cadey/xesite
cadey
/
xesite
1
0
Fork 1
Code Issues Pull Requests 1 Projects Releases Wiki Activity

add these old posts from greedo (#31) 

* add these old posts from greedo

* Update textile-to-markdown-literate-haskell-2017-02-08.markdown
This commit is contained in:
Cadey Ratio 2019-04-12 14:35:00 -07:00 committed by GitHub
parent d647610dae
commit e1fdc6841b
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 243 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

182
blog/textile-to-markdown-literate-haskell-2017-02-08.markdown Normal file
View File

@ -0,0 +1,182 @@
---
title: textile-conversion Main
date: 2017-02-08
---
# textile-conversion Main
Author's Note: this was intended to be documentation for a service that never ended
up being implemented. It was going to help [Derpibooru](https://derpibooru.org)
convert its existing markup to [Markdown](https://github.github.com/gfm/). This
never happened.
This program listens on port 5000 and serves an unchecked-path web handler that
converts Derpibooru Textile via HTML into Markdown, using a two-step process.
The first step is to have SimpleTextile emit a HTML AST of the comment.
The second is to have Pandoc turn that HTML into Markdown.
This is intended to be helpful during Derpi's migration from Textile.
## Pragmas
The following pragma tells the compiler to automagically tease string literals
into whatever type they need to be. For more information on this, see [this page][hs-ovs].
```haskell
{-# LANGUAGE OverloadedStrings #-}
module Main where
```
## Imports
In order to accomplish our task, we need to import some libraries.
```haskell
import Data.String.Conv (toS)
import Network.Wai
import Network.HTTP.Simple
import Network.HTTP.Types
import Network.Wai.Handler.Warp (run)
import System.Environment (lookupEnv)
import Text.Pandoc
import Text.Pandoc.Error (PandocError, handleError)
```
## Helper Functions
getEnvDefault queries an environment variable, returning a default value if it
is unset.
```haskell
getEnvDefault :: String -> String -> IO String
getEnvDefault name default' = do
envvar <- lookupEnv name
case envvar of
Nothing -> pure default'
Just x -> pure x
```
---
htmlToMarkdown uses Pandoc to convert a HTML input string into the equivalent
Markdown. The `Either` type is used here in place of raising an exception.
```haskell
htmlToMarkdown :: String -> Either PandocError String
htmlToMarkdown inp = do
let
corpus = readHtml def inp
case corpus of
Left x -> Left x
Right x -> pure $ writeMarkdown def x
```
## Web Application
Now we are getting into the meat of the situation. This is the main
[Application][wai-application].
```haskell
toMarkdown :: Application
```
First, let's use a [guard][guards] to ensure that we are only accepting `POST`
requests. If the request is not a `POST` request, return [HTTP error code 405][http-4xx].
```haskell
toMarkdown req respond
| requestMethod req /= methodPost =
respond $ responseLBS
status405
[("Content-Type", "text/plain")]
"Not allowed"
```
Otherwise, this is a `POST` request, so we should:
1. Unpack the data from the post body of the HTTP request
2. Send the data to the Sinatra app for conversion from Textile to HTML
3. Take the resulting HTML and feed it to `htmlToMarkdown`
4. Respond with the resulting Markdown.
We use [http-conduit][http-conduit] to contact the Sinatra app.
```haskell
| otherwise = do
body <- requestBody req
targetHost <- getEnvDefault "TARGET_SERVER" "http://127.0.0.1:9292"
remoteRequest' <- parseRequest ("POST " ++ targetHost ++ "/textile/html")
```
The `($)` operator is a synonym for calling functions. It is defined in the [Prelude][dolla]
as `f $ x = f x` and is mainly used for omitting parentheses. Here it is used
to combine HTTP request settings into one big request.
Additionally we use a custom [Manager][manager] to avoid any issues with
request timeouts, as those are not important for the scope of this tool.
```haskell
let settings = defaultManagerSettings { managerResponseTimeout = Nothing }
manager <- newManager settings
let remoteRequest = setRequestBodyLBS (toS body)
$ setRequestManager manager
$ remoteRequest'
```
Now it is time to send off the request and unpack the response.
```haskell
response <- httpLBS remoteRequest
```
If the sinatra app failed to deal with this properly for some reason, report
its error as `text/plain` and return `400`.
```haskell
if getResponseStatusCode response /= 200
then respond $ responseLBS
status400
[("Content-Type", "text/plain")]
$ toS $ getResponseBody response
else do
let rbody = toS $ getResponseBody response
```
Convert the result body into Markdown. If there is an error, respond with a `400`
and the contents of that error.
```haskell
let mbody = htmlToMarkdown rbody
case mbody of
Left x ->
respond $ responseLBS
status400
[("Content-Type", "text/plain")]
$ toS $ show x
Right x -> do
respond $ responseLBS
status200
[("Content-Type", "text/markdown")]
$ toS x
```
Now we bootstrap it all by running the `toMarkdown` Application on port `5000`.
No other code is needed.
```haskell
main :: IO ()
main =
run 5000 toMarkdown
```
[hs-ovs]: https://ocharles.org.uk/blog/posts/2014-12-17-overloaded-strings.html
[wai-application]: https://hackage.haskell.org/package/wai
[guards]: https://en.wikibooks.org/wiki/Haskell/Control_structures
[http-4xx]: https://en.wikipedia.org/wiki/List_of_HTTP_status_codes#4xx_Client_Error
[http-conduit]: https://www.stackage.org/haddock/lts-6.6/http-conduit-2.1.11/Network-HTTP-Simple.html
[dolla]: https://hackage.haskell.org/package/base-4.9.0.0/docs/Prelude.html#v:-36-
[manger]: https://www.stackage.org/haddock/lts-6.5/http-client-0.4.29/Network-HTTP-Client.html#g:3

