Listens for
- webhooks from GitHub/GitLab/Bitbucket.
+ webhooks from GitHub/GitLab/Bitbucket/Azure Devops.
Listens for
- webhooks from GitHub/GitLab/Bitbucket.
+ webhooks from GitHub/GitLab/Bitbucket/Azure Devops.
Runs terraform
commands remotely and comments back with their output.
diff --git a/runatlantis.io/docs/access-credentials.md b/runatlantis.io/docs/access-credentials.md
index a09df508b..abe02174c 100644
--- a/runatlantis.io/docs/access-credentials.md
+++ b/runatlantis.io/docs/access-credentials.md
@@ -1,5 +1,6 @@
# Git Host Access Credentials
-This page describes how to create credentials for your Git host (GitHub, GitLab or Bitbucket)
+This page describes how to create credentials for your Git host (GitHub, GitLab, Bitbucket, or Azure Devops)
+
that Atlantis will use to make API calls.
[[toc]]
@@ -19,6 +20,7 @@ generate an access token. Read on for the instructions for your specific Git hos
* [GitLab](#gitlab)
* [Bitbucket Cloud (bitbucket.org)](#bitbucket-cloud-bitbucket-org)
* [Bitbucket Server (aka Stash)](#bitbucket-server-aka-stash)
+* [Azure Devops](#azure-devops)
### GitHub
- Create a Personal Access Token by following: [https://help.github.com/articles/creating-a-personal-access-token-for-the-command-line/#creating-a-token](https://help.github.com/articles/creating-a-personal-access-token-for-the-command-line/#creating-a-token)
@@ -48,5 +50,14 @@ Your Atlantis user must also have "Write permissions" (for repos in an organizat
- Give the token **Read** Project permissions and **Write** Pull request permissions
- Click **Create** and record the access token
+### Azure Devops
+- Create a Personal access token by following [https://docs.microsoft.com/en-us/azure/devops/organizations/accounts/use-personal-access-tokens-to-authenticate?view=azure-devops](https://docs.microsoft.com/en-us/azure/devops/organizations/accounts/use-personal-access-tokens-to-authenticate?view=azure-devops)
+- Label the password "atlantis"
+- The minimum scopes required for this token are:
+ - Code (Read)
+ - Code (Status)
+ - Work Items (Read & write)
+- Record the access token
+
## Next Steps
Once you've got your user and access token, you're ready to create a webhook secret. See [Creating a Webhook Secret](webhook-secrets.html).
diff --git a/runatlantis.io/docs/apply-requirements.md b/runatlantis.io/docs/apply-requirements.md
index 1d30ca1f9..75189dc3b 100644
--- a/runatlantis.io/docs/apply-requirements.md
+++ b/runatlantis.io/docs/apply-requirements.md
@@ -49,6 +49,7 @@ Each VCS provider has different rules around who can approve:
* **Bitbucket Cloud (bitbucket.org)** – A user can approve their own pull request but
Atlantis does not count that as an approval and requires an approval from at least one user that
is not the author of the pull request
+* **Azure Devops** – **All builtin groups include the "Contribute to pull requests"** permission and can approve a pull request
:::tip Tip
If you want to require **certain people** to approve the pull request, look at the
@@ -120,6 +121,20 @@ merge. We don't check anything else because Bitbucket's API doesn't support it.
If you need a specific check, please
[open an issue](https://github.com/runatlantis/atlantis/issues/new).
+#### Azure Devops
+In Azure Devops, all pull requests are mergeable unless there is a conflict. You can set a pull request to "Complete" right away, or set "Auto-Complete", which will merge after all branch policies are met. See [Review code with pull requests](https://docs.microsoft.com/en-us/azure/devops/repos/git/pull-requests?view=azure-devops).
+
+[Branch policies](https://docs.microsoft.com/en-us/azure/devops/repos/git/branch-policies?view=azure-devops) can:
+* Require a minimum number of reviewers
+* Allow users to approve their own changes
+* Allow completion even if some reviewers vote "Waiting" or "Reject"
+* Reset code reviewer votes when there are new changes
+* Require a specfic merge strategy (squash, rebase, etc.)
+
+::: warning
+At this time, the Azure Devops client only supports merging using the default 'no fast-forward' strategy. Make sure your branch policies permit this type of merge.
+:::
+
## Setting Apply Requirements
As mentioned above, you can set apply requirements via flags, in `repos.yaml`, or in `atlantis.yaml` if `repos.yaml`
allows the override.
@@ -186,3 +201,4 @@ request can run the actual `atlantis apply` command.
* For more information on GitHub pull request reviews and approvals see: [https://help.github.com/articles/about-pull-request-reviews/](https://help.github.com/articles/about-pull-request-reviews/)
* For more information on GitLab merge request reviews and approvals (only supported on GitLab Enterprise) see: [https://docs.gitlab.com/ee/user/project/merge_requests/merge_request_approvals.html](https://docs.gitlab.com/ee/user/project/merge_requests/merge_request_approvals.html).
* For more information on Bitbucket pull request reviews and approvals see: [https://confluence.atlassian.com/bitbucket/pull-requests-and-code-review-223220593.html](https://confluence.atlassian.com/bitbucket/pull-requests-and-code-review-223220593.html)
+* For more information on Azure Devops pull request reviews and approvals see: [https://docs.microsoft.com/en-us/azure/devops/repos/git/pull-requests-overview?view=azure-devops](https://docs.microsoft.com/en-us/azure/devops/repos/git/pull-requests-overview?view=azure-devops)
diff --git a/runatlantis.io/docs/configuring-webhooks.md b/runatlantis.io/docs/configuring-webhooks.md
index 51aba9dd1..6899d0021 100644
--- a/runatlantis.io/docs/configuring-webhooks.md
+++ b/runatlantis.io/docs/configuring-webhooks.md
@@ -85,7 +85,40 @@ If you're using GitLab, navigate to your project's home page in GitLab
- Click **Save**
- See [Next Steps](#next-steps)
+## Azure Devops
+Webhooks are installed at the [team project](https://docs.microsoft.com/en-us/azure/devops/organizations/projects/about-projects?view=azure-devops) level, but may be restricted to only fire based on events pertaining to [specific repos](https://docs.microsoft.com/en-us/azure/devops/service-hooks/services/webhooks?view=azure-devops) within the team project.
+
+- Navigate anywhere within a team project, ie: `https://dev.azure.com/orgName/projectName/_git/repoName`
+- Select **Project settings** in the lower-left corner
+- Select **Service hooks**
+ - If you see the message "You do not have sufficient permissions to view or configure subscriptions." you need to ensure your user is a member of either the organization's "Project Collection Administrators" group or the project's "Project Administrators" group.
+ - To add your user to the Project Collection Build Administrators group, navigate to the organization level, click **Organization Settings** and then click **Permissions**. You should be at `https://dev.azure.com/
+
+```
+
+To pass through mostly unaltered (it gained a rel="nofollow" which is a good thing for user generated content):
+```html
+
+
+
+```
+
+It protects sites from [XSS](http://en.wikipedia.org/wiki/Cross-site_scripting) attacks. There are many [vectors for an XSS attack](https://www.owasp.org/index.php/XSS_Filter_Evasion_Cheat_Sheet) and the best way to mitigate the risk is to sanitize user input against a known safe list of HTML elements and attributes.
+
+You should **always** run bluemonday **after** any other processing.
+
+If you use [blackfriday](https://github.com/russross/blackfriday) or [Pandoc](http://johnmacfarlane.net/pandoc/) then bluemonday should be run after these steps. This ensures that no insecure HTML is introduced later in your process.
+
+bluemonday is heavily inspired by both the [OWASP Java HTML Sanitizer](https://code.google.com/p/owasp-java-html-sanitizer/) and the [HTML Purifier](http://htmlpurifier.org/).
+
+## Technical Summary
+
+Whitelist based, you need to either build a policy describing the HTML elements and attributes to permit (and the `regexp` patterns of attributes), or use one of the supplied policies representing good defaults.
+
+The policy containing the whitelist is applied using a fast non-validating, forward only, token-based parser implemented in the [Go net/html library](https://godoc.org/golang.org/x/net/html) by the core Go team.
+
+We expect to be supplied with well-formatted HTML (closing elements for every applicable open element, nested correctly) and so we do not focus on repairing badly nested or incomplete HTML. We focus on simply ensuring that whatever elements do exist are described in the policy whitelist and that attributes and links are safe for use on your web page. [GIGO](http://en.wikipedia.org/wiki/Garbage_in,_garbage_out) does apply and if you feed it bad HTML bluemonday is not tasked with figuring out how to make it good again.
+
+### Supported Go Versions
+
+bluemonday is tested against Go 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, and tip.
+
+We do not support Go 1.0 as we depend on `golang.org/x/net/html` which includes a reference to `io.ErrNoProgress` which did not exist in Go 1.0.
+
+## Is it production ready?
+
+*Yes*
+
+We are using bluemonday in production having migrated from the widely used and heavily field tested OWASP Java HTML Sanitizer.
+
+We are passing our extensive test suite (including AntiSamy tests as well as tests for any issues raised). Check for any [unresolved issues](https://github.com/microcosm-cc/bluemonday/issues?page=1&state=open) to see whether anything may be a blocker for you.
+
+We invite pull requests and issues to help us ensure we are offering comprehensive protection against various attacks via user generated content.
