Fix XR indent_adjust misfiring on 'template data timeout' and 'template options timeout' (follow-up to #205)

[btarno]

Problem

#208 (fixing #205) added a negative lookahead so the XR indent_adjust rule no longer misfires on the bare template timeout <s> leaf inside flow exporter-map ... version v9. However, IOS-XR also renders two sibling leaf forms in the same block:

flow exporter-map EXPORTER1
 version v9
  options interface-table timeout 300
  options sampler-table timeout 300
  template data timeout 15
  template options timeout 15
 !
 transport udp 2055
!
interface Loopback0
 ipv4 address 10.0.0.1 255.255.255.255

template data timeout 15 and template options timeout 15 do not match (?!\s+timeout), so the rule still fires, and since no end-template ever arrives, every subsequent line of the config is nested under the timeout leaf. On a real production IOS-XR config (Cisco 8000, ~1900 lines) the tree collapses from 249 top-level children to 110, and everything after the flow exporter-map (interfaces, router bgp, router isis, ...) silently stops matching for negations, merges, config_to_get_to, and future(). Any XR config that exports NetFlow/IPFIX with explicit template timeouts hits this.

Fix

Extend the lookahead to cover all three leaf forms from the flow exporter-map version block grammar (template timeout, template data timeout, template options timeout):

^\s*template(?!\s+(?:data\s+|options\s+)?timeout)

Real template <name> ... end-template blocks still match the rule; the existing test_template_block_indent_adjust covers that and still passes.

Testing

• Added test_flow_exporter_template_data_options_timeout_indent_adjust, modeled on the Update the XR driver per github issues #205 and #206 #208 test: asserts the sections following the exporter-map stay at the top level, the timeout leaves parse as children of version v9, and a remediation across the config is unaffected. The parse assertions fail on master and pass with this change.
• Full suite: 619 passed. ruff check, ruff format --check, and mypy clean on the touched files.
• Verified against a real ~1900-line IOS-XR 25.2.2 config: tree parses to the correct 249 top-level children and section negation (no interface ...) behaves correctly again.

[btarno]

ready for review

[jtdub]

Nice! Thanks Brad! I'll try to give this a look today.

[jtdub]

LGTM