From: Adam Sampson <ats@offog.org>
Date: Mon, 11 Jun 2018 19:30:33 +0000 (+0100)
Subject: Override pytidylib's BASE_OPTIONS explicitly.
X-Git-Url: http://git.ozo.com/?a=commitdiff_plain;h=1ca4d2f1ad8c66d4235876b7eb39657a115ac2bb;p=rawdog%2F.git

Override pytidylib's BASE_OPTIONS explicitly.

This variable has moved from tidylib.BASE_OPTIONS to
tidylib.tidy.BASE_OPTIONS as of pytidylib 0.3.2.
---

diff --git a/NEWS b/NEWS
index 2e7d5e2..161290b 100644
--- a/NEWS
+++ b/NEWS
@@ -7,6 +7,9 @@ of the feed in question).
 Add the --find option, which shows what feedscanner returns for a given
 URL (as --dump does for feedparser).
 
+The location of BASE_OPTIONS has changed in pytidylib 0.3.2; rather than
+trying to change it, override the corresponding options explicitly.
+
 - rawdog 2.22
 
 When handling an HTTP 301 redirect response, check whether the new
diff --git a/rawdoglib/rawdog.py b/rawdoglib/rawdog.py
index 5c4eaa8..5731593 100644
--- a/rawdoglib/rawdog.py
+++ b/rawdoglib/rawdog.py
@@ -134,20 +134,27 @@ def sanitise_html(html, baseurl, inline, config):
 			html = "<p>" + html
 
 	if config["tidyhtml"]:
+		# This must include: options where the default value in tidy
+		# has changed at some point, and pytidylib's BASE_OPTIONS which
+		# it would otherwise set if we didn't specify them.
 		args = {
 			"numeric_entities": 1,
+			# In tidy 0.99 these are ASCII; in tidy 5, UTF-8.
 			"input_encoding": "ascii",
 			"output_encoding": "ascii",
 			"output_html": 1,
 			"output_xhtml": 0,
 			"output_xml": 0,
+			"indent": 0,
+			"tidy-mark": 0,
+			"alt-text": "",
+			"doctype": "strict",
+			"force-output": 1,
 			"wrap": 0,
 			}
 		call_hook("mxtidy_args", config, args, baseurl, inline)
 		call_hook("tidy_args", config, args, baseurl, inline)
 		if tidylib is not None:
-			# Disable PyTidyLib's somewhat unhelpful defaults.
-			tidylib.BASE_OPTIONS = {}
 			output = tidylib.tidy_document(html, args)[0]
 		elif mxtidy is not None:
 			output = mxtidy.tidy(html, None, None, **args)[2]