termite: Add patch to avoid matching trailing punctuation in URLs

Patch based on: thestinger/termite#516 While this might "break" some URL's in the wild, in return this fixes matching of URL's in Nix code. Without this patch the trailing semicolon is included as part of the URL, which creates invalid URL's for nearly all meta.homepage values.
NixOS · Dec 20, 2017 · c433fef · c433fef
1 parent 8e6d563
commit c433fef
Show file tree

Hide file tree

Showing 2 changed files with 30 additions and 0 deletions.
diff --git a/pkgs/applications/misc/termite/default.nix b/pkgs/applications/misc/termite/default.nix
@@ -13,6 +13,9 @@ let
       sha256 = "02cn70ygl93ghhkhs3xdxn5b1yadc255v3yp8cmhhyzsv5027hvj";
     };
 
+    # https://github.com/thestinger/termite/pull/516
+    patches = [ ./url_regexp_trailing.patch ];
+
     postPatch = "sed '1i#include <math.h>' -i termite.cc";
 
     makeFlags = [ "VERSION=v${version}" "PREFIX=" "DESTDIR=$(out)" ];

diff --git a/pkgs/applications/misc/termite/url_regexp_trailing.patch b/pkgs/applications/misc/termite/url_regexp_trailing.patch
@@ -0,0 +1,27 @@
+Based on https://github.com/thestinger/termite/pull/516
+Modified to apply to v13
+
+From 65a454ffa8e681f3f14729cba7c42e1570a85e8a Mon Sep 17 00:00:00 2001
+From: Paul Baecher <pbaecher@gmail.com>
+Date: Thu, 7 Sep 2017 22:58:51 +0200
+Subject: [PATCH] Do not match punctuation at the end of URLs
+
+Punctuation at the end of URLs is most likely part of natural language
+or markup (for example in Markdown). Do not match it as part of the URL.
+---
+ url_regex.hh | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/url_regex.hh b/url_regex.hh
+index 2ec6be8..3039b53 100644
+--- a/url_regex.hh
++++ b/url_regex.hh
+@@ -9,7 +9,7 @@
+ #define PORT            "(?:\\:[[:digit:]]{1,5})?"
+ #define SCHEME          "(?:[[:alpha:]][+-.[:alnum:]]*:)"
+ #define USERPASS        USERCHARS_CLASS "+(?:\\:" PASSCHARS_CLASS "+)?"
+-#define URLPATH         "(?:/[[:alnum:]\\Q-_.!~*'();/?:@&=+$,#%\\E]*)?"
++#define URLPATH         "(?:/[[:alnum:]\\Q-_.!~*'();/?:@&=+$,#%\\E]*(?<![\\Q.,:;()!?\\E]))?"
+
+ const char * const url_regex = SCHEME "//(?:" USERPASS "\\@)?" HOST PORT URLPATH;
+