+
+## Usage
+
+Install in your `${GOPATH}` using `go get -u github.com/microcosm-cc/bluemonday`
+
+Then call it:
+```go
+package main
+
+import (
+ "fmt"
+
+ "github.com/microcosm-cc/bluemonday"
+)
+
+func main() {
+ // Do this once for each unique policy, and use the policy for the life of the program
+ // Policy creation/editing is not safe to use in multiple goroutines
+ p := bluemonday.UGCPolicy()
+
+ // The policy can then be used to sanitize lots of input and it is safe to use the policy in multiple goroutines
+ html := p.Sanitize(
+ `Google`,
+ )
+
+ // Output:
+ // Google
+ fmt.Println(html)
+}
+```
+
+We offer three ways to call Sanitize:
+```go
+p.Sanitize(string) string
+p.SanitizeBytes([]byte) []byte
+p.SanitizeReader(io.Reader) bytes.Buffer
+```
+
+If you are obsessed about performance, `p.SanitizeReader(r).Bytes()` will return a `[]byte` without performing any unnecessary casting of the inputs or outputs. Though the difference is so negligible you should never need to care.
+
+You can build your own policies:
+```go
+package main
+
+import (
+ "fmt"
+
+ "github.com/microcosm-cc/bluemonday"
+)
+
+func main() {
+ p := bluemonday.NewPolicy()
+
+ // Require URLs to be parseable by net/url.Parse and either:
+ // mailto: http:// or https://
+ p.AllowStandardURLs()
+
+ // We only allow and
+ p.AllowAttrs("href").OnElements("a")
+ p.AllowElements("p")
+
+ html := p.Sanitize(
+ `Google`,
+ )
+
+ // Output:
+ // Google
+ fmt.Println(html)
+}
+```
+
+We ship two default policies:
+
+1. `bluemonday.StrictPolicy()` which can be thought of as equivalent to stripping all HTML elements and their attributes as it has nothing on its whitelist. An example usage scenario would be blog post titles where HTML tags are not expected at all and if they are then the elements *and* the content of the elements should be stripped. This is a *very* strict policy.
+2. `bluemonday.UGCPolicy()` which allows a broad selection of HTML elements and attributes that are safe for user generated content. Note that this policy does *not* whitelist iframes, object, embed, styles, script, etc. An example usage scenario would be blog post bodies where a variety of formatting is expected along with the potential for TABLEs and IMGs.
+
+## Policy Building
+
+The essence of building a policy is to determine which HTML elements and attributes are considered safe for your scenario. OWASP provide an [XSS prevention cheat sheet](https://www.owasp.org/index.php/XSS_(Cross_Site_Scripting)_Prevention_Cheat_Sheet) to help explain the risks, but essentially:
+
+1. Avoid anything other than the standard HTML elements
+1. Avoid `script`, `style`, `iframe`, `object`, `embed`, `base` elements that allow code to be executed by the client or third party content to be included that can execute code
+1. Avoid anything other than plain HTML attributes with values matched to a regexp
+
+Basically, you should be able to describe what HTML is fine for your scenario. If you do not have confidence that you can describe your policy please consider using one of the shipped policies such as `bluemonday.UGCPolicy()`.
+
+To create a new policy:
+```go
+p := bluemonday.NewPolicy()
+```
+
+To add elements to a policy either add just the elements:
+```go
+p.AllowElements("b", "strong")
+```
+
+Or add elements as a virtue of adding an attribute:
+```go
+// Not the recommended pattern, see the recommendation on using .Matching() below
+p.AllowAttrs("nowrap").OnElements("td", "th")
+```
+
+Attributes can either be added to all elements:
+```go
+p.AllowAttrs("dir").Matching(regexp.MustCompile("(?i)rtl|ltr")).Globally()
+```
+
+Or attributes can be added to specific elements:
+```go
+// Not the recommended pattern, see the recommendation on using .Matching() below
+p.AllowAttrs("value").OnElements("li")
+```
+
+It is **always** recommended that an attribute be made to match a pattern. XSS in HTML attributes is very easy otherwise:
+```go
+// \p{L} matches unicode letters, \p{N} matches unicode numbers
+p.AllowAttrs("title").Matching(regexp.MustCompile(`[\p{L}\p{N}\s\-_',:\[\]!\./\\\(\)&]*`)).Globally()
+```
+
+You can stop at any time and call .Sanitize():
+```go
+// string htmlIn passed in from a HTTP POST
+htmlOut := p.Sanitize(htmlIn)
+```
+
+And you can take any existing policy and extend it:
+```go
+p := bluemonday.UGCPolicy()
+p.AllowElements("fieldset", "select", "option")
+```
+
+### Links
+
+Links are difficult beasts to sanitise safely and also one of the biggest attack vectors for malicious content.
+
+It is possible to do this:
+```go
+p.AllowAttrs("href").Matching(regexp.MustCompile(`(?i)mailto|https?`)).OnElements("a")
+```
+
+But that will not protect you as the regular expression is insufficient in this case to have prevented a malformed value doing something unexpected.
+
+We provide some additional global options for safely working with links.
+
+`RequireParseableURLs` will ensure that URLs are parseable by Go's `net/url` package:
+```go
+p.RequireParseableURLs(true)
+```
+
+If you have enabled parseable URLs then the following option will `AllowRelativeURLs`. By default this is disabled (bluemonday is a whitelist tool... you need to explicitly tell us to permit things) and when disabled it will prevent all local and scheme relative URLs (i.e. `href="localpage.html"`, `href="../home.html"` and even `href="//www.google.com"` are relative):
+```go
+p.AllowRelativeURLs(true)
+```
+
+If you have enabled parseable URLs then you can whitelist the schemes (commonly called protocol when thinking of `http` and `https`) that are permitted. Bear in mind that allowing relative URLs in the above option will allow for a blank scheme:
+```go
+p.AllowURLSchemes("mailto", "http", "https")
+```
+
+Regardless of whether you have enabled parseable URLs, you can force all URLs to have a rel="nofollow" attribute. This will be added if it does not exist, but only when the `href` is valid:
+```go
+// This applies to "a" "area" "link" elements that have a "href" attribute
+p.RequireNoFollowOnLinks(true)
+```
+
+We provide a convenience method that applies all of the above, but you will still need to whitelist the linkable elements for the URL rules to be applied to:
+```go
+p.AllowStandardURLs()
+p.AllowAttrs("cite").OnElements("blockquote", "q")
+p.AllowAttrs("href").OnElements("a", "area")
+p.AllowAttrs("src").OnElements("img")
+```
+
+An additional complexity regarding links is the data URI as defined in [RFC2397](http://tools.ietf.org/html/rfc2397). The data URI allows for images to be served inline using this format:
+
+```html
+ Hello World
+```
+
+We have provided a helper to verify the mimetype followed by base64 content of data URIs links:
+
+```go
+p.AllowDataURIImages()
+```
+
+That helper will enable GIF, JPEG, PNG and WEBP images.
+
+It should be noted that there is a potential [security](http://palizine.plynt.com/issues/2010Oct/bypass-xss-filters/) [risk](https://capec.mitre.org/data/definitions/244.html) with the use of data URI links. You should only enable data URI links if you already trust the content.
+
+We also have some features to help deal with user generated content:
+```go
+p.AddTargetBlankToFullyQualifiedLinks(true)
+```
+
+This will ensure that anchor `` links that are fully qualified (the href destination includes a host name) will get `target="_blank"` added to them.
+
+Additionally any link that has `target="_blank"` after the policy has been applied will also have the `rel` attribute adjusted to add `noopener`. This means a link may start like `` and will end up as ``. It is important to note that the addition of `noopener` is a security feature and not an issue. There is an unfortunate feature to browsers that a browser window opened as a result of `target="_blank"` can still control the opener (your web page) and this protects against that. The background to this can be found here: [https://dev.to/ben/the-targetblank-vulnerability-by-example](https://dev.to/ben/the-targetblank-vulnerability-by-example)
+
+### Policy Building Helpers
+
+We also bundle some helpers to simplify policy building:
+```go
+
+// Permits the "dir", "id", "lang", "title" attributes globally
+p.AllowStandardAttributes()
+
+// Permits the "img" element and its standard attributes
+p.AllowImages()
+
+// Permits ordered and unordered lists, and also definition lists
+p.AllowLists()
+
+// Permits HTML tables and all applicable elements and non-styling attributes
+p.AllowTables()
+```
+
+### Invalid Instructions
+
+The following are invalid:
+```go
+// This does not say where the attributes are allowed, you need to add
+// .Globally() or .OnElements(...)
+// This will be ignored without error.
+p.AllowAttrs("value")
+
+// This does not say where the attributes are allowed, you need to add
+// .Globally() or .OnElements(...)
+// This will be ignored without error.
+p.AllowAttrs(
+ "type",
+).Matching(
+ regexp.MustCompile("(?i)^(circle|disc|square|a|A|i|I|1)$"),
+)
+```
+
+Both examples exhibit the same issue, they declare attributes but do not then specify whether they are whitelisted globally or only on specific elements (and which elements). Attributes belong to one or more elements, and the policy needs to declare this.
+
+## Limitations
+
+We are not yet including any tools to help whitelist and sanitize CSS. Which means that unless you wish to do the heavy lifting in a single regular expression (inadvisable), **you should not allow the "style" attribute anywhere**.
+
+It is not the job of bluemonday to fix your bad HTML, it is merely the job of bluemonday to prevent malicious HTML getting through. If you have mismatched HTML elements, or non-conforming nesting of elements, those will remain. But if you have well-structured HTML bluemonday will not break it.
