From 383986e60df346bb03c2f02dd5f7bc81f5d20a79 Mon Sep 17 00:00:00 2001
From: Ryan Oldenburg <ryan@guzba.com>
Date: Wed, 31 Jan 2024 00:27:07 -0600
Subject: [PATCH] updated readme, libcurl small changes

---
 README.md                              | 58 +++++++++++++++++---------
 src/puppy/platforms/linux/platform.nim | 38 ++++++++---------
 2 files changed, 58 insertions(+), 38 deletions(-)

diff --git a/README.md b/README.md
index 30df6da..7efd3b4 100644
--- a/README.md
+++ b/README.md
@@ -4,50 +4,70 @@
 
 `nimble install puppy`
 
-![Github Actions](https://github.com/treeform/puppy/workflows/Github%20Actions/badge.svg)
+![Github Actions](https://github.com/treeform/puppy/actions/workflows/build.yml/badge.svg)
 
 [API reference](https://treeform.github.io/puppy)
 
 ## About
 
-Puppy does not use Nim's HTTP stack, instead it uses `WinHttp` API on Windows , `AppKit` on macOS, and `libcurl` on Linux. Because Puppy uses system APIs, there is no need to ship extra `*.dll`s, `cacert.pem`, or forget to pass the `-d:ssl` flag. This also has the effect of producing slightly smaller binaires.
+Puppy makes HTTP requests easy!
 
-Furthermore, Puppy supports gzip transparently right out of the box.
+With Puppy you can make HTTP requests without needing to pass the `-d:ssl` flag or shipping extra `*.dll`s and `cacerts.pem` on Windows. Puppy avoids these gotchas by using system APIs instead of Nim's HTTP stack.
+
+Some other highlights of Puppy are:
+
+* Supports gzip'ed responses out of the box
+* Make an HTTP request using a one-line `proc` call
 
 OS    |  Method
 ----- | ---------------------------
 Win32 | WinHttp WinHttpRequest
 macOS | AppKit NSMutableURLRequest
-linux | libcurl easy_perform
+Linux | libcurl easy_perform
 
 *Curently does not support async*
 
-```nim
-import puppy
+## Easy mode
 
+```nim
 echo fetch("http://neverssl.com/")
 ```
 
-Will raise `PuppyError` if the response status code is not `200`.
+A call to `fetch` will raise PuppyError if the response status code is not 200.
+
+## More request types
+
+Make a basic GET request:
+
+```nim
+import puppy
+
+let response = get("https://www.google.com/")
+```
 
 Need to pass headers?
 
 ```nim
 import puppy
 
-echo fetch(
+let response = get(
   "http://neverssl.com/",
   headers = @[("User-Agent", "Nim 1.0")]
 )
 ```
 
-Need a more complex API?
-* verbs: GET, POST, PUT, UPDATE, DELETE..
-* headers: User-Agent, Content-Type..
-* response code: 200, 404, 500..
-* response headers: Content-Type..
+Easy one-line procs for your favorite verbs:
+
+```nim
+discard get(url, headers)
+discard post(url, headers, body)
+discard put(url, headers, body)
+discard patch(url, headers, body)
+discard delete(url, headers)
+discard head(url, headers)
+```
 
-Use these instead.
+## Working with responses
 
 ```nim
 Response* = ref object
@@ -56,8 +76,6 @@ Response* = ref object
   body*: string
 ```
 
-Usage examples:
-
 ```nim
 import puppy
 
@@ -82,7 +100,7 @@ echo response.headers
 echo response.body.len
 ```
 
-## Examples
+## More examples
 
 Using multipart/form-data:
 
@@ -109,6 +127,8 @@ let response = post("Your API endpoint here", headers, body)
 
 See the [examples/](https://github.com/treeform/puppy) folder for more examples.
 
-## Always use Libcurl
+## Always use libcurl
+
+You can pass `-d:puppyLibcurl` to force use of `libcurl` even on Windows and macOS. This is useful if for some reason the native OS API is not working.
 
-You can pass `-d:puppyLibcurl` to force use of `libcurl` even on windows and macOS. This is useful to debug, if the some reason native OS API does not work. Libcurl is usually installed on macOS but requires a `curl.dll` on windows.
+Libcurl is typically ready-to-use on macOS and Linux. On Windows you'll need to grab the latest libcurl DLL from https://curl.se/windows/, rename it to libcurl.dll, and put it in the same directory as your executable.
diff --git a/src/puppy/platforms/linux/platform.nim b/src/puppy/platforms/linux/platform.nim
index 793998a..8c98cc5 100644
--- a/src/puppy/platforms/linux/platform.nim
+++ b/src/puppy/platforms/linux/platform.nim
@@ -15,26 +15,26 @@ type StringWrap = object
   ## some sort of wrapper to be passed to C.
   str: string
 
+{.push stackTrace: off.}
+
+proc curlWriteFn(
+  buffer: cstring,
+  size: int,
+  count: int,
+  outstream: pointer
+): int {.cdecl.} =
+  let
+    outbuf = cast[ptr StringWrap](outstream)
+    i = outbuf.str.len
+  outbuf.str.setLen(outbuf.str.len + count)
+  copyMem(outbuf.str[i].addr, buffer, count)
+  result = size * count
+
+{.pop.}
+
 proc internalFetch*(req: Request): Response {.raises: [PuppyError].} =
   result = Response()
 
-  {.push stackTrace: off.}
-
-  proc curlWriteFn(
-    buffer: cstring,
-    size: int,
-    count: int,
-    outstream: pointer
-  ): int {.cdecl.} =
-    let
-      outbuf = cast[ptr StringWrap](outstream)
-      i = outbuf.str.len
-    outbuf.str.setLen(outbuf.str.len + count)
-    copyMem(outbuf.str[i].addr, buffer, count)
-    result = size * count
-
-  {.pop.}
-
   var strings: seq[string]
   strings.add $req.url
   strings.add req.verb.toUpperAscii()
@@ -65,9 +65,9 @@ proc internalFetch*(req: Request): Response {.raises: [PuppyError].} =
 
     discard curl.easy_setopt(OPT_HTTPHEADER, headerList)
 
-    if req.verb.toUpperAscii() == "HEAD":
+    if cmpIgnoreCase(req.verb, "HEAD") == 0:
       discard curl.easy_setopt(OPT_NOBODY, 1)
-    elif req.verb.toUpperAscii() == "POST" or req.body.len > 0:
+    elif cmpIgnoreCase(req.verb, "POST") == 0 or req.body.len > 0:
       discard curl.easy_setopt(OPT_POSTFIELDSIZE, req.body.len)
       discard curl.easy_setopt(OPT_POSTFIELDS, req.body.cstring)