carlmontanari
commented
6 months ago hey @hroyd 👋
I'm pretty sure this is actually unrelated to lxml parsing -- this result should only ever be set if we dont think the actual chunk size matches what the router told us it should be.
what version of scrapli_netconf are you using? using latest this looks like it "should" work as intended, here is a down and dirty script that takes that chunk and rips out the response parsing bits into a simple little setup:
import re
CHUNK_MATCH_1_1 = re.compile(rb"(?P<size>\d+)\n(?P<content>.*?)^#", flags=re.M | re.S)
RAW = rb"""#303
<software-code-version>TiMOS-B-21.10.R9 both/hops64 Nokia 7750 SR Copyright(c) 2000-2022 Nokia.\nAll rights reserved. All use subject to applicable license agreements.\nBuilt on Wed Nov 30 13:05:52 PST 2022 by builder in /builds/c/2110B/R9/panos/main/sros\n</software-code-version>
"""
import re
CHUNK_MATCH_1_1 = re.compile(rb"(?P<size>\d+)\n(?P<content>.*?)^#", flags=re.M | re.S)
RAW = rb"""#303
<software-code-version>TiMOS-B-21.10.R9 both/hops64 Nokia 7750 SR Copyright(c) 2000-2022 Nokia.\nAll rights reserved. All use subject to applicable license agreements.\nBuilt on Wed Nov 30 13:05:52 PST 2022 by builder in /builds/c/2110B/R9/panos/main/sros\n</software-code-version>
#"""
def main():
chunk_matches = re.finditer(pattern=CHUNK_MATCH_1_1, string=RAW)
for chunk_match in chunk_matches:
size = int(chunk_match.groupdict().get("size", 0))
chunk = chunk_match.groupdict().get("content", "")
actual_size = len(chunk)
rstripped_size = len(chunk.rstrip())
trailing_newline_count = actual_size - rstripped_size
if trailing_newline_count > 1:
extraneous_trailing_newline_count = trailing_newline_count - 1
else:
extraneous_trailing_newline_count = 1
trimmed_newline_len = actual_size - extraneous_trailing_newline_count
if rstripped_size == 0:
actual_size = size
if size == actual_size:
print("matches actual")
elif size == rstripped_size:
print("matches rstripped")
elif size == trimmed_newline_len:
print("matches trimmed")
else:
print("no matches")
if __name__ == "__main__":
main()
tl;dr you can read the comments in the response parsing bits there's a bit of a saga about all this, but I think it should all have been sorted out by now (and appears so in this case or at least I hope).
hroyd
joshbenz
Hello,
We switched from nclient to scrapli for netconf operations. We issue a get request against a Nokia router for the card state, which contains some text with special characters, and this causes scrapli to throw an error message presumable in lxml parsing. The response.xml_result is exactly what we expect, but do not want to see that error message, or think it needs to be handled better in some way?
Thanks
Dean
Return element length invalid, expected 303 got 301 for element: b' <software-code-version>TiMOS-B-21.10.R9 both/hops64 Nokia 7750 SR Copyright (c) 2000-2022 Nokia.\nAll rights reserved. All use subject to applicable license agreements.\nBuilt on Wed Nov 30 13:05:52 PST 2022 by builder in /builds/c/2110B/R9/panos/main/sros\n</software-code-version>\n\n'