+
+## TODO
+
+* Add support for CSS sanitisation to allow some CSS properties based on a whitelist, possibly using the [Gorilla CSS3 scanner](http://www.gorillatoolkit.org/pkg/css/scanner) - PRs welcome so long as testing covers XSS and demonstrates safety first
+* Investigate whether devs want to blacklist elements and attributes. This would allow devs to take an existing policy (such as the `bluemonday.UGCPolicy()` ) that encapsulates 90% of what they're looking for but does more than they need, and to remove the extra things they do not want to make it 100% what they want
+* Investigate whether devs want a validating HTML mode, in which the HTML elements are not just transformed into a balanced tree (every start tag has a closing tag at the correct depth) but also that elements and character data appear only in their allowed context (i.e. that a `table` element isn't a descendent of a `caption`, that `colgroup`, `thead`, `tbody`, `tfoot` and `tr` are permitted, and that character data is not permitted)
+
+## Development
+
+If you have cloned this repo you will probably need the dependency:
+
+`go get golang.org/x/net/html`
+
+Gophers can use their familiar tools:
+
+`go build`
+
+`go test`
+
+I personally use a Makefile as it spares typing the same args over and over whilst providing consistency for those of us who jump from language to language and enjoy just typing `make` in a project directory and watch magic happen.
+
+`make` will build, vet, test and install the library.
+
+`make clean` will remove the library from a *single* `${GOPATH}/pkg` directory tree
+
+`make test` will run the tests
+
+`make cover` will run the tests and *open a browser window* with the coverage report
+
+`make lint` will run golint (install via `go get github.com/golang/lint/golint`)
+
+## Long term goals
+
+1. Open the code to adversarial peer review similar to the [Attack Review Ground Rules](https://code.google.com/p/owasp-java-html-sanitizer/wiki/AttackReviewGroundRules)
+1. Raise funds and pay for an external security review
diff --git a/vendor/github.com/microcosm-cc/bluemonday/doc.go b/vendor/github.com/microcosm-cc/bluemonday/doc.go
new file mode 100644
index 000000000..71dab6089
--- /dev/null
+++ b/vendor/github.com/microcosm-cc/bluemonday/doc.go
@@ -0,0 +1,104 @@
+// Copyright (c) 2014, David Kitchen
+
+
+To pass through mostly unaltered (it gained a rel="nofollow"):
+
+
+
+
+
+The primary purpose of bluemonday is to take potentially unsafe user generated
+content (from things like Markdown, HTML WYSIWYG tools, etc) and make it safe
+for you to put on your website.
+
+It protects sites against XSS (http://en.wikipedia.org/wiki/Cross-site_scripting)
+and other malicious content that a user interface may deliver. There are many
+vectors for an XSS attack (https://www.owasp.org/index.php/XSS_Filter_Evasion_Cheat_Sheet)
+and the safest thing to do is to sanitize user input against a known safe list
+of HTML elements and attributes.
+
+Note: You should always run bluemonday after any other processing.
+
+If you use blackfriday (https://github.com/russross/blackfriday) or
+Pandoc (http://johnmacfarlane.net/pandoc/) then bluemonday should be run after
+these steps. This ensures that no insecure HTML is introduced later in your
+process.
+
+bluemonday is heavily inspired by both the OWASP Java HTML Sanitizer
+(https://code.google.com/p/owasp-java-html-sanitizer/) and the HTML Purifier
+(http://htmlpurifier.org/).
+
+We ship two default policies, one is bluemonday.StrictPolicy() and can be
+thought of as equivalent to stripping all HTML elements and their attributes as
+it has nothing on its whitelist.
+
+The other is bluemonday.UGCPolicy() and allows a broad selection of HTML
+elements and attributes that are safe for user generated content. Note that
+this policy does not whitelist iframes, object, embed, styles, script, etc.
+
+The essence of building a policy is to determine which HTML elements and
+attributes are considered safe for your scenario. OWASP provide an XSS
+prevention cheat sheet ( https://www.google.com/search?q=xss+prevention+cheat+sheet )
+to help explain the risks, but essentially:
+
+ 1. Avoid whitelisting anything other than plain HTML elements
+ 2. Avoid whitelisting `script`, `style`, `iframe`, `object`, `embed`, `base`
+ elements
+ 3. Avoid whitelisting anything other than plain HTML elements with simple
+ values that you can match to a regexp
+*/
+package bluemonday
diff --git a/vendor/github.com/microcosm-cc/bluemonday/go.mod b/vendor/github.com/microcosm-cc/bluemonday/go.mod
new file mode 100644
index 000000000..fa8453c5f
--- /dev/null
+++ b/vendor/github.com/microcosm-cc/bluemonday/go.mod
@@ -0,0 +1,5 @@
+module github.com/microcosm-cc/bluemonday
+
+go 1.9
+
+require golang.org/x/net v0.0.0-20181220203305-927f97764cc3
diff --git a/vendor/github.com/microcosm-cc/bluemonday/go.sum b/vendor/github.com/microcosm-cc/bluemonday/go.sum
new file mode 100644
index 000000000..bee241d1c
--- /dev/null
+++ b/vendor/github.com/microcosm-cc/bluemonday/go.sum
@@ -0,0 +1,2 @@
+golang.org/x/net v0.0.0-20181220203305-927f97764cc3 h1:eH6Eip3UpmR+yM/qI9Ijluzb1bNv/cAU/n+6l8tRSis=
+golang.org/x/net v0.0.0-20181220203305-927f97764cc3/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
diff --git a/vendor/github.com/microcosm-cc/bluemonday/helpers.go b/vendor/github.com/microcosm-cc/bluemonday/helpers.go
new file mode 100644
index 000000000..dfa5868da
--- /dev/null
+++ b/vendor/github.com/microcosm-cc/bluemonday/helpers.go
@@ -0,0 +1,297 @@
+// Copyright (c) 2014, David Kitchen . To prevent those being removed we maintain
+ // a list of elements that are allowed to have no attributes and that will
+ // be maintained in the output HTML.
+ setOfElementsAllowedWithoutAttrs map[string]struct{}
+
+ setOfElementsToSkipContent map[string]struct{}
+}
+
+type attrPolicy struct {
+
+ // optional pattern to match, when not nil the regexp needs to match
+ // otherwise the attribute is removed
+ regexp *regexp.Regexp
+}
+
+type attrPolicyBuilder struct {
+ p *Policy
+
+ attrNames []string
+ regexp *regexp.Regexp
+ allowEmpty bool
+}
+
+type urlPolicy func(url *url.URL) (allowUrl bool)
+
+// init initializes the maps if this has not been done already
+func (p *Policy) init() {
+ if !p.initialized {
+ p.elsAndAttrs = make(map[string]map[string]attrPolicy)
+ p.globalAttrs = make(map[string]attrPolicy)
+ p.allowURLSchemes = make(map[string]urlPolicy)
+ p.setOfElementsAllowedWithoutAttrs = make(map[string]struct{})
+ p.setOfElementsToSkipContent = make(map[string]struct{})
+ p.initialized = true
+ }
+}
+
+// NewPolicy returns a blank policy with nothing whitelisted or permitted. This
+// is the recommended way to start building a policy and you should now use
+// AllowAttrs() and/or AllowElements() to construct the whitelist of HTML
+// elements and attributes.
+func NewPolicy() *Policy {
+
+ p := Policy{}
+
+ p.addDefaultElementsWithoutAttrs()
+ p.addDefaultSkipElementContent()
+
+ return &p
+}
+
+// AllowAttrs takes a range of HTML attribute names and returns an
+// attribute policy builder that allows you to specify the pattern and scope of
+// the whitelisted attribute.
+//
+// The attribute policy is only added to the core policy when either Globally()
+// or OnElements(...) are called.
+func (p *Policy) AllowAttrs(attrNames ...string) *attrPolicyBuilder {
+
+ p.init()
+
+ abp := attrPolicyBuilder{
+ p: p,
+ allowEmpty: false,
+ }
+
+ for _, attrName := range attrNames {
+ abp.attrNames = append(abp.attrNames, strings.ToLower(attrName))
+ }
+
+ return &abp
+}
+
+// AllowDataAttributes whitelists all data attributes. We can't specify the name
+// of each attribute exactly as they are customized.
+//
+// NOTE: These values are not sanitized and applications that evaluate or process
+// them without checking and verification of the input may be at risk if this option
+// is enabled. This is a 'caveat emptor' option and the person enabling this option
+// needs to fully understand the potential impact with regards to whatever application
+// will be consuming the sanitized HTML afterwards, i.e. if you know you put a link in a
+// data attribute and use that to automatically load some new window then you're giving
+// the author of a HTML fragment the means to open a malicious destination automatically.
+// Use with care!
+func (p *Policy) AllowDataAttributes() {
+ p.allowDataAttributes = true
+}
+
+// AllowNoAttrs says that attributes on element are optional.
+//
+// The attribute policy is only added to the core policy when OnElements(...)
+// are called.
+func (p *Policy) AllowNoAttrs() *attrPolicyBuilder {
+
+ p.init()
+
+ abp := attrPolicyBuilder{
+ p: p,
+ allowEmpty: true,
+ }
+ return &abp
+}
+
+// AllowNoAttrs says that attributes on element are optional.
+//
+// The attribute policy is only added to the core policy when OnElements(...)
