On Fri, Mar 14, 2014 at 9:12 AM, Dave Page <dpage@pgadmin.org> wrote:
> On Thu, Mar 13, 2014 at 7:24 PM, Andreas 'ads' Scherbaum
> <adsmail@wars-nicht.de> wrote:
>> On 03/04/2014 09:29 AM, Andreas 'ads' Scherbaum wrote:
>>>
>>> On 03/03/2014 02:44 PM, Dave Page wrote:
>>>>
>>>> On Mon, Mar 3, 2014 at 1:05 PM, Andreas 'ads' Scherbaum
>>>> <adsmail@wars-nicht.de> wrote:
>>>>>
>>>>> On 03/03/2014 11:32 AM, Dave Page wrote:
>>>>>>
>>>>>>
>>>>>> On Mon, Mar 3, 2014 at 10:20 AM, Andreas 'ads' Scherbaum
>>>>>> <adsmail@wars-nicht.de> wrote:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Yes, I know how to fix this. My question is: why is "configure" not
>>>>>>> checking
>>>>>>> for this library, when it is required to build pgAdmin3?
>>>>>>
>>>>>>
>>>>>>
>>>>>> Noone wrote the code I guess. Patches welcome :-)
>>>>>
>>>>>
>>>>>
>>>>> Just checking the code: why is it even requiring "crypto" when compiling
>>>>> without --with-openssl?
>>>>
>>>>
>>>> That's actually the likely reason why we don't do that test. It seems
>>>> to be required because we're linking with libpq which is linked with
>>>> OpenSSL, apps linking to it (on some platforms) also need to link with
>>>> libcypto. It may also be that an element of bitrot has crept into that
>>>> configure option - I very much doubt anyone has tested without SSL
>>>> support in a long time.
>>>
>>>
>>> Given what you said, attached is a patch which requires "openssl/ssl.h"
>>> to be available. If not, configure bails out with an error message.
>>
>>
>> Any feedback regarding this patch?
>
> Not from me at least (yet). I've been up to my eyeballs all week at work.
I took a quick look at this (and then passed the buck to Ashesh for
additional review and commit :-p ) - I think it's fine except that the
test needs to be inside the OpenSSL block above. See attached.
--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake
EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company