Boost logo

Boost :

From: Robert Ramey (ramey_at_[hidden])
Date: 2024-04-10 19:19:18


On 4/10/24 11:36 AM, René Ferdinand Rivera Morell via Boost wrote:
> As Andrey said.. You confused two different build systems. But I'm
> replying solely for B2. :-)
>

I'm talking about the version which gets built when one invokes the
"bootstrap" script.

> Do you have a suggestion to encourage you to update to the up to date
> releases as they happen?

If there were some sort of versioning that might work. Suppose b2 boost
1.85 is all working. Suppose that b2 contains the check - "guaranteed
to work on boost builds up to version 1.85". When b2 starts up, it
checks the version of boost it's running on. If the current version of
b2 can't work with the current version of boost, it emits an helpful
error message and shuts down.

This actually begs the main question: Why is b2 coupled to the version
of Boost in the first place? I don't have time to argue