Ticket20218 rebased #1670
Ticket20218 rebased #1670
Conversation
- Check all fields that might change in a routerstatus - Document the refactoring Signed-off-by: Victor Nepveu <victor.nepveu@imt-atlantique.net>
Pull Request Test Coverage Report for Build 7807
|
src/test/test_nodelist.c
Outdated
| rs.is_flagged_running = 1; | ||
| ASSERT_CHANGED(); | ||
|
|
||
| // Isn't this obsolete? |
Named and Unnamed are no longer listed in the consensus (proposal 235).
But some of the client and authority code is still present:
https://gitweb.torproject.org/torspec.git/tree/proposals/282-remove-named-from-consensus.txt
If we want to remove this flag from client code, let's remove all instances, in another ticket?
And let's add a consensus method that removes the authority code? In another ticket?
And if we're looking to remove unused consensus code in a new consensus method, "package" lines can go too:
https://gitweb.torproject.org/torspec.git/tree/proposals/301-dont-vote-on-package-fingerprints.txt
src/test/test_nodelist.c
Outdated
| rs.has_bandwidth = 1; | ||
| ASSERT_CHANGED(); | ||
|
|
||
| // Does not actually matter unless exitsummary changes. |
It actually doesn't matter at all, exitsummary is not output via the control port routerstatus format. (See exitsummary below for details.)
routerstatus_has_changed() Is only checking for changes that get output over the control port:
It only checks for fields that are output by control port.
This should be kept in sync with the struct routerstatus_t
and the printing function routerstatus_format_entry in
NS_CONTROL_PORT mode.
Let's rename routerstatus_has_changed() to a name that reflects its purpose as a control-port only function.
In fact, has_exitsummary is not used in any routerstatus format. Perhaps it's used in microdescriptors?
|
|
||
| // Does not actually matter? Shouldn't it? XXXX | ||
| rs.bw_is_unmeasured = 1; | ||
| ASSERT_SAME(); |
This is a NS_CONTROL_PORT format change function, and this field is only output in NS_V3_VOTE format:
tor/src/feature/nodelist/fmt_routerstatus.c
Line 164 in 01fdc32
In fact, bw_is_unmeasured is not used in any routerstatus format. Perhaps it's used by other code?
| // XXX Shoudn't this count as a change? | ||
| rs.has_guardfraction = 1; | ||
| rs.guardfraction_percentage = 22; | ||
| ASSERT_SAME(); |
This is a NS_CONTROL_PORT format change function, and this field is only output in NS_V3_VOTE format:
tor/src/feature/nodelist/fmt_routerstatus.c
Line 169 in 01fdc32
|
|
||
| // Shouldn't this count as a change? | ||
| rs.exitsummary = (char*)"accept 1-2"; | ||
| ASSERT_SAME(); |
exitsummary is not used by in any routerstatus format. Perhaps it's used by other code?
… questions. Answer those questions.
Now they also check whether output of routerstatus_format_entry() has changed.
This is an automated commit, generated by this command:
./scripts/maint/rename_c_identifier.py \
routerstatus_has_changed routerstatus_has_visibly_changed
It was generated with --no-verify, since it introduces a wide line.
I'll fix it in a subsequent commit.
|
Resolved merge conflict in #1673. |
No description provided.
The text was updated successfully, but these errors were encountered: