Search in GTK installer takes too much time

[browny]

TC 11.03 64-bit in Wndows 11 and Windows Server 2019.
File size of GTK runtime installer is below 50 MB (about 300 MB unpacked); archive test in TC and 7-zip GUI shows no errors.
Entered into the installer with Ctrl+PgDn
Search text: bullet-
Marked are only ANSI, ASCII, UTF-16 and UTF-8

TC finds something, but keeps running with ~4% of CPU - search has to be stopped manually.

Edit. TC was left running on Windows Server 2019 in RDP session while this report was being written. It took quite a few minutes but then it seems the search has completed. Disk usage or speed are not an issue, Windows 11 uses SSD.

Edit 2. This installer most probably is a solid archive, unpacking a single file is quite slow. Just a guess, multiple unpackings, every time starting from the beginning of the file might be the reason.

[ghisler(Author)]

When you press Ctrl+B in the archive, you can see that it contains 12682 files. It takes quite a while to unpack each of them and then search in them individually. I will check whether there is a way to optimize it.

Can you check how the file was opened by pressing Alt+Enter on a file inside of it? Here it opened with Total7zip.wcx.

[browny]

Plugin Total7zip.wcx64.
For a comparison, source code in a 75 MB .zip, unpacked 285 MB, 7500 files.
I did not bother to use stopwatch, because search with exactly the same parameters completes in about 3 seconds.

For a solid archive it should be optimal to unpack and search/compare in the order of files in the archive.

[ghisler(Author)]

Plugin Total7zip.wcx64.

I see - moved to plugins forum.

Update: I have analyzed the file, it is NOT a 7zip archive! It's a Nullsoft installer (NSIS) file. The Total7zip plugin can unpack it via 7z dll, but the plugin is slow with handling solid archives.

[browny]

Unpacking of the whole contents of the installer from SSD to HDD took about 1 minute only.
Searching in the unpacked subdirectories took about 2 seconds.
If unpacking does not need to be repeated several times from the beginning of the file, what could make search so slow?