+// are called.
+func (abp *attrPolicyBuilder) AllowNoAttrs() *attrPolicyBuilder {
+
+ abp.allowEmpty = true
+
+ return abp
+}
+
+// Matching allows a regular expression to be applied to a nascent attribute
+// policy, and returns the attribute policy. Calling this more than once will
+// replace the existing regexp.
+func (abp *attrPolicyBuilder) Matching(regex *regexp.Regexp) *attrPolicyBuilder {
+
+ abp.regexp = regex
+
+ return abp
+}
+
+// OnElements will bind an attribute policy to a given range of HTML elements
+// and return the updated policy
+func (abp *attrPolicyBuilder) OnElements(elements ...string) *Policy {
+
+ for _, element := range elements {
+ element = strings.ToLower(element)
+
+ for _, attr := range abp.attrNames {
+
+ if _, ok := abp.p.elsAndAttrs[element]; !ok {
+ abp.p.elsAndAttrs[element] = make(map[string]attrPolicy)
+ }
+
+ ap := attrPolicy{}
+ if abp.regexp != nil {
+ ap.regexp = abp.regexp
+ }
+
+ abp.p.elsAndAttrs[element][attr] = ap
+ }
+
+ if abp.allowEmpty {
+ abp.p.setOfElementsAllowedWithoutAttrs[element] = struct{}{}
+
+ if _, ok := abp.p.elsAndAttrs[element]; !ok {
+ abp.p.elsAndAttrs[element] = make(map[string]attrPolicy)
+ }
+ }
+ }
+
+ return abp.p
+}
+
+// Globally will bind an attribute policy to all HTML elements and return the
+// updated policy
+func (abp *attrPolicyBuilder) Globally() *Policy {
+
+ for _, attr := range abp.attrNames {
+ if _, ok := abp.p.globalAttrs[attr]; !ok {
+ abp.p.globalAttrs[attr] = attrPolicy{}
+ }
+
+ ap := attrPolicy{}
+ if abp.regexp != nil {
+ ap.regexp = abp.regexp
+ }
+
+ abp.p.globalAttrs[attr] = ap
+ }
+
+ return abp.p
+}
+
+// AllowElements will append HTML elements to the whitelist without applying an
+// attribute policy to those elements (the elements are permitted
+// sans-attributes)
+func (p *Policy) AllowElements(names ...string) *Policy {
+ p.init()
+
+ for _, element := range names {
+ element = strings.ToLower(element)
+
+ if _, ok := p.elsAndAttrs[element]; !ok {
+ p.elsAndAttrs[element] = make(map[string]attrPolicy)
+ }
+ }
+
+ return p
+}
+
+// RequireNoFollowOnLinks will result in all tags having a rel="nofollow"
+// added to them if one does not already exist
+//
+// Note: This requires p.RequireParseableURLs(true) and will enable it.
+func (p *Policy) RequireNoFollowOnLinks(require bool) *Policy {
+
+ p.requireNoFollow = require
+ p.requireParseableURLs = true
+
+ return p
+}
+
+// RequireNoFollowOnFullyQualifiedLinks will result in all tags that point
+// to a non-local destination (i.e. starts with a protocol and has a host)
+// having a rel="nofollow" added to them if one does not already exist
+//
+// Note: This requires p.RequireParseableURLs(true) and will enable it.
+func (p *Policy) RequireNoFollowOnFullyQualifiedLinks(require bool) *Policy {
+
+ p.requireNoFollowFullyQualifiedLinks = require
+ p.requireParseableURLs = true
+
+ return p
+}
+
+// AddTargetBlankToFullyQualifiedLinks will result in all tags that point
+// to a non-local destination (i.e. starts with a protocol and has a host)
+// having a target="_blank" added to them if one does not already exist
+//
+// Note: This requires p.RequireParseableURLs(true) and will enable it.
+func (p *Policy) AddTargetBlankToFullyQualifiedLinks(require bool) *Policy {
+
+ p.addTargetBlankToFullyQualifiedLinks = require
+ p.requireParseableURLs = true
+
+ return p
+}
+
+// RequireParseableURLs will result in all URLs requiring that they be parseable
+// by "net/url" url.Parse()
+// This applies to:
+// - a.href
+// - area.href
+// - blockquote.cite
+// - img.src
+// - link.href
+// - script.src
+func (p *Policy) RequireParseableURLs(require bool) *Policy {
+
+ p.requireParseableURLs = require
+
+ return p
+}
+
+// AllowRelativeURLs enables RequireParseableURLs and then permits URLs that
+// are parseable, have no schema information and url.IsAbs() returns false
+// This permits local URLs
+func (p *Policy) AllowRelativeURLs(require bool) *Policy {
+
+ p.RequireParseableURLs(true)
+ p.allowRelativeURLs = require
+
+ return p
+}
+
+// AllowURLSchemes will append URL schemes to the whitelist
+// Example: p.AllowURLSchemes("mailto", "http", "https")
+func (p *Policy) AllowURLSchemes(schemes ...string) *Policy {
+ p.init()
+
+ p.RequireParseableURLs(true)
+
+ for _, scheme := range schemes {
+ scheme = strings.ToLower(scheme)
+
+ // Allow all URLs with matching scheme.
+ p.allowURLSchemes[scheme] = nil
+ }
+
+ return p
+}
+
+// AllowURLSchemeWithCustomPolicy will append URL schemes with
+// a custom URL policy to the whitelist.
+// Only the URLs with matching schema and urlPolicy(url)
+// returning true will be allowed.
+func (p *Policy) AllowURLSchemeWithCustomPolicy(
+ scheme string,
+ urlPolicy func(url *url.URL) (allowUrl bool),
+) *Policy {
+
+ p.init()
+
+ p.RequireParseableURLs(true)
+
+ scheme = strings.ToLower(scheme)
+
+ p.allowURLSchemes[scheme] = urlPolicy
+
+ return p
+}
+
+// AddSpaceWhenStrippingTag states whether to add a single space " " when
+// removing tags that are not whitelisted by the policy.
+//
+// This is useful if you expect to strip tags in dense markup and may lose the
+// value of whitespace.
+//
+// For example: "
is valid, but isn't valid as the "dir" attr
+// is mandatory
+func (p *Policy) addDefaultElementsWithoutAttrs() {
+ p.init()
+
+ p.setOfElementsAllowedWithoutAttrs["abbr"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["acronym"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["address"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["article"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["aside"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["audio"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["b"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["bdi"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["blockquote"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["body"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["br"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["button"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["canvas"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["caption"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["center"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["cite"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["code"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["col"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["colgroup"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["datalist"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["dd"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["del"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["details"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["dfn"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["div"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["dl"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["dt"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["em"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["fieldset"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["figcaption"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["figure"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["footer"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["h1"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["h2"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["h3"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["h4"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["h5"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["h6"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["head"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["header"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["hgroup"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["hr"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["html"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["i"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["ins"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["kbd"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["li"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["mark"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["marquee"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["nav"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["ol"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["optgroup"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["option"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["p"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["pre"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["q"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["rp"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["rt"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["ruby"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["s"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["samp"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["script"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["section"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["select"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["small"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["span"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["strike"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["strong"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["style"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["sub"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["summary"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["sup"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["svg"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["table"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["tbody"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["td"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["textarea"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["tfoot"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["th"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["thead"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["title"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["time"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["tr"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["tt"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["u"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["ul"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["var"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["video"] = struct{}{}
+ p.setOfElementsAllowedWithoutAttrs["wbr"] = struct{}{}
+
+}
+
+// addDefaultSkipElementContent adds the HTML elements that we should skip
+// rendering the character content of, if the element itself is not allowed.
+// This is all character data that the end user would not normally see.
+// i.e. if we exclude a tag.
+func (p *Policy) addDefaultSkipElementContent() {
+ p.init()
+
+ p.setOfElementsToSkipContent["frame"] = struct{}{}
+ p.setOfElementsToSkipContent["frameset"] = struct{}{}
+ p.setOfElementsToSkipContent["iframe"] = struct{}{}
+ p.setOfElementsToSkipContent["noembed"] = struct{}{}
+ p.setOfElementsToSkipContent["noframes"] = struct{}{}
+ p.setOfElementsToSkipContent["noscript"] = struct{}{}
+ p.setOfElementsToSkipContent["nostyle"] = struct{}{}
+ p.setOfElementsToSkipContent["object"] = struct{}{}
+ p.setOfElementsToSkipContent["script"] = struct{}{}
+ p.setOfElementsToSkipContent["style"] = struct{}{}
+ p.setOfElementsToSkipContent["title"] = struct{}{}
+}
diff --git a/vendor/github.com/microcosm-cc/bluemonday/sanitize.go b/vendor/github.com/microcosm-cc/bluemonday/sanitize.go
new file mode 100644
index 000000000..65ed89b79
--- /dev/null
+++ b/vendor/github.com/microcosm-cc/bluemonday/sanitize.go
@@ -0,0 +1,581 @@
+// Copyright (c) 2014, David Kitchen
are treated as start tags, except that
+ // hasSelfClosingToken is set while they are being processed.
+ hasSelfClosingToken bool
+ // doc is the document root element.
+ doc *Node
+ // The stack of open elements (section 12.2.3.2) and active formatting
+ // elements (section 12.2.3.3).
+ oe, afe nodeStack
+ // Element pointers (section 12.2.3.4).
+ head, form *Node
+ // Other parsing state flags (section 12.2.3.5).
