libgcrypt20 backported to stretch

[Matteo F. Vescovi]

Hi!

On 2017-10-11 at 18:47 (+0200), Andreas Metzler wrote:

[...]

> Hello Matteo,
>
> remmina Seems to be a popular package. ;-)
> Raphaël had contacted us at libgcrypt20 at pdo about a month with exactly
> the same question.
> http://lists.alioth.debian.org/pipermail/pkg-gnutls-maint/2017-September/008096.html
>
> We had some back and fore, but the main point is that it is not
> necessary to backport libgcrypt. All that is needed is simple patch to
> libssh to fix its testsuite failure due to incorrect assumptions on how
> gcrypt versions look like (not necessarily [0-9]+.[0-9]+.[0-9]+").
> Suggested patch is here:
> http://lists.alioth.debian.org/pipermail/pkg-gnutls-maint/2017-September/008101.html
>
> I hope that helps.

It does; libssh builds fine with your simple patch[1].
Laurent, is it ok for you to use that patch and upload the resulting
~bpo9 package to backports?

> As libgcrypt has its share of security issues due to its usage scope I
> would strongly prefer to not have a bpo unless it is unavoidable.

I see.

Thanks for your help.

Cheers.


[1] http://debomatic-amd64.debian.net/distribution#stretch-backports/libssh/0.8.0~20170825.94fa1e38-1~bpo9+1/buildlog

-- 
Matteo F. Vescovi || Debian Developer
GnuPG KeyID: 4096R/0x8062398983B2CF7A
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 987 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/pkg-gnutls-maint/attachments/20171011/dbe1d3ba/attachment.sig>