Close comics during scan

[nichwall]

This PR partially addresses some of the CBR/CBZ issues.

As mentioned in #2750, the memory does not get freed until after the comics have completed parsing. This PR adds the archive.close() call at the end of the extraction functions.

I tested this by:

• Grabbing a number of random images from my personal photos
• Created a 1.8 GB CBZ and CBR comic file (no cover image, just ran the directory of images through the zip and rar bash utilities)
• Ran a scan without these changes (server crashed during scan of second book)
• Ran a scan including these changes (server completed scan of both books)

I don't have any "real" comics, so additional testing should be done on edge before this is part of a full release. My homemade comics did not actually load in the ereader, but were parsed during the scan.

[advplyr]

Thanks!

[Nda70]

Hi, first comment here and french, be kind.
last edge image from 9h ago still occupy all ram in 1mn and finish by crash.
i'm scannig comics big ,but not huge, library (cbr, cbz, pdf)
Docker 2.12.3 hosted on synology NAS.
On debug log nothing appear, it begin a new parsing and log stop, even crash_log.txt isn't created.
daily logs appear normally, but scans and crash ones are not.
Last log an container manager says:
wasm streaming compile failed: typeerror: failed to parse url from /client/dist/libarchive/wasm-gen//libarchive.wasm
then
falling back to arraybuffer instantiation.
then crash.
Hope it could helps, thanks for your work here.
Have a nice day.

[advorsky73]

Same here.... Still crashing with around 400 cbr/cbz files (i have 64gb ram on that machine though)