[
https://issues.apache.org/jira/browse/NUTCH-1733?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sebastian Nagel updated NUTCH-1733:
-----------------------------------
Attachment: charset_bom_utf16_html5.html
Hi [~jlafitte], yes that's true: in a hex-editor it turns out, that it's a
space preceded by a BOM. Interestingly, it does not happen for UTF-16 (attached
sample file). However, inside the function {{sniffCharacterEncoding(byte[]
content)}} it's not possible to change {{content}}. We could open a separate
issue for the leading BOM (but, luckily, HTML with BOM appears considerably
rarely).
> parse-html to support HTML5 charset definitions
> -----------------------------------------------
>
> Key: NUTCH-1733
> URL: https://issues.apache.org/jira/browse/NUTCH-1733
> Project: Nutch
> Issue Type: Bug
> Components: parser
> Affects Versions: 1.8, 2.2.1
> Reporter: Sebastian Nagel
> Fix For: 2.3, 1.9
>
> Attachments: NUTCH-1733-trunk.patch, charset_bom_html5.html,
> charset_bom_utf16_html5.html, charset_html5.html
>
>
> HTML 5 allows to specify the character encoding of a page per
> * {{<meta charset="...">}}
> * Unicode Byte Order Mark (BOM)
> These are allowed in addition to previous HTTP/http-equiv Content-Type, see
> [[1|http://www.w3.org/TR/2011/WD-html5-diff-20110405/#character-encoding]].
> Parse-html ignores both meta charset and BOM, falls back to the default
> encoding (cp1252). Parse-tika sets the encoding appropriately.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
