Mirror
/
monolith
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Use a shared client initialized in main.rs

This commit is contained in:
Emi Simpson 2019-12-09 21:13:25 -05:00
parent 35f5e1353d
commit 65d0eab793
No known key found for this signature in database
GPG Key ID: 68FAB2E2E6DFC98B
3 changed files with 36 additions and 35 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
src/html.rs
View File

@ -7,6 +7,7 @@ use html5ever::tree_builder::{Attribute, TreeSink};
use html5ever::{local_name, namespace_url, ns}; use html5ever::{local_name, namespace_url, ns};
use http::retrieve_asset; use http::retrieve_asset;
use js::attr_is_event_handler; use js::attr_is_event_handler;
use reqwest::Client;
use std::collections::HashMap; use std::collections::HashMap;
use std::default::Default; use std::default::Default;
use utils::{data_to_dataurl, is_valid_url, resolve_url, url_has_protocol}; use utils::{data_to_dataurl, is_valid_url, resolve_url, url_has_protocol};
@ -45,14 +46,13 @@ pub fn is_icon(attr_value: &str) -> bool {
pub fn walk_and_embed_assets( pub fn walk_and_embed_assets(
cache: &mut HashMap<String, String>, cache: &mut HashMap<String, String>,
client: &Client,
url: &str, url: &str,
node: &Handle, node: &Handle,
opt_no_css: bool, opt_no_css: bool,
opt_no_js: bool, opt_no_js: bool,
opt_no_images: bool, opt_no_images: bool,
opt_user_agent: &str,
opt_silent: bool, opt_silent: bool,
opt_insecure: bool,
opt_no_frames: bool, opt_no_frames: bool,
) { ) {
match node.data { match node.data {
@ -61,14 +61,13 @@ pub fn walk_and_embed_assets(
for child in node.children.borrow().iter() { for child in node.children.borrow().iter() {
walk_and_embed_assets( walk_and_embed_assets(
cache, cache,
client,
&url, &url,
child, child,
opt_no_css, opt_no_css,
opt_no_js, opt_no_js,
opt_no_images, opt_no_images,
opt_user_agent,
opt_silent, opt_silent,
opt_insecure,
opt_no_frames, opt_no_frames,
); );
} }
@ -107,12 +106,11 @@ pub fn walk_and_embed_assets(
.unwrap_or(EMPTY_STRING.clone()); .unwrap_or(EMPTY_STRING.clone());
let (favicon_dataurl, _) = retrieve_asset( let (favicon_dataurl, _) = retrieve_asset(
cache, cache,
client,
&href_full_url, &href_full_url,
true, true,
"", "",
opt_user_agent,
opt_silent, opt_silent,
opt_insecure,
) )
.unwrap_or((EMPTY_STRING.clone(), EMPTY_STRING.clone())); .unwrap_or((EMPTY_STRING.clone(), EMPTY_STRING.clone()));
attr.value.clear(); attr.value.clear();
@ -131,12 +129,11 @@ pub fn walk_and_embed_assets(
.unwrap_or(EMPTY_STRING.clone()); .unwrap_or(EMPTY_STRING.clone());
let (css_dataurl, _) = retrieve_asset( let (css_dataurl, _) = retrieve_asset(
cache, cache,
client,
&href_full_url, &href_full_url,
true, true,
"text/css", "text/css",
opt_user_agent,
opt_silent, opt_silent,
opt_insecure,
) )
.unwrap_or((EMPTY_STRING.clone(), EMPTY_STRING.clone())); .unwrap_or((EMPTY_STRING.clone(), EMPTY_STRING.clone()));
attr.value.clear(); attr.value.clear();
@ -174,12 +171,11 @@ pub fn walk_and_embed_assets(
resolve_url(&url, &value).unwrap_or(EMPTY_STRING.clone()); resolve_url(&url, &value).unwrap_or(EMPTY_STRING.clone());
let (img_dataurl, _) = retrieve_asset( let (img_dataurl, _) = retrieve_asset(
cache, cache,
