Support ' and () in url

since they are valid and can be found in url (ex: kibana), support
them.

Signed-off-by: Gaetan Semet <[email protected]>

Author: gsemet

guake/globals.py

@@ -72,10 +72,12 @@ def is_run_from_git_workdir():
 # TODO this is not as fancy as as it could be
 # pylint: disable=anomalous-backslash-in-string
 TERMINAL_MATCH_TAGS = ('schema', 'http', 'https', 'email', 'ftp')
+# Beware this is a PRCE (Perl) regular expression, not a Python one!
+# Edit: use regex101.com with PCRE syntax
 TERMINAL_MATCH_EXPRS = [
-    "(news:|telnet:|nntp:|file:\/|https?:|ftps?:|webcal:)\/\/"
-    "([-[:alnum:]]+(:[-[:alnum:],?;.:\/!%$^*&~\"#']+)?\@)?[-[:alnum:]]+"
-    "(\.[-[:alnum:]]+)*(:[0-9]{1,5})?(\/[-[:alnum:]_$.+!*(),;:@&=?\/~#%]*[^]'.>) \t\r\n,\\\"])?",
+    "(news:|telnet:|nntp:|file:\/|https?:|ftps?:|webcal:)\/\/([-[:alnum:]]+"
+    "(:[-[:alnum:],?;.:\/!%$^\*&~\"#']+)?\@)?[-[:alnum:]]+(\.[-[:alnum:]]+)*"
+    "(:[0-9]{1,5})?(\/[-[:alnum:]_$.+!*(),;:@&=?\/~#'%]*[^].> \t\r\n,\\\"])?",
     "(www|ftp)[-[:alnum:]]*\.[-[:alnum:]]+(\.[-[:alnum:]]+)*(:[0-9]{1,5})?"
     "(\/[-[:alnum:]_$.+!*(),;:@&=?\/~#%]*[^]'.>) \t\r\n,\\\"])?",
     "(mailto:)?[-[:alnum:]][-[:alnum:].]*@[-[:alnum:]]+\.[-[:alnum:]]+(\\.[-[:alnum:]]+)*"

releasenotes/notes/hyperlinks-778efab6774df2e6.yaml

@@ -0,0 +1,17 @@
+fixes:
+  - |
+    URL with ``'`` (simple quote) and ``()`` (parenthesis) are now captured by hyperlink matcher.
+    This may causes some issues with log and so that use parenthesis *around* hyperlinks,
+    but since parenthesis and quotes are valid characters inside a URL, like for instance
+    URL created by Kibana, they deserve the right to be shown as proper url in Guake.
+
+    User can still select the URL in the terminal if he wishes to capture the exact url, before
+    doing a Ctrl+click or a right click.
+
+    For developers, it is advised to end the URL with a character that cannot be used in URL, such
+    as space, tab, new line. Ending with a dot (``.``) or a comma (``,``) will not be seen as part
+    of the URL by Guake, so most logs and traces that adds a dot or a comma at the end of the URL
+    might still work.
+
+known_issues:
+  - Multiline url are sometimes not handled correctly.

scripts/weird_urls.sh

@@ -0,0 +1,5 @@
+
+
+
+#!/bin/bash
+echo "blab blab la : http://mykibana.server/app/kibana#/discover?_g=(refreshInterval:(display:Off,pause:!f,value:0),time:(from:now-24h,mode:quick,to:now))&_a=(columns:!(levelname,syslog_message,extra.flat),interval:auto,query:(language:lucene,query:'uuid:%22076a4d47-4af5-4808-ab7d-1d5e6983e7c6%22'),sort:!('@timestamp',asc))"