where are the testing docs ?

Dennis Clarke dclarke at blastwave.org
Mon May 10 17:43:45 UTC 2021

On 5/10/21 01:55, @lbutlr wrote:
> On 06 May 2021, at 09:57, Dennis Clarke via bind-users <bind-users at lists.isc.org> wrote:
>> I do NOT trust a build result where I had to go hacking into all the
>> Makefiles just to get it to build. You install without doing testing?
> That's a very strange definition of "hacking". Setting makefile [preferences and options is not in and way "hacking".

I realize you are being a jerk on purpose but regardless :

    1) 9.11.26 builds perfectly out of the box with no issues

    2) 9.11.27 fails to build with a bucket of undefined symbols

    3) 9.11.28 fails to build in the same manner

    4) 9.11.29 why waste my time looking here ?

    5) 9.11.30 does not even exist ... please play again

    6) 9.11.31 fails to build with a bucket of undefined symbols

    7) dig around madly into 9.11.26 to see if *something* has gone
       wonky thereafter ... rebuild it and watch everything "just work"

    8) change some compiler flags, look at the CPPFLAGS and begin
       digging into the Makefiles to see where things have gone bork

    9) find that the Makefile in bin/tools is in fact bork bork bork

   10) compare the Makefile in bin/tools with the results from 9.11.26

   11) find possible bork botk bork and begin to hack in some silly
       edits to get past the bin/tools portion of this mess

   12) that works and now other things break, so begin a pile of sed
       and grep and awk and such over ALL the Makefiles everywhere and
       determine that in fact yes they are all borked slightly

I call all of those four days of work a pile of hack. On an old legacy
platform that no one wants to keep running anymore, and it just keeps
running and running. I don't know what you call it. I just say that
releases after 9.11.26 are borked. However only slightly and in places
no one would look at anyways.

Dennis Clarke
UNIX and Linux spoken
GreyBeard and suspenders optional

More information about the bind-users mailing list