client,
&src_full_url, &src_full_url,
true, true,
"", "",
opt_user_agent,
opt_silent, opt_silent,
opt_insecure,
) )
.unwrap_or((EMPTY_STRING.clone(), EMPTY_STRING.clone())); .unwrap_or((EMPTY_STRING.clone(), EMPTY_STRING.clone()));
attr.value.clear(); attr.value.clear();
@ -208,12 +204,11 @@ pub fn walk_and_embed_assets(
.unwrap_or(EMPTY_STRING.clone()); .unwrap_or(EMPTY_STRING.clone());
let (source_dataurl, _) = retrieve_asset( let (source_dataurl, _) = retrieve_asset(
cache, cache,
client,
&srcset_full_url, &srcset_full_url,
true, true,
"", "",
opt_user_agent,
opt_silent, opt_silent,
opt_insecure,
) )
.unwrap_or((EMPTY_STRING.clone(), EMPTY_STRING.clone())); .unwrap_or((EMPTY_STRING.clone(), EMPTY_STRING.clone()));
attr.value.clear(); attr.value.clear();
@ -255,12 +250,11 @@ pub fn walk_and_embed_assets(
.unwrap_or(EMPTY_STRING.clone()); .unwrap_or(EMPTY_STRING.clone());
let (js_dataurl, _) = retrieve_asset( let (js_dataurl, _) = retrieve_asset(
cache, cache,
client,
&src_full_url, &src_full_url,
true, true,
"application/javascript", "application/javascript",
opt_user_agent,
opt_silent, opt_silent,
opt_insecure,
) )
.unwrap_or((EMPTY_STRING.clone(), EMPTY_STRING.clone())); .unwrap_or((EMPTY_STRING.clone(), EMPTY_STRING.clone()));
attr.value.clear(); attr.value.clear();
@ -309,25 +303,23 @@ pub fn walk_and_embed_assets(
resolve_url(&url, &iframe_src).unwrap_or(EMPTY_STRING.clone()); resolve_url(&url, &iframe_src).unwrap_or(EMPTY_STRING.clone());
let (iframe_data, iframe_final_url) = retrieve_asset( let (iframe_data, iframe_final_url) = retrieve_asset(
cache, cache,
client,
&src_full_url, &src_full_url,
false, false,
"text/html", "text/html",
opt_user_agent,
opt_silent, opt_silent,
opt_insecure,
) )
.unwrap_or((EMPTY_STRING.clone(), src_full_url)); .unwrap_or((EMPTY_STRING.clone(), src_full_url));
let dom = html_to_dom(&iframe_data); let dom = html_to_dom(&iframe_data);
walk_and_embed_assets( walk_and_embed_assets(
cache, cache,
client,
&iframe_final_url, &iframe_final_url,
&dom.document, &dom.document,
opt_no_css, opt_no_css,
opt_no_js, opt_no_js,
opt_no_images, opt_no_images,
opt_user_agent,
opt_silent, opt_silent,
opt_insecure,
opt_no_frames, opt_no_frames,
); );
let mut buf: Vec<u8> = Vec::new(); let mut buf: Vec<u8> = Vec::new();
@ -355,12 +347,11 @@ pub fn walk_and_embed_assets(
.unwrap_or(EMPTY_STRING.clone()); .unwrap_or(EMPTY_STRING.clone());
let (poster_dataurl, _) = retrieve_asset( let (poster_dataurl, _) = retrieve_asset(
cache, cache,
client,
&poster_full_url, &poster_full_url,
true, true,
"", "",
opt_user_agent,
opt_silent, opt_silent,
opt_insecure,
) )
.unwrap_or((poster_full_url, EMPTY_STRING.clone())); .unwrap_or((poster_full_url, EMPTY_STRING.clone()));
attr.value.clear(); attr.value.clear();
@ -404,14 +395,13 @@ pub fn walk_and_embed_assets(
for child in node.children.borrow().iter() { for child in node.children.borrow().iter() {
walk_and_embed_assets( walk_and_embed_assets(
cache, cache,
client,
&url, &url,
child, child,
opt_no_css, opt_no_css,
opt_no_js, opt_no_js,
opt_no_images, opt_no_images,
opt_user_agent,
opt_silent, opt_silent,
opt_insecure,
opt_no_frames, opt_no_frames,
); );
} }