+ scripting, framesetOK bool
+ // im is the current insertion mode.
+ im insertionMode
+ // originalIM is the insertion mode to go back to after completing a text
+ // or inTableText insertion mode.
+ originalIM insertionMode
+ // fosterParenting is whether new elements should be inserted according to
+ // the foster parenting rules (section 12.2.5.3).
+ fosterParenting bool
+ // quirks is whether the parser is operating in "quirks mode."
+ quirks bool
+ // fragment is whether the parser is parsing an HTML fragment.
+ fragment bool
+ // context is the context element when parsing an HTML fragment
+ // (section 12.4).
+ context *Node
+}
+
+func (p *parser) top() *Node {
+ if n := p.oe.top(); n != nil {
+ return n
+ }
+ return p.doc
+}
+
+// Stop tags for use in popUntil. These come from section 12.2.3.2.
+var (
+ defaultScopeStopTags = map[string][]a.Atom{
+ "": {a.Applet, a.Caption, a.Html, a.Table, a.Td, a.Th, a.Marquee, a.Object, a.Template},
+ "math": {a.AnnotationXml, a.Mi, a.Mn, a.Mo, a.Ms, a.Mtext},
+ "svg": {a.Desc, a.ForeignObject, a.Title},
+ }
+)
+
+type scope int
+
+const (
+ defaultScope scope = iota
+ listItemScope
+ buttonScope
+ tableScope
+ tableRowScope
+ tableBodyScope
+ selectScope
+)
+
+// popUntil pops the stack of open elements at the highest element whose tag
+// is in matchTags, provided there is no higher element in the scope's stop
+// tags (as defined in section 12.2.3.2). It returns whether or not there was
+// such an element. If there was not, popUntil leaves the stack unchanged.
+//
+// For example, the set of stop tags for table scope is: "html", "table". If
+// the stack was:
+// ["html", "body", "font", "table", "b", "i", "u"]
+// then popUntil(tableScope, "font") would return false, but
+// popUntil(tableScope, "i") would return true and the stack would become:
+// ["html", "body", "font", "table", "b"]
+//
+// If an element's tag is in both the stop tags and matchTags, then the stack
+// will be popped and the function returns true (provided, of course, there was
+// no higher element in the stack that was also in the stop tags). For example,
+// popUntil(tableScope, "table") returns true and leaves:
+// ["html", "body", "font"]
+func (p *parser) popUntil(s scope, matchTags ...a.Atom) bool {
+ if i := p.indexOfElementInScope(s, matchTags...); i != -1 {
+ p.oe = p.oe[:i]
+ return true
+ }
+ return false
+}
+
+// indexOfElementInScope returns the index in p.oe of the highest element whose
+// tag is in matchTags that is in scope. If no matching element is in scope, it
+// returns -1.
+func (p *parser) indexOfElementInScope(s scope, matchTags ...a.Atom) int {
+ for i := len(p.oe) - 1; i >= 0; i-- {
+ tagAtom := p.oe[i].DataAtom
+ if p.oe[i].Namespace == "" {
+ for _, t := range matchTags {
+ if t == tagAtom {
+ return i
+ }
+ }
+ switch s {
+ case defaultScope:
+ // No-op.
+ case listItemScope:
+ if tagAtom == a.Ol || tagAtom == a.Ul {
+ return -1
+ }
+ case buttonScope:
+ if tagAtom == a.Button {
+ return -1
+ }
+ case tableScope:
+ if tagAtom == a.Html || tagAtom == a.Table {
+ return -1
+ }
+ case selectScope:
+ if tagAtom != a.Optgroup && tagAtom != a.Option {
+ return -1
+ }
+ default:
+ panic("unreachable")
+ }
+ }
+ switch s {
+ case defaultScope, listItemScope, buttonScope:
+ for _, t := range defaultScopeStopTags[p.oe[i].Namespace] {
+ if t == tagAtom {
+ return -1
+ }
+ }
+ }
+ }
+ return -1
+}
+
+// elementInScope is like popUntil, except that it doesn't modify the stack of
+// open elements.
+func (p *parser) elementInScope(s scope, matchTags ...a.Atom) bool {
+ return p.indexOfElementInScope(s, matchTags...) != -1
+}
+
+// clearStackToContext pops elements off the stack of open elements until a
+// scope-defined element is found.
+func (p *parser) clearStackToContext(s scope) {
+ for i := len(p.oe) - 1; i >= 0; i-- {
+ tagAtom := p.oe[i].DataAtom
+ switch s {
+ case tableScope:
+ if tagAtom == a.Html || tagAtom == a.Table {
+ p.oe = p.oe[:i+1]
+ return
+ }
+ case tableRowScope:
+ if tagAtom == a.Html || tagAtom == a.Tr {
+ p.oe = p.oe[:i+1]
+ return
+ }
+ case tableBodyScope:
+ if tagAtom == a.Html || tagAtom == a.Tbody || tagAtom == a.Tfoot || tagAtom == a.Thead {
+ p.oe = p.oe[:i+1]
+ return
+ }
+ default:
+ panic("unreachable")
+ }
+ }
+}
+
+// generateImpliedEndTags pops nodes off the stack of open elements as long as
+// the top node has a tag name of dd, dt, li, option, optgroup, p, rp, or rt.
+// If exceptions are specified, nodes with that name will not be popped off.
+func (p *parser) generateImpliedEndTags(exceptions ...string) {
+ var i int
+loop:
+ for i = len(p.oe) - 1; i >= 0; i-- {
+ n := p.oe[i]
+ if n.Type == ElementNode {
+ switch n.DataAtom {
+ case a.Dd, a.Dt, a.Li, a.Option, a.Optgroup, a.P, a.Rp, a.Rt:
+ for _, except := range exceptions {
+ if n.Data == except {
+ break loop
+ }
+ }
+ continue
+ }
+ }
+ break
+ }
+
+ p.oe = p.oe[:i+1]
+}
+
+// addChild adds a child node n to the top element, and pushes n onto the stack
+// of open elements if it is an element node.
+func (p *parser) addChild(n *Node) {
+ if p.shouldFosterParent() {
+ p.fosterParent(n)
+ } else {
+ p.top().AppendChild(n)
+ }
+
+ if n.Type == ElementNode {
+ p.oe = append(p.oe, n)
+ }
+}
+
+// shouldFosterParent returns whether the next node to be added should be
+// foster parented.
+func (p *parser) shouldFosterParent() bool {
+ if p.fosterParenting {
+ switch p.top().DataAtom {
+ case a.Table, a.Tbody, a.Tfoot, a.Thead, a.Tr:
+ return true
+ }
+ }
+ return false
+}
+
+// fosterParent adds a child node according to the foster parenting rules.
+// Section 12.2.5.3, "foster parenting".
+func (p *parser) fosterParent(n *Node) {
+ var table, parent, prev *Node
+ var i int
+ for i = len(p.oe) - 1; i >= 0; i-- {
+ if p.oe[i].DataAtom == a.Table {
+ table = p.oe[i]
+ break
+ }
+ }
+
+ if table == nil {
+ // The foster parent is the html element.
+ parent = p.oe[0]
+ } else {
+ parent = table.Parent
+ }
+ if parent == nil {
+ parent = p.oe[i-1]
+ }
+
+ if table != nil {
+ prev = table.PrevSibling
+ } else {
+ prev = parent.LastChild
+ }
+ if prev != nil && prev.Type == TextNode && n.Type == TextNode {
+ prev.Data += n.Data
+ return
+ }
+
+ parent.InsertBefore(n, table)
+}
+
+// addText adds text to the preceding node if it is a text node, or else it
+// calls addChild with a new text node.
+func (p *parser) addText(text string) {
+ if text == "" {
+ return
+ }
+
+ if p.shouldFosterParent() {
+ p.fosterParent(&Node{
+ Type: TextNode,
+ Data: text,
+ })
+ return
+ }
+
+ t := p.top()
+ if n := t.LastChild; n != nil && n.Type == TextNode {
+ n.Data += text
+ return
+ }
+ p.addChild(&Node{
+ Type: TextNode,
+ Data: text,
+ })
+}
+
+// addElement adds a child element based on the current token.
+func (p *parser) addElement() {
+ p.addChild(&Node{
+ Type: ElementNode,
+ DataAtom: p.tok.DataAtom,
+ Data: p.tok.Data,
+ Attr: p.tok.Attr,
+ })
+}
+
+// Section 12.2.3.3.
+func (p *parser) addFormattingElement() {
+ tagAtom, attr := p.tok.DataAtom, p.tok.Attr
+ p.addElement()
+
+ // Implement the Noah's Ark clause, but with three per family instead of two.
+ identicalElements := 0
+findIdenticalElements:
+ for i := len(p.afe) - 1; i >= 0; i-- {
+ n := p.afe[i]
+ if n.Type == scopeMarkerNode {
+ break
+ }
+ if n.Type != ElementNode {
+ continue
+ }
+ if n.Namespace != "" {
+ continue
+ }
+ if n.DataAtom != tagAtom {
+ continue
+ }
+ if len(n.Attr) != len(attr) {
+ continue
+ }
+ compareAttributes:
+ for _, t0 := range n.Attr {
+ for _, t1 := range attr {
+ if t0.Key == t1.Key && t0.Namespace == t1.Namespace && t0.Val == t1.Val {
+ // Found a match for this attribute, continue with the next attribute.
