|
Boost : |
From: Robert Ramey (ramey_at_[hidden])
Date: 2023-11-06 19:44:11
On 11/6/23 10:15 AM, René Ferdinand Rivera Morell via Boost wrote:
> On Mon, Nov 6, 2023 at 10:17â¯AM Robert Ramey via Boost
> <boost_at_[hidden]> wrote:
>>
>> Apparently, no one is responsable for maintaining the test script.
>> Maybe can fix that.
>
> Indeed...https://lists.boost.org/Archives/boost//2017/07/237250.php
>
I don't remember seeing this, but I guess it addresses my original question.
I did try to promote the idea that boost tools be subjected to the same
development requirements as boost libraries regarding: reviews, tests,
documentation, etc.... But the idea didn't catch on.
I also tried to organize a effort to get take a more formal approach to
implementation of CMake support. That idea also fizzled. Peter stepped
up and unveiled solution as a fait compli implement CMakeLists.txt in
all the boost libraries. So job was done. Or was it. I already had
CMake build/test implemented in my libraries so I assumed that the
official one was more or less similar. I didn't think about it much
after that. Recently I came to understand that the current boost usage
of CMake doesn't include the tests. Wow - what is the point of that?
If we eliminated b2 and depended totally on CMake then ... we be running
no tests and we'd have no disputes about testing/CI . The normal boost
development procededure would have caught this requirement oversite at
the beginning.
Sooo... I renew my periodic and probably pointless exhortation to start
subjecting our tools to the same standard that we demand of of our
libraries and implement procedures to guarantee that.
Robert Ramey
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk