Properly parse file:/// URIs

Previously, file URIs were lumped in with full uris that would be
used for HTTP, FTP, etc.  This caused file:/// uri parser to ignore
any file with a root dir that had a character that was not a valid
hostname character to be ignored.
This commit is contained in:
Matt Rose 2024-03-11 15:17:51 -04:00
parent 51e3bd2519
commit 4e6e293771

View File

@ -300,6 +300,7 @@ class Terminal(Gtk.VBox):
registry.load_plugins(force)
def _add_regex(self, name, re):
dbg(f"adding regex: {re}")
match = -1
if regex.FLAGS_PCRE2:
try:
@ -323,19 +324,23 @@ class Terminal(Gtk.VBox):
passchars = "-A-Za-z0-9,?;.:/!%$^*&~\"#'"
hostchars = r"-A-Za-z0-9:\[\]"
pathchars = "-A-Za-z0-9_$.+!*(),;:@&=?/~#%'"
schemes = "(news:|telnet:|nntp:|file:/|https?:|ftps?:|webcal:|ssh:)"
schemes = "(news:|telnet:|nntp:|https?:|ftps?:|webcal:|ssh:)"
user = "[" + userchars + "]+(:[" + passchars + "]+)?"
urlpath = "/[" + pathchars + "]*[^]'.}>) \t\r\n,\\\"]"
lboundry = "\\b"
rboundry = "\\b"
re = (lboundry + "file:/" + "//?(:[0-9]+)?(" + urlpath + ")" +
rboundry + "/?")
self._add_regex('file', re)
re = (lboundry + schemes +
"//(" + user + "@)?[" + hostchars +".]+(:[0-9]+)?(" +
urlpath + ")?" + rboundry + "/?")
self._add_regex('full_uri', re)
if self.matches['full_uri'] == -1:
if self.matches['full_uri'] == -1 or self.matches['file'] == -1:
err ('Terminal::update_url_matches: Failed adding URL matches')
else:
re = (lboundry +