61
blog/webirc-protocol-2017-04-12.markdown Normal file
View File

@ -0,0 +1,61 @@
---
title: "IRCv3.2 `webirc` Extension"
date: 2017-04-12
---
# IRCv3.1 `webirc` Extension
This document does not describe a new IRCv3 standard. It is designed to
document how the existing `WEBIRC` mechanism works so there is a specification
to test things against. This is known to be implemented by all major IRC
daemons as of the time of this writing.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD",
"SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be
interpreted as described in RFC 2119.
Summary
-------
The `WEBIRC` verb allows a connecting IRC client to spoof its origin IP address
so that a user connecting via a gateway of some kind may have accountability
for their actions and bans against them do not affect unintended users of said
gateway.
This protocol verb must be sent before the initial `NICK` and `USER` handshake
and may be advertised as the client capability `webirc`. The remote server may
send a pre-connection `NOTICE` clarifying that the user has their specified IP
address and reverse DNS. Gateway implementors must not let the user set their
own IP address as part of connection negotiations.
Formatting
----------
The `WEBIRC` verb must be used as such:
WEBIRC <password> <client ident> <client reverse DNS> <client IP address>
Access to `WEBIRC` must be protected by a password to prevent abuse. If the
password the client gives fails, the IRC daemon should disconnect the client
with an appropriate error message. IRC daemon authors should also restruct the
use of the `WEBIRC` verb to a specific IP address and may force the use of
a specific identd reply.
Example Session
---------------
>> WEBIRC snowflower Mibbit anonyhash.mibbit.com 127.0.0.1
>> NICK mib_4002
>> USER Mibbit x x :http://mibbit.com AJAX IRC Client
<< :hostname.domain.tld 001 mib_4002 :Welcome to ShadowNET mib_4002!
Limitations
-----------
In order for this to be secure, the relay server must be trusted by the IRC
server. A remote server may kill off clients that fail the password and host
check, but this is not required.
---
This was recovered from an old backup of my site data on 2019-04-12.