12
src/http.rs
View File

@ -1,17 +1,15 @@
use reqwest::header::{CONTENT_TYPE, USER_AGENT}; use reqwest::header::CONTENT_TYPE;
use reqwest::Client; use reqwest::Client;
use std::collections::HashMap; use std::collections::HashMap;
use std::time::Duration;
use utils::{data_to_dataurl, is_data_url}; use utils::{data_to_dataurl, is_data_url};
pub fn retrieve_asset( pub fn retrieve_asset(
cache: &mut HashMap<String, String>, cache: &mut HashMap<String, String>,
client: &Client,
url: &str, url: &str,
as_dataurl: bool, as_dataurl: bool,
mime: &str, mime: &str,
opt_user_agent: &str,
opt_silent: bool, opt_silent: bool,
opt_insecure: bool,
) -> Result<(String, String), reqwest::Error> { ) -> Result<(String, String), reqwest::Error> {
if is_data_url(&url).unwrap() { if is_data_url(&url).unwrap() {
Ok((url.to_string(), url.to_string())) Ok((url.to_string(), url.to_string()))
@ -25,11 +23,7 @@ pub fn retrieve_asset(
Ok((data.to_string(), url.to_string())) Ok((data.to_string(), url.to_string()))
} else { } else {
// url not in cache, we request it // url not in cache, we request it
let client = Client::builder() let mut response = client.get(url).send()?;
.timeout(Duration::from_secs(10))
.danger_accept_invalid_certs(opt_insecure)
.build()?;
let mut response = client.get(url).header(USER_AGENT, opt_user_agent).send()?;
if !opt_silent { if !opt_silent {
if url == response.url().as_str() { if url == response.url().as_str() {

25
src/main.rs
View File

@ -1,6 +1,7 @@
#[macro_use] #[macro_use]
extern crate clap; extern crate clap;
extern crate monolith; extern crate monolith;
extern crate reqwest;
mod args; mod args;
@ -8,34 +9,50 @@ use args::AppArgs;
use monolith::html::{html_to_dom, stringify_document, walk_and_embed_assets}; use monolith::html::{html_to_dom, stringify_document, walk_and_embed_assets};
use monolith::http::retrieve_asset; use monolith::http::retrieve_asset;
use monolith::utils::is_valid_url; use monolith::utils::is_valid_url;
use reqwest::header::{HeaderMap, HeaderValue, USER_AGENT};
use std::collections::HashMap; use std::collections::HashMap;
use std::time::Duration;
fn main() { fn main() {
let app_args = AppArgs::get(); let app_args = AppArgs::get();
let cache = &mut HashMap::new(); let cache = &mut HashMap::new();
if is_valid_url(app_args.url_target.as_str()) { if is_valid_url(app_args.url_target.as_str()) {
// Initialize client
let mut header_map = HeaderMap::new();
match HeaderValue::from_str(&app_args.user_agent) {
Ok(header) => header_map.insert(USER_AGENT, header),
Err(err) => {
eprintln!("Invalid user agent! {}", err);
return;
}
};
let client = reqwest::Client::builder()
.timeout(Duration::from_secs(10))
.danger_accept_invalid_certs(app_args.insecure)
.default_headers(header_map)
.build()
.expect("Failed to initialize HTTP client");
let (data, final_url) = retrieve_asset( let (data, final_url) = retrieve_asset(
cache, cache,
&client,
app_args.url_target.as_str(), app_args.url_target.as_str(),
false, false,
"", "",
app_args.user_agent.as_str(),
app_args.silent, app_args.silent,
app_args.insecure,
) )
.unwrap(); .unwrap();
let dom = html_to_dom(&data); let dom = html_to_dom(&data);
walk_and_embed_assets( walk_and_embed_assets(
cache, cache,
&client,
&final_url, &final_url,
&dom.document, &dom.document,
app_args.no_css, app_args.no_css,
app_args.no_js, app_args.no_js,
app_args.no_images, app_args.no_images,
app_args.user_agent.as_str(),
app_args.silent, app_args.silent,
app_args.insecure,
app_args.no_frames, app_args.no_frames,
); );