Pull request #3128: JS Normalization: single pass processing

Pull request #3128: JS Normalization: single pass processing

Merge in SNORT/snort3 from ~OSHUMEIK/snort3:spp to master

Squashed commit of the following:

commit f09974f5dca6d48223f441e61ccd1b7676fd64e2
Author: Oleksii Shumeiko <oshumeik@cisco.com>
Date:   Fri Oct 22 15:55:56 2021 +0300

    utils: correct Normalizer's output upon the next scan

    The output stream buffer was updated with a special-case code to speed up
    getting the output size.

commit 0f66f7491fcd07c44934a4a473d26354dd39a859
Author: Oleksii Shumeiko <oshumeik@cisco.com>
Date:   Mon Oct 18 16:23:35 2021 +0300

    http_inspect: eliminate cumulative js data processing

    Input data is fed by portions (script_detection, chunked HTTP) to JSNormalizer.
    Output data is accumulated in output stream buffer, which resides in
    JSNormalizer context. Accumulated output data is deleted at the end of PDU.

commit 7fe0cc81badb99a2a732c74cddc1aa042e40cbd2
Author: Oleksii Shumeiko <oshumeik@cisco.com>
Date:   Fri Oct 15 16:50:09 2021 +0300

    utils: add get methods to peek in internal buffer