Issue extracting archives

Since last Beta I noticed some issues when extracting archives.

When extracting (using context entry of DO, not WinRAR etc.), either extracting-window opens very late or extracting finishes w/o any windows (large archives which take some time to extract).

Extracting with WinRAR works fine and I tested different archives.

Which kind(s) of archives?

Are there some examples we can try with?

Zip and RAR.

The examples unfortunately are all 1GB+, but it seems that it happens randomly - yesterday I extracted an archive and long time nothing happened (no window), today same archive was extracted w/o problems.

I think I need to investigate for more details, but maybe others noticed same behaviour.

Could be caused by antivirus the first time the archive is scanned, maybe.

Or by caching the folder listing or disk data so it only takes a long time to open the archive the first time.

AV is not installed and Win-10-Defender would cause high CPU, which is not the case. The archives include only a few files and 2-3 folders.

I did not much changes to my sys, but I'll restore it next days to a "clean" state and have a look if it occurs again. Thanks.

After some investigation: DO needs very long time to count the files in archives (even if only a few files are included). It took always xx seconds and nothing is accessing the file (no antivirus, no high task usage, no other access to hdd).

Some archive formats (especially RAR prior to RAR v5) can take a long time to list a lot of files, since they have to scan through the entire archive. Same no matter which tool you open them with (but remember the second one you try will probably benefit from the first one causing the data to be cached in memory).

Sure, but not zips with one 1-2 files (~1GB and larger).

Edit: This behaviour never happened in previous versions.

Are you sure, and tested with the same zips?

It's possible there are some changes to do with whether Opus depends on the zip central directory vs individual file records (since some tools write the wrong information to one or the other), but I don't remember anything related to that changing in quite a long time and I suspect the difference you are seeing is being triggered by an outside factor.

If you can verify different versions are much slower/faster with the same zips, we can look into it. We'd need to know how to obtain or create similar zips as well.