Actions
Bug #6236
closedfix duality: SDP fmtp parameter 'octet-align' vs 'octet-aligned'
Status:
Resolved
Priority:
Normal
Assignee:
-
Target version:
-
Start date:
10/26/2023
Due date:
% Done:
0%
Spec Reference:
Description
across all osmocom CNI source files:
$ grep -h 'octet-align\>' $(find . -name "*.[hc]") | grep -v '^[[:blank:]]*[/*]' | wc -l 36 $ grep -h 'octet-aligned' $(find . -name "*.[hc]") | grep -v '^[[:blank:]]*[/*]' | wc -l 24
- which one is the correct version?
- is the SDP world out there de-facto using both versions? if yes always query both versions for compat.
- decide for one version that we put in outgoing SDP and apply in our code base
Updated by neels 6 months ago
https://www.rfc-editor.org/rfc/rfc4867#section-8.1 is clear:
the correct version is:
octet-align=1
Updated by neels 6 months ago
Though in flow texts, it talks of "octet-aligned mode":
octet-align: Permissible values are 0 and 1. If 1, octet-aligned
operation SHALL be used. If 0 or if not present,
bandwidth-efficient operation is employed.
on the very line of the definition of 'octet-align', it speaks of "octet-aligned operation". No wonder the confusion.
Updated by neels 6 months ago
- Status changed from New to Resolved
ok, after looking around, actually all places handling SDP use the correct format.
Some regression test output looks like it were trouble, but since it's only printf to log it is actually irrelevant.
Many places use 'octet-aligned' in log output and doc strings, which is perfectly fine and makes sense.
One place in osmo-cbc uses octet-aligned in the context of asn.1...
so just a bunch of false positives.
Actions