- Oct 12, 2022
-
-
Robert Franklin authored
-
- Oct 06, 2022
-
-
Robert Franklin authored
added first level which just prints the individual converter matches, rather than details of the conversions - this provides a simple way to get a summary of all changes to be made, one per line also did some code tidying
-
Robert Franklin authored
-
Robert Franklin authored
-
Robert Franklin authored
when using the first level of debugging the convert (a single -V) then only a single line displaying the matches will be printed for each match, so the blank lines just waste space -- this removes those, when using that higher debug levels on the convert (e.g. -VV) will leave the blank lines as there'll be multiple lines for each hit
-
Robert Franklin authored
was using a mixture of 'debug_msg += ["..."]' and 'debug_msg.append("...")' - settled on the latter
-
Robert Franklin authored
this just displays the remove/update/trigger match and not any of the other details, giving a useful summary of the changes being made this bumps the debug levels for other -V options one level down
-
- Oct 03, 2022
-
-
Robert Franklin authored
interface VRF changes correct/corrupt 'ip helper-address' commands so they are removed, prior to the change, then reinstated afterwards this update also improves error messages from exceptions in converter methods, as well as the trigger() method calls
-
Robert Franklin authored
trigger() is only called when a converter is called is triggered explicitly by another converter but not because of other changes, so there will be no change - as such, only the old/new configuration is required and these will be the same, so only 'new' is supplied
-
Robert Franklin authored
these now print out the old/remove/update/new arguments to help debug why the crash is happening; previously these would only be printed if a suitable debugging level is enabled
-
- Sep 30, 2022
-
-
Robert Franklin authored
previously displayed entire path to class
-
Robert Franklin authored
-
Robert Franklin authored
changing (or setting) the VRF on an interface causes 'ip helper-address' commands to be corrupter (well, changed, as required) this update causes those commands to be removed from the configuration, the VRF changed and then the commands re-applied
-
Robert Franklin authored
-
Robert Franklin authored
was incorrectly matching a 'vrf' field as global
-
Robert Franklin authored
converter flag which specifis a converter should always fire a trigger, even if the converter methods don't generate any output
-
Robert Franklin authored
there is still a bug when changing VRF but commit this as an clean up, prior to fixing that
-
Robert Franklin authored
ipv6 address, ip pim bsr-border and ipv6 pim bsr border all get removed by changing the VRF - moved those after VRF change
-
- Sep 09, 2022
-
-
Robert Franklin authored
-
Robert Franklin authored
enter() methods weren't correctly returning a list (of one line, typically) following the previous changes to tidy up these uses
-
- Sep 05, 2022
-
-
Robert Franklin authored
bug fix: set ip next-hop now handles multiple IP addresses
-
Robert Franklin authored
regular expression didn't match a space between addresses!
-
- Sep 02, 2022
-
-
Robert Franklin authored
this is actually how the help says the script should work (well, just says 'debug mode', from a long time ago!) but this was lost at some point
-
- Aug 31, 2022
-
-
Robert Franklin authored
route-maps are now handled by the parser and converter interface / ip[v6] policy route-map commands are now handled by the parser and converter no longer deletes non-existent old ACLs, prefix-lists and track objects when creating new ones error messages from converter methods now show the configuration dictionaries when raising the exception converter get_ext() method now supports wildcard arguments optimised usage of converter enter() methods to avoid unnecessary expansion removed unnecessary debugging output left in code when not debugging
-
- Aug 30, 2022
-
-
Robert Franklin authored
also, improved the service-policy interface command with a _cmd() method
-
Robert Franklin authored
the syntax can be a unpleasant, in some cases but more efficient
-
Robert Franklin authored
was expanding and recreating the list of commands unnecessarily
-
Robert Franklin authored
this ensures two things: first - removing all the old clauses before adding the new ones ensures that conflicting entries are never attempted to be created, and second - avoids resetting the action of a route-map entry ('permit' or 'deny') to the old type, if removing an old clause after setting a new one
-
Robert Franklin authored
when list was previously empty
-
Robert Franklin authored
-
Robert Franklin authored
correctly handles changing between the two and 'global' now handled as an empty string (which makes comparison using deepdiff() work)
-
Robert Franklin authored
-
Robert Franklin authored
-
Robert Franklin authored
-
Robert Franklin authored
still need to do 'set'
-
Robert Franklin authored
also, fixed plain 'set ip next-hop' to not match these commands
-
Robert Franklin authored
-
Robert Franklin authored
-
Robert Franklin authored
-
Robert Franklin authored
-