[Aaron]

That filter isn't long enough to trigger a slowdown on my machine. If you remove the filter and repeat the attempt is it still slow for you? Also, which OS and OS version are you running?

[zw963]

If you remove the filter and repeat the attempt is it still slow for you?

No, very quickly, but if with it, almost need 4 seconds to finish exclude a file.

Also, which OS and OS version are you running?

Arch Linux, i guess it should slow for another distro too.

[Aaron]

It's a combination of the filter complexity and the items it needs to apply to in the current view. Improving the performance in these scenarios is something on our wishlist, and I'll add your filter string to our test cases as an example of a shorter filter that can be slow.