varia/etherpump
14
0
Fork 0
Code Issues 15 Pull Requests Releases Activity

Add back try/catch to match dhtml code block

Browse Source
This commit is contained in:
Luke Murphy 2020-12-12 16:42:57 +01:00
parent cc0bcb5af0
commit 6cd15a6256
No known key found for this signature in database
GPG Key ID: 5E2EF5A63E3718CC
1 changed files with 43 additions and 37 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

80
etherpump/commands/pull.py
View File

@ -470,22 +470,25 @@ async def handle_pad(args, padid, data, info, session):
downloaded_html = True
if html["_code"] == 200:
html_body = html["data"]["html"]
ver["path"] = p + ".raw.html"
ver["url"] = quote(ver["path"])
doc = html5lib.parse(
html_body, treebuilder="etree", namespaceHTMLElements=False
)
html5tidy(
doc,
indent=True,
title=padid,
scripts=args.script,
links=links,
)
async with await trio.open_file(ver["path"], "w") as f:
output = ET.tostring(doc, method="html", encoding="unicode")
await f.write(output)
try:
html_body = html["data"]["html"]
ver["path"] = p + ".raw.html"
ver["url"] = quote(ver["path"])
doc = html5lib.parse(
html_body, treebuilder="etree", namespaceHTMLElements=False
)
html5tidy(
doc,
indent=True,
title=padid,
scripts=args.script,
links=links,
)
async with await trio.open_file(ver["path"], "w") as f:
output = ET.tostring(doc, method="html", encoding="unicode")
await f.write(output)
except TypeError:
ver["message"] = html["message"]
if args.all or args.magicwords:
if not downloaded_html:
@ -495,28 +498,31 @@ async def handle_pad(args, padid, data, info, session):
ver["code"] = html["_code"]
if html["_code"] == 200:
html_body = html["data"]["html"]
ver["path"] = p + ".magicwords.html"
ver["url"] = quote(ver["path"])
for magic_word in magic_words:
replace_word = (
"<span class='highlight'>" + magic_word + "</span>"
try:
html_body = html["data"]["html"]
ver["path"] = p + ".magicwords.html"
ver["url"] = quote(ver["path"])
for magic_word in magic_words:
replace_word = (
"<span class='highlight'>" + magic_word + "</span>"
)
if magic_word in html_body:
html_body = html_body.replace(magic_word, replace_word)
doc = html5lib.parse(
html_body, treebuilder="etree", namespaceHTMLElements=False
)
if magic_word in html_body:
html_body = html_body.replace(magic_word, replace_word)
doc = html5lib.parse(
html_body, treebuilder="etree", namespaceHTMLElements=False
)
html5tidy(
doc,
indent=True,
title=padid,
scripts=args.script,
links=links,
)
async with await trio.open_file(ver["path"], "w") as f:
output = ET.tostring(doc, method="html", encoding="unicode")
await f.write(output)
html5tidy(
doc,
indent=True,
title=padid,
scripts=args.script,
links=links,
)
async with await trio.open_file(ver["path"], "w") as f:
output = ET.tostring(doc, method="html", encoding="unicode")
await f.write(output)
except TypeError:
ver["message"] = html["message"]
# output meta
if args.all or args.meta: