| 1 |
On 09/30/2015 01:29 PM, Kristian Fiskerstrand wrote: |
| 2 |
> On 09/30/2015 01:27 PM, hasufell wrote: |
| 3 |
>> On 09/30/2015 01:22 PM, Rich Freeman wrote: |
| 4 |
>>> On Wed, Sep 30, 2015 at 2:35 AM, Paweł Hajdan, Jr. |
| 5 |
>>> <phajdan.jr@g.o> wrote: |
| 6 |
>>>> On 9/29/15 3:32 PM, Rich Freeman wrote: |
| 7 |
> |
| 8 |
> .. |
| 9 |
> |
| 10 |
>>> Perhaps the in-between solution would be for forking upstreams |
| 11 |
>>> to preserve the same symbol names as long as the APIs are |
| 12 |
>>> identical, and change them when they are not. I don't really see |
| 13 |
>>> that having any more impact on downstream consumers than silently |
| 14 |
>>> changing the APIs and it would probably get rid of the symbol |
| 15 |
>>> collision problem. |
| 16 |
>>> |
| 17 |
> |
| 18 |
>> Again: can you take that to libressl mailing list or start another |
| 19 |
>> thread? |
| 20 |
> |
| 21 |
> |
| 22 |
> The way I see it this is relevant to the discussion at hand. Before |
| 23 |
> implementing any system wide change to support LibreSSL, in order to |
| 24 |
> avoid future issues, a proper cost/benefit analysis and discussion is |
| 25 |
> in order. |
| 26 |
> |
| 27 |
|
| 28 |
Since I am almost the only one working on it, I am not sure what you |
| 29 |
mean with that. |
| 30 |
|
| 31 |
If you have anything to add to the import discussion please do so. But |
| 32 |
this thread is not about the general course of LibreSSL, because we are |
| 33 |
not forking it. And I doubt that anyone here is going to do that, so I |
| 34 |
don't see how such discussion are relevant for us here. |
| 35 |
|
| 36 |
The relevant information that impacts us as distributors has already |
| 37 |
been added to this thread and the consequences are as follows: |
| 38 |
* virtual does not make sense |
| 39 |
* libressl USE flag is necessary |
| 40 |
* libressl has to conflict with openssl |
| 41 |
|
| 42 |
|
| 43 |
> Do we have an overview of what functionality and other pros (hereunder |
| 44 |
> security gains that is not fixed in OpenSSL) is gained by implementing |
| 45 |
> global LibreSSL support? |
| 46 |
> |
| 47 |
|
| 48 |
Yes, it is widely known: |
| 49 |
https://en.wikipedia.org/wiki/LibreSSL#Security_and_vulnerabilities |
| 50 |
|
| 51 |
> Or is this just increasing our maintenance, and security tracking, etc |
| 52 |
> burdens without any strong benefits? |
| 53 |
> |
| 54 |
|
| 55 |
It increases maintenance burden in the same way that clang/libav/qt5 |
| 56 |
support increases maintenance burden. |
Archives