+ continue compareAttributes
+ }
+ }
+ // If we get here, there is no attribute that matches a.
+ // Therefore the element is not identical to the new one.
+ continue findIdenticalElements
+ }
+
+ identicalElements++
+ if identicalElements >= 3 {
+ p.afe.remove(n)
+ }
+ }
+
+ p.afe = append(p.afe, p.top())
+}
+
+// Section 12.2.3.3.
+func (p *parser) clearActiveFormattingElements() {
+ for {
+ n := p.afe.pop()
+ if len(p.afe) == 0 || n.Type == scopeMarkerNode {
+ return
+ }
+ }
+}
+
+// Section 12.2.3.3.
+func (p *parser) reconstructActiveFormattingElements() {
+ n := p.afe.top()
+ if n == nil {
+ return
+ }
+ if n.Type == scopeMarkerNode || p.oe.index(n) != -1 {
+ return
+ }
+ i := len(p.afe) - 1
+ for n.Type != scopeMarkerNode && p.oe.index(n) == -1 {
+ if i == 0 {
+ i = -1
+ break
+ }
+ i--
+ n = p.afe[i]
+ }
+ for {
+ i++
+ clone := p.afe[i].clone()
+ p.addChild(clone)
+ p.afe[i] = clone
+ if i == len(p.afe)-1 {
+ break
+ }
+ }
+}
+
+// Section 12.2.4.
+func (p *parser) acknowledgeSelfClosingTag() {
+ p.hasSelfClosingToken = false
+}
+
+// An insertion mode (section 12.2.3.1) is the state transition function from
+// a particular state in the HTML5 parser's state machine. It updates the
+// parser's fields depending on parser.tok (where ErrorToken means EOF).
+// It returns whether the token was consumed.
+type insertionMode func(*parser) bool
+
+// setOriginalIM sets the insertion mode to return to after completing a text or
+// inTableText insertion mode.
+// Section 12.2.3.1, "using the rules for".
+func (p *parser) setOriginalIM() {
+ if p.originalIM != nil {
+ panic("html: bad parser state: originalIM was set twice")
+ }
+ p.originalIM = p.im
+}
+
+// Section 12.2.3.1, "reset the insertion mode".
+func (p *parser) resetInsertionMode() {
+ for i := len(p.oe) - 1; i >= 0; i-- {
+ n := p.oe[i]
+ if i == 0 && p.context != nil {
+ n = p.context
+ }
+
+ switch n.DataAtom {
+ case a.Select:
+ p.im = inSelectIM
+ case a.Td, a.Th:
+ p.im = inCellIM
+ case a.Tr:
+ p.im = inRowIM
+ case a.Tbody, a.Thead, a.Tfoot:
+ p.im = inTableBodyIM
+ case a.Caption:
+ p.im = inCaptionIM
+ case a.Colgroup:
+ p.im = inColumnGroupIM
+ case a.Table:
+ p.im = inTableIM
+ case a.Head:
+ p.im = inBodyIM
+ case a.Body:
+ p.im = inBodyIM
+ case a.Frameset:
+ p.im = inFramesetIM
+ case a.Html:
+ p.im = beforeHeadIM
+ default:
+ continue
+ }
+ return
+ }
+ p.im = inBodyIM
+}
+
+const whitespace = " \t\r\n\f"
+
+// Section 12.2.5.4.1.
+func initialIM(p *parser) bool {
+ switch p.tok.Type {
+ case TextToken:
+ p.tok.Data = strings.TrimLeft(p.tok.Data, whitespace)
+ if len(p.tok.Data) == 0 {
+ // It was all whitespace, so ignore it.
+ return true
+ }
+ case CommentToken:
+ p.doc.AppendChild(&Node{
+ Type: CommentNode,
+ Data: p.tok.Data,
+ })
+ return true
+ case DoctypeToken:
+ n, quirks := parseDoctype(p.tok.Data)
+ p.doc.AppendChild(n)
+ p.quirks = quirks
+ p.im = beforeHTMLIM
+ return true
+ }
+ p.quirks = true
+ p.im = beforeHTMLIM
+ return false
+}
+
+// Section 12.2.5.4.2.
+func beforeHTMLIM(p *parser) bool {
+ switch p.tok.Type {
+ case DoctypeToken:
+ // Ignore the token.
+ return true
+ case TextToken:
+ p.tok.Data = strings.TrimLeft(p.tok.Data, whitespace)
+ if len(p.tok.Data) == 0 {
+ // It was all whitespace, so ignore it.
+ return true
+ }
+ case StartTagToken:
+ if p.tok.DataAtom == a.Html {
+ p.addElement()
+ p.im = beforeHeadIM
+ return true
+ }
+ case EndTagToken:
+ switch p.tok.DataAtom {
+ case a.Head, a.Body, a.Html, a.Br:
+ p.parseImpliedToken(StartTagToken, a.Html, a.Html.String())
+ return false
+ default:
+ // Ignore the token.
+ return true
+ }
+ case CommentToken:
+ p.doc.AppendChild(&Node{
+ Type: CommentNode,
+ Data: p.tok.Data,
+ })
+ return true
+ }
+ p.parseImpliedToken(StartTagToken, a.Html, a.Html.String())
+ return false
+}
+
+// Section 12.2.5.4.3.
+func beforeHeadIM(p *parser) bool {
+ switch p.tok.Type {
+ case TextToken:
+ p.tok.Data = strings.TrimLeft(p.tok.Data, whitespace)
+ if len(p.tok.Data) == 0 {
+ // It was all whitespace, so ignore it.
+ return true
+ }
+ case StartTagToken:
+ switch p.tok.DataAtom {
+ case a.Head:
+ p.addElement()
+ p.head = p.top()
+ p.im = inHeadIM
+ return true
+ case a.Html:
+ return inBodyIM(p)
+ }
+ case EndTagToken:
+ switch p.tok.DataAtom {
+ case a.Head, a.Body, a.Html, a.Br:
+ p.parseImpliedToken(StartTagToken, a.Head, a.Head.String())
+ return false
+ default:
+ // Ignore the token.
+ return true
+ }
+ case CommentToken:
+ p.addChild(&Node{
+ Type: CommentNode,
+ Data: p.tok.Data,
+ })
+ return true
+ case DoctypeToken:
+ // Ignore the token.
+ return true
+ }
+
+ p.parseImpliedToken(StartTagToken, a.Head, a.Head.String())
+ return false
+}
+
+// Section 12.2.5.4.4.
+func inHeadIM(p *parser) bool {
+ switch p.tok.Type {
+ case TextToken:
+ s := strings.TrimLeft(p.tok.Data, whitespace)
+ if len(s) < len(p.tok.Data) {
+ // Add the initial whitespace to the current node.
+ p.addText(p.tok.Data[:len(p.tok.Data)-len(s)])
+ if s == "" {
+ return true
+ }
+ p.tok.Data = s
+ }
+ case StartTagToken:
+ switch p.tok.DataAtom {
+ case a.Html:
+ return inBodyIM(p)
+ case a.Base, a.Basefont, a.Bgsound, a.Command, a.Link, a.Meta:
+ p.addElement()
+ p.oe.pop()
+ p.acknowledgeSelfClosingTag()
+ return true
+ case a.Script, a.Title, a.Noscript, a.Noframes, a.Style:
+ p.addElement()
+ p.setOriginalIM()
+ p.im = textIM
+ return true
+ case a.Head:
+ // Ignore the token.
+ return true
+ }
+ case EndTagToken:
+ switch p.tok.DataAtom {
+ case a.Head:
+ n := p.oe.pop()
+ if n.DataAtom != a.Head {
+ panic("html: bad parser state: element not found, in the in-head insertion mode")
+ }
+ p.im = afterHeadIM
+ return true
+ case a.Body, a.Html, a.Br:
+ p.parseImpliedToken(EndTagToken, a.Head, a.Head.String())
+ return false
+ default:
+ // Ignore the token.
+ return true
+ }
+ case CommentToken:
+ p.addChild(&Node{
+ Type: CommentNode,
+ Data: p.tok.Data,
+ })
+ return true
+ case DoctypeToken:
+ // Ignore the token.
+ return true
+ }
+
+ p.parseImpliedToken(EndTagToken, a.Head, a.Head.String())
+ return false
+}
+
+// Section 12.2.5.4.6.
+func afterHeadIM(p *parser) bool {
+ switch p.tok.Type {
+ case TextToken:
+ s := strings.TrimLeft(p.tok.Data, whitespace)
+ if len(s) < len(p.tok.Data) {
+ // Add the initial whitespace to the current node.
+ p.addText(p.tok.Data[:len(p.tok.Data)-len(s)])
+ if s == "" {
+ return true
+ }
+ p.tok.Data = s
+ }
+ case StartTagToken:
+ switch p.tok.DataAtom {
+ case a.Html:
+ return inBodyIM(p)
+ case a.Body:
+ p.addElement()
+ p.framesetOK = false
+ p.im = inBodyIM
+ return true
+ case a.Frameset:
+ p.addElement()
+ p.im = inFramesetIM
+ return true
+ case a.Base, a.Basefont, a.Bgsound, a.Link, a.Meta, a.Noframes, a.Script, a.Style, a.Title:
+ p.oe = append(p.oe, p.head)
+ defer p.oe.remove(p.head)
+ return inHeadIM(p)
+ case a.Head:
+ // Ignore the token.
