control: owner -1 !
control: tags -1 moreinfo
Hello,
looking at nanomsg diff:
diff -Nru nanomsg-1.1.5+dfsg/src/nn.h nanomsg-1.2+dfsg/src/nn.h
--- nanomsg-1.1.5+dfsg/src/nn.h 2018-10-15 15:50:59.000000000 +0200
+++ nanomsg-1.2+dfsg/src/nn.h 2022-04-19 01:29:26.000000000 +0200
@@ -2,7 +2,7 @@
Copyright (c) 2012-2014 Martin Sustrik All rights reserved.
Copyright (c) 2013 GoPivotal, Inc. All rights reserved.
Copyright (c) 2015-2016 Jack R. Dunaway. All rights reserved.
- Copyright 2017 Garrett D'Amore <garrett@damore.org>
+ Copyright 2022 Staysail Systems, Inc. <info@staysail.tech>
+++ nanomsg-1.2+dfsg/src/transports/utils/literal.c 2022-04-19 01:29:26.000000000 +0200
@@ -1,5 +1,6 @@
/*
Copyright (c) 2012-2013 Martin Sustrik All rights reserved.
+ Copyright 2022 Staysail Systems, Inc.
+++ nanomsg-1.2+dfsg/src/utils/stopwatch.h 2022-04-19 01:29:26.000000000 +0200
@@ -1,6 +1,6 @@
/*
Copyright (c) 2012 Martin Sustrik All rights reserved.
- Copyright 2015 Garrett D'Amore <garrett@damore.org>
+ Copyright 2022 Staysail Systems, Inc.
These looks missing
But the main reason for the moreinfo tag is that I checked the whole diff file, and there is just a new symbols added.
void nn_literal_link_local_resolve(struct in6_addr *in6addr, int64_t *sin6_scope_id, char *addr)
And looks not part of the public API exposed, but just called from nn_literal_resolve function.
If the symbols are added and not changed, we should be ABI safe, so no transition is needed, right?
Please double check for ABI stability, because if this is true, we can save an useless trip to new queue.
I leave to you to create a new symlink for backward compatibility, to revert the library soname bump, or to keep things as
is.
Also, since the transition will likely involve a couple of packages, it might be a good point to just
follow what upstream named and let things go that way (and maybe open an upstream bug asking to avoid
changing soname if the API is stable)
G.
Attachment:
OpenPGP_signature.asc
Description: OpenPGP digital signature