BIND 10 #1640: incomplete pc file for botan and wrong filename for botan-config

[BIND 10 Development]

#1640: incomplete pc file for botan and wrong filename for botan-config
-------------------------------------+-------------------------------------
                   Reporter:  jreed  |                 Owner:  jreed
                       Type:         |                Status:  reviewing
  defect                             |             Milestone:
                   Priority:  low    |  Sprint-20120320
                  Component:  build  |            Resolution:
  system                             |             Sensitive:  0
                   Keywords:         |           Sub-Project:  Core
            Defect Severity:  N/A    |  Estimated Difficulty:  4
Feature Depending on Ticket:         |           Total Hours:  0
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------
Changes (by jelte):

 * owner:  kevin_tes => jreed


Comment:

 Oh ok apparently I misunderstood part of the problem, i thought it was
 about the -config script lying, which we have also seen... (but I think
 that has since been fixed by the maintainer, and that system did not have
 a .pc file for botan)

 pkg-config was preferred since that is the official successor intended to
 solve this config script crap :p

 Of course if it is lying, it is a bug that should be filed with the
 maintainer, but still something we need to work around. Not sure whether
 we should prefer the -config approach or just scratch the entire use of
 pkgconfig. Has anyone seen a system where the .pc file is installed but
 the config script is not?

 Reversing the checks for now (first scripts, then pkgconfig).

 Also copied your suggested error message.

 BTW, the triple repeat there is some autoconf magic done by
 AC_CHECK_HEADERS I think.

 Reassigning ticket to jeremy with the request to pull update and try again
 :)

 BTW, if this fixes it for some, but then breaks it again on other systems,
 there is yet another approach we can take, and that is to try every
 possible combination, and actually do a compile test for each, instead of
 just assuming we get correct data. And if THAT still fails we could then
 also include switching some well-known flags on an off. But I really don't
 want to go there :)

-- 
Ticket URL: <http://bind10.isc.org/ticket/1640#comment:21>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development