Create showapi.com.xml

[gloomy-ghost]

No description provided.

[jeremyn]

Here's what I would do:

<!--
	200 with plain HTTP, 404 with HTTPS:
	- docs.showapi.com
	  example: http://docs.showapi.com/jdk7/index.html

	503 with plain HTTP, refused with HTTPS:
	- wap.showapi.com

	No working URL known:
	- static.showapi.com

-->
<ruleset name="ShowAPI.com">
	<target host="showapi.com" />
	<target host="www.showapi.com" />
	<target host="bbs.showapi.com" />
	<target host="route.showapi.com" />

	<rule from="^http://bbs\.showapi\.com/(archiver|data|source|static)/" to="https://bbs.showapi.com/$1/" />
		<test url="http://bbs.showapi.com/archiver/" />
		<test url="http://bbs.showapi.com/data/" />
		<test url="http://bbs.showapi.com/source/" />
		<test url="http://bbs.showapi.com/static/" />
		<!-- This exception disables the implicit test created by the target. -->
		<exclusion pattern="^http://bbs\.showapi\.com/$" />
	<rule from="^http://(www\.|route\.)?showapi.com/" to="https://$1showapi.com/" />
</ruleset>

Also the filename would be ShowAPI.com.xml.

Do you want to add dn-showapi.qbox.me to Qbox.me.xml?

[gloomy-ghost]

I was using exceptions to keep the plain rewriting rule, otherwise the rules are basically just repeating subdomains that are already included in targets. Maybe complicated rules that don't do anything else than http>https redirecting would be a good method if there were only wildcard targets.

@Hainish Would it be possible to add something likes a no_plain_rewriting attribute in targets so that a target which is only partially redirected by rule won't be redirected by http:>https: rule in the end of the ruleset?

[jeremyn]

The ruleset I wrote in #7957 (comment) is simple: for these domains, do these rewrites. This is a lot easier to understand in my opinion than saying rewrite everything except these other certain things.

I'm opposed to something like no_plain_rewriting because that means that to figure out what the rewrite logic is, I can't just look at the rule, I have to also possibly look at attributes of the targets.

[Hainish]

I agree with @jeremyn on this - adding an attribute which changes how rewrites are performed also complicates things for our downstream dependencies, like Brave browser or libhttpseverywhere.

[jeremyn]

@gloomy-ghost Can we finish this?

[gloomy-ghost]

Is there a url for actual resource under http://bbs.showapi.com/source/? I can't find any.

BTW, I still think that listing all subdomains in a plain rewriting rule is repeating the targets' jobs. If everything is in rules, I can replace all targets with a wildcard and it won't change anything. Then why there are target tags?

[jeremyn]

I can't find anything under http://bbs.showapi.com/source/ either, but it returns a 200 even though it's empty so that's good enough for me.

I don't know how to answer the question in your second paragraph. Maybe a better system would merge the concept of target and rule.

[jeremyn]

• https://docs.showapi.com is Different HTTP/HTTPS content, not 404
• The general case *.showapi.com works for me (they all redirect to www), so you can remove it from the top comment
• Also u5335.showapi.com redirects to https://www.showapi.com like the general * case, so you can delete that from the top comment too
• The u(\d+)\.showapi\.com rule here doesn't have a target, so please either delete the rule or add a target
• I would move the rule for ^ here into the larger rule here, using (www\.|admin\.|api\.| ...)? but if you prefer the style you have now, that's fine with me.

[jeremyn]

I made the changes I described in #7957 (comment).

@J0WI Can you review this please?