[
https://issues.apache.org/jira/browse/KNOX-578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16563932#comment-16563932
]
Wei Han commented on KNOX-578:
------------------------------
[~jeffreyr97] [~kminder] Was there a follow up on this JIRA? Looks like I'm
facing a similar issue with jobhistory UI, which has the following meta refresh
tag:
{quote} <meta http-equiv="refresh" content="1;
url=http://hadoopmaster007:19888/jobhistory/logs/hadoopworker3064:8041/container_e50_1529521645530_115338_01_000001/container_e50_1529521645530_115338_01_000001/xxxx";>
{quote}
> Supporting Regex patterns on rewrite patters accross Html Filtering beyon
> embeded JS
> ------------------------------------------------------------------------------------
>
> Key: KNOX-578
> URL: https://issues.apache.org/jira/browse/KNOX-578
> Project: Apache Knox
> Issue Type: Improvement
> Components: Server
> Affects Versions: 0.7.0
> Environment: All
> Reporter: Jeffrey E Rodriguez
> Priority: Major
> Fix For: Future
>
>
> The reason for the this is that we already have some Regex support
> on embedded Javascript but it doesn't seem to work on some other cases like
> in Html elements like.
> meta element which is used by Yarn UI for example where some
> of the UI pages uses them to do a meta refresh.
> e.g.
> Let say we want the following outbound html content to go through Knox
> <meta http-equiv="refresh" content="0; url=http://service";>
> So we would like to rewrite it to:
> <meta http-equiv="refresh" content="0;
> url=http://knox_host:knox_port/gateway/default/service";>
> I can match the entire "0; url=http://service"; with a rule dir="OUT"
> but I don't have control on how to parse the refresh time from the url so if
> the page changes tomorrow I respect it in the rewrite.
> If Regex would be supported then we could use something like (\d);\s*(.*) to
> capture the refresh time in the rule and use
> it in the rewrite template.
> The case above is a simple one, but having better control through Regex on
> how to parse patterns and use matched groups would be desirable.
> Kevin Minder, said the following:
> "Something like
> <apply-regex path=“{xpath}” regex=“{regex}” rule=“{rule}”/>
> Where the rule would be applied to say group(1) of the regex. Might even be
> able to build that into the normal <apply/> with an implied regex=“.*”.
> "
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