+ return true
+ }
+ case EndTagToken:
+ switch p.tok.DataAtom {
+ case a.Body, a.Html, a.Br:
+ // Drop down to creating an implied tag.
+ default:
+ // Ignore the token.
+ return true
+ }
+ case CommentToken:
+ p.addChild(&Node{
+ Type: CommentNode,
+ Data: p.tok.Data,
+ })
+ return true
+ case DoctypeToken:
+ // Ignore the token.
+ return true
+ }
+
+ p.parseImpliedToken(StartTagToken, a.Body, a.Body.String())
+ p.framesetOK = true
+ return false
+}
+
+// copyAttributes copies attributes of src not found on dst to dst.
+func copyAttributes(dst *Node, src Token) {
+ if len(src.Attr) == 0 {
+ return
+ }
+ attr := map[string]string{}
+ for _, t := range dst.Attr {
+ attr[t.Key] = t.Val
+ }
+ for _, t := range src.Attr {
+ if _, ok := attr[t.Key]; !ok {
+ dst.Attr = append(dst.Attr, t)
+ attr[t.Key] = t.Val
+ }
+ }
+}
+
+// Section 12.2.5.4.7.
+func inBodyIM(p *parser) bool {
+ switch p.tok.Type {
+ case TextToken:
+ d := p.tok.Data
+ switch n := p.oe.top(); n.DataAtom {
+ case a.Pre, a.Listing:
+ if n.FirstChild == nil {
+ // Ignore a newline at the start of a block.
+ if d != "" && d[0] == '\r' {
+ d = d[1:]
+ }
+ if d != "" && d[0] == '\n' {
+ d = d[1:]
+ }
+ }
+ }
+ d = strings.Replace(d, "\x00", "", -1)
+ if d == "" {
+ return true
+ }
+ p.reconstructActiveFormattingElements()
+ p.addText(d)
+ if p.framesetOK && strings.TrimLeft(d, whitespace) != "" {
+ // There were non-whitespace characters inserted.
+ p.framesetOK = false
+ }
+ case StartTagToken:
+ switch p.tok.DataAtom {
+ case a.Html:
+ copyAttributes(p.oe[0], p.tok)
+ case a.Base, a.Basefont, a.Bgsound, a.Command, a.Link, a.Meta, a.Noframes, a.Script, a.Style, a.Title:
+ return inHeadIM(p)
+ case a.Body:
+ if len(p.oe) >= 2 {
+ body := p.oe[1]
+ if body.Type == ElementNode && body.DataAtom == a.Body {
+ p.framesetOK = false
+ copyAttributes(body, p.tok)
+ }
+ }
+ case a.Frameset:
+ if !p.framesetOK || len(p.oe) < 2 || p.oe[1].DataAtom != a.Body {
+ // Ignore the token.
+ return true
+ }
+ body := p.oe[1]
+ if body.Parent != nil {
+ body.Parent.RemoveChild(body)
+ }
+ p.oe = p.oe[:1]
+ p.addElement()
+ p.im = inFramesetIM
+ return true
+ case a.Address, a.Article, a.Aside, a.Blockquote, a.Center, a.Details, a.Dir, a.Div, a.Dl, a.Fieldset, a.Figcaption, a.Figure, a.Footer, a.Header, a.Hgroup, a.Menu, a.Nav, a.Ol, a.P, a.Section, a.Summary, a.Ul:
+ p.popUntil(buttonScope, a.P)
+ p.addElement()
+ case a.H1, a.H2, a.H3, a.H4, a.H5, a.H6:
+ p.popUntil(buttonScope, a.P)
+ switch n := p.top(); n.DataAtom {
+ case a.H1, a.H2, a.H3, a.H4, a.H5, a.H6:
+ p.oe.pop()
+ }
+ p.addElement()
+ case a.Pre, a.Listing:
+ p.popUntil(buttonScope, a.P)
+ p.addElement()
+ // The newline, if any, will be dealt with by the TextToken case.
+ p.framesetOK = false
+ case a.Form:
+ if p.form == nil {
+ p.popUntil(buttonScope, a.P)
+ p.addElement()
+ p.form = p.top()
+ }
+ case a.Li:
+ p.framesetOK = false
+ for i := len(p.oe) - 1; i >= 0; i-- {
+ node := p.oe[i]
+ switch node.DataAtom {
+ case a.Li:
+ p.oe = p.oe[:i]
+ case a.Address, a.Div, a.P:
+ continue
+ default:
+ if !isSpecialElement(node) {
+ continue
+ }
+ }
+ break
+ }
+ p.popUntil(buttonScope, a.P)
+ p.addElement()
+ case a.Dd, a.Dt:
+ p.framesetOK = false
+ for i := len(p.oe) - 1; i >= 0; i-- {
+ node := p.oe[i]
+ switch node.DataAtom {
+ case a.Dd, a.Dt:
+ p.oe = p.oe[:i]
+ case a.Address, a.Div, a.P:
+ continue
+ default:
+ if !isSpecialElement(node) {
+ continue
+ }
+ }
+ break
+ }
+ p.popUntil(buttonScope, a.P)
+ p.addElement()
+ case a.Plaintext:
+ p.popUntil(buttonScope, a.P)
+ p.addElement()
+ case a.Button:
+ p.popUntil(defaultScope, a.Button)
+ p.reconstructActiveFormattingElements()
+ p.addElement()
+ p.framesetOK = false
+ case a.A:
+ for i := len(p.afe) - 1; i >= 0 && p.afe[i].Type != scopeMarkerNode; i-- {
+ if n := p.afe[i]; n.Type == ElementNode && n.DataAtom == a.A {
+ p.inBodyEndTagFormatting(a.A)
+ p.oe.remove(n)
+ p.afe.remove(n)
+ break
+ }
+ }
+ p.reconstructActiveFormattingElements()
+ p.addFormattingElement()
+ case a.B, a.Big, a.Code, a.Em, a.Font, a.I, a.S, a.Small, a.Strike, a.Strong, a.Tt, a.U:
+ p.reconstructActiveFormattingElements()
+ p.addFormattingElement()
+ case a.Nobr:
+ p.reconstructActiveFormattingElements()
+ if p.elementInScope(defaultScope, a.Nobr) {
+ p.inBodyEndTagFormatting(a.Nobr)
+ p.reconstructActiveFormattingElements()
+ }
+ p.addFormattingElement()
+ case a.Applet, a.Marquee, a.Object:
+ p.reconstructActiveFormattingElements()
+ p.addElement()
+ p.afe = append(p.afe, &scopeMarker)
+ p.framesetOK = false
+ case a.Table:
+ if !p.quirks {
+ p.popUntil(buttonScope, a.P)
+ }
+ p.addElement()
+ p.framesetOK = false
+ p.im = inTableIM
+ return true
+ case a.Area, a.Br, a.Embed, a.Img, a.Input, a.Keygen, a.Wbr:
+ p.reconstructActiveFormattingElements()
+ p.addElement()
+ p.oe.pop()
+ p.acknowledgeSelfClosingTag()
+ if p.tok.DataAtom == a.Input {
+ for _, t := range p.tok.Attr {
+ if t.Key == "type" {
+ if strings.ToLower(t.Val) == "hidden" {
+ // Skip setting framesetOK = false
+ return true
+ }
+ }
+ }
+ }
+ p.framesetOK = false
+ case a.Param, a.Source, a.Track:
+ p.addElement()
+ p.oe.pop()
+ p.acknowledgeSelfClosingTag()
+ case a.Hr:
+ p.popUntil(buttonScope, a.P)
+ p.addElement()
+ p.oe.pop()
+ p.acknowledgeSelfClosingTag()
+ p.framesetOK = false
+ case a.Image:
+ p.tok.DataAtom = a.Img
+ p.tok.Data = a.Img.String()
+ return false
+ case a.Isindex:
+ if p.form != nil {
+ // Ignore the token.
+ return true
+ }
+ action := ""
+ prompt := "This is a searchable index. Enter search keywords: "
+ attr := []Attribute{{Key: "name", Val: "isindex"}}
+ for _, t := range p.tok.Attr {
+ switch t.Key {
+ case "action":
+ action = t.Val
+ case "name":
+ // Ignore the attribute.
