Cut URLs before adding to automcomp at >

[jubalh]

First I tried with g_uri_parse() and g_uri_to_string() but then I learned that GUri validation API is only for things that are part of a proper URL.

Let's cut the string at > since they are sometimes enclosed in <>.

Fix #1877

[jubalh]

Maybe adjusting the regex used at https://github.com/profanity-im/profanity/pull/1883/files#diff-1a7db58e738a75a1ee3630328ced5287e85e1f30168c13d6d1d680d5f7c2a830L1313 would be better. If anyone is a regex master here .. ;)

[sjaeckel]

Maybe adjusting the regex used at https://github.com/profanity-im/profanity/pull/1883/files#diff-1a7db58e738a75a1ee3630328ced5287e85e1f30168c13d6d1d680d5f7c2a830L1313 would be better. If anyone is a regex master here .. ;)

... now we have two problems ...

$ git diff -U1
diff --git a/src/ui/window_list.c b/src/ui/window_list.c
index a00fcb43..2206c6d8 100644
--- a/src/ui/window_list.c
+++ b/src/ui/window_list.c
@@ -1312,3 +1312,3 @@ wins_add_urls_ac(const ProfWin* const win, const ProfMessage* const message, con
 
-    regex = g_regex_new("(https?|aesgcm)://\\S+", 0, 0, NULL);
+    regex = g_regex_new("(https?|aesgcm)://[\\w\\-.~:/?#\\[\\]@!$&'()*+,;=%]+", 0, 0, NULL);
     g_regex_match(regex, message->plain, 0, &match_info);

No regex master here, https://stackoverflow.com/q/43588699 did most of the work, I only had to adapt the escaping.

[sjaeckel]

Could you please add a comment that references the SO thread?