+ case "prompt":
+ prompt = t.Val
+ default:
+ attr = append(attr, t)
+ }
+ }
+ p.acknowledgeSelfClosingTag()
+ p.popUntil(buttonScope, a.P)
+ p.parseImpliedToken(StartTagToken, a.Form, a.Form.String())
+ if action != "" {
+ p.form.Attr = []Attribute{{Key: "action", Val: action}}
+ }
+ p.parseImpliedToken(StartTagToken, a.Hr, a.Hr.String())
+ p.parseImpliedToken(StartTagToken, a.Label, a.Label.String())
+ p.addText(prompt)
+ p.addChild(&Node{
+ Type: ElementNode,
+ DataAtom: a.Input,
+ Data: a.Input.String(),
+ Attr: attr,
+ })
+ p.oe.pop()
+ p.parseImpliedToken(EndTagToken, a.Label, a.Label.String())
+ p.parseImpliedToken(StartTagToken, a.Hr, a.Hr.String())
+ p.parseImpliedToken(EndTagToken, a.Form, a.Form.String())
+ case a.Textarea:
+ p.addElement()
+ p.setOriginalIM()
+ p.framesetOK = false
+ p.im = textIM
+ case a.Xmp:
+ p.popUntil(buttonScope, a.P)
+ p.reconstructActiveFormattingElements()
+ p.framesetOK = false
+ p.addElement()
+ p.setOriginalIM()
+ p.im = textIM
+ case a.Iframe:
+ p.framesetOK = false
+ p.addElement()
+ p.setOriginalIM()
+ p.im = textIM
+ case a.Noembed, a.Noscript:
+ p.addElement()
+ p.setOriginalIM()
+ p.im = textIM
+ case a.Select:
+ p.reconstructActiveFormattingElements()
+ p.addElement()
+ p.framesetOK = false
+ p.im = inSelectIM
+ return true
+ case a.Optgroup, a.Option:
+ if p.top().DataAtom == a.Option {
+ p.oe.pop()
+ }
+ p.reconstructActiveFormattingElements()
+ p.addElement()
+ case a.Rp, a.Rt:
+ if p.elementInScope(defaultScope, a.Ruby) {
+ p.generateImpliedEndTags()
+ }
+ p.addElement()
+ case a.Math, a.Svg:
+ p.reconstructActiveFormattingElements()
+ if p.tok.DataAtom == a.Math {
+ adjustAttributeNames(p.tok.Attr, mathMLAttributeAdjustments)
+ } else {
+ adjustAttributeNames(p.tok.Attr, svgAttributeAdjustments)
+ }
+ adjustForeignAttributes(p.tok.Attr)
+ p.addElement()
+ p.top().Namespace = p.tok.Data
+ if p.hasSelfClosingToken {
+ p.oe.pop()
+ p.acknowledgeSelfClosingTag()
+ }
+ return true
+ case a.Caption, a.Col, a.Colgroup, a.Frame, a.Head, a.Tbody, a.Td, a.Tfoot, a.Th, a.Thead, a.Tr:
+ // Ignore the token.
+ default:
+ p.reconstructActiveFormattingElements()
+ p.addElement()
+ }
+ case EndTagToken:
+ switch p.tok.DataAtom {
+ case a.Body:
+ if p.elementInScope(defaultScope, a.Body) {
+ p.im = afterBodyIM
+ }
+ case a.Html:
+ if p.elementInScope(defaultScope, a.Body) {
+ p.parseImpliedToken(EndTagToken, a.Body, a.Body.String())
+ return false
+ }
+ return true
+ case a.Address, a.Article, a.Aside, a.Blockquote, a.Button, a.Center, a.Details, a.Dir, a.Div, a.Dl, a.Fieldset, a.Figcaption, a.Figure, a.Footer, a.Header, a.Hgroup, a.Listing, a.Menu, a.Nav, a.Ol, a.Pre, a.Section, a.Summary, a.Ul:
+ p.popUntil(defaultScope, p.tok.DataAtom)
+ case a.Form:
+ node := p.form
+ p.form = nil
+ i := p.indexOfElementInScope(defaultScope, a.Form)
+ if node == nil || i == -1 || p.oe[i] != node {
+ // Ignore the token.
+ return true
+ }
+ p.generateImpliedEndTags()
+ p.oe.remove(node)
+ case a.P:
+ if !p.elementInScope(buttonScope, a.P) {
+ p.parseImpliedToken(StartTagToken, a.P, a.P.String())
+ }
+ p.popUntil(buttonScope, a.P)
+ case a.Li:
+ p.popUntil(listItemScope, a.Li)
+ case a.Dd, a.Dt:
+ p.popUntil(defaultScope, p.tok.DataAtom)
+ case a.H1, a.H2, a.H3, a.H4, a.H5, a.H6:
+ p.popUntil(defaultScope, a.H1, a.H2, a.H3, a.H4, a.H5, a.H6)
+ case a.A, a.B, a.Big, a.Code, a.Em, a.Font, a.I, a.Nobr, a.S, a.Small, a.Strike, a.Strong, a.Tt, a.U:
+ p.inBodyEndTagFormatting(p.tok.DataAtom)
+ case a.Applet, a.Marquee, a.Object:
+ if p.popUntil(defaultScope, p.tok.DataAtom) {
+ p.clearActiveFormattingElements()
+ }
+ case a.Br:
+ p.tok.Type = StartTagToken
+ return false
+ default:
+ p.inBodyEndTagOther(p.tok.DataAtom)
+ }
+ case CommentToken:
+ p.addChild(&Node{
+ Type: CommentNode,
+ Data: p.tok.Data,
+ })
+ }
+
+ return true
+}
+
+func (p *parser) inBodyEndTagFormatting(tagAtom a.Atom) {
+ // This is the "adoption agency" algorithm, described at
+ // https://html.spec.whatwg.org/multipage/syntax.html#adoptionAgency
+
+ // TODO: this is a fairly literal line-by-line translation of that algorithm.
+ // Once the code successfully parses the comprehensive test suite, we should
+ // refactor this code to be more idiomatic.
+
+ // Steps 1-4. The outer loop.
+ for i := 0; i < 8; i++ {
+ // Step 5. Find the formatting element.
+ var formattingElement *Node
+ for j := len(p.afe) - 1; j >= 0; j-- {
+ if p.afe[j].Type == scopeMarkerNode {
+ break
+ }
+ if p.afe[j].DataAtom == tagAtom {
+ formattingElement = p.afe[j]
+ break
+ }
+ }
+ if formattingElement == nil {
+ p.inBodyEndTagOther(tagAtom)
+ return
+ }
+ feIndex := p.oe.index(formattingElement)
+ if feIndex == -1 {
+ p.afe.remove(formattingElement)
+ return
+ }
+ if !p.elementInScope(defaultScope, tagAtom) {
+ // Ignore the tag.
+ return
+ }
+
+ // Steps 9-10. Find the furthest block.
+ var furthestBlock *Node
+ for _, e := range p.oe[feIndex:] {
+ if isSpecialElement(e) {
+ furthestBlock = e
+ break
+ }
+ }
+ if furthestBlock == nil {
+ e := p.oe.pop()
+ for e != formattingElement {
+ e = p.oe.pop()
+ }
+ p.afe.remove(e)
+ return
+ }
+
+ // Steps 11-12. Find the common ancestor and bookmark node.
+ commonAncestor := p.oe[feIndex-1]
+ bookmark := p.afe.index(formattingElement)
+
+ // Step 13. The inner loop. Find the lastNode to reparent.
+ lastNode := furthestBlock
+ node := furthestBlock
+ x := p.oe.index(node)
+ // Steps 13.1-13.2
+ for j := 0; j < 3; j++ {
+ // Step 13.3.
+ x--
+ node = p.oe[x]
+ // Step 13.4 - 13.5.
+ if p.afe.index(node) == -1 {
+ p.oe.remove(node)
+ continue
+ }
+ // Step 13.6.
+ if node == formattingElement {
+ break
+ }
+ // Step 13.7.
+ clone := node.clone()
+ p.afe[p.afe.index(node)] = clone
+ p.oe[p.oe.index(node)] = clone
+ node = clone
+ // Step 13.8.
+ if lastNode == furthestBlock {
+ bookmark = p.afe.index(node) + 1
+ }
+ // Step 13.9.
+ if lastNode.Parent != nil {
+ lastNode.Parent.RemoveChild(lastNode)
+ }
+ node.AppendChild(lastNode)
+ // Step 13.10.
+ lastNode = node
+ }
+
+ // Step 14. Reparent lastNode to the common ancestor,
+ // or for misnested table nodes, to the foster parent.
+ if lastNode.Parent != nil {
+ lastNode.Parent.RemoveChild(lastNode)
+ }
+ switch commonAncestor.DataAtom {
+ case a.Table, a.Tbody, a.Tfoot, a.Thead, a.Tr:
+ p.fosterParent(lastNode)
+ default:
+ commonAncestor.AppendChild(lastNode)
+ }
+
+ // Steps 15-17. Reparent nodes from the furthest block's children
+ // to a clone of the formatting element.
+ clone := formattingElement.clone()
+ reparentChildren(clone, furthestBlock)
+ furthestBlock.AppendChild(clone)
+
+ // Step 18. Fix up the list of active formatting elements.
+ if oldLoc := p.afe.index(formattingElement); oldLoc != -1 && oldLoc < bookmark {
+ // Move the bookmark with the rest of the list.
+ bookmark--
+ }
+ p.afe.remove(formattingElement)
+ p.afe.insert(bookmark, clone)
+
+ // Step 19. Fix up the stack of open elements.
+ p.oe.remove(formattingElement)
+ p.oe.insert(p.oe.index(furthestBlock)+1, clone)
+ }
+}
+
+// inBodyEndTagOther performs the "any other end tag" algorithm for inBodyIM.
+// "Any other end tag" handling from 12.2.5.5 The rules for parsing tokens in foreign content
+// https://html.spec.whatwg.org/multipage/syntax.html#parsing-main-inforeign
+func (p *parser) inBodyEndTagOther(tagAtom a.Atom) {
+ for i := len(p.oe) - 1; i >= 0; i-- {
+ if p.oe[i].DataAtom == tagAtom {
+ p.oe = p.oe[:i]
+ break
+ }
+ if isSpecialElement(p.oe[i]) {
+ break
+ }
+ }
+}
+
+// Section 12.2.5.4.8.
+func textIM(p *parser) bool {
+ switch p.tok.Type {
+ case ErrorToken:
+ p.oe.pop()
+ case TextToken:
+ d := p.tok.Data
+ if n := p.oe.top(); n.DataAtom == a.Textarea && n.FirstChild == nil {
+ // Ignore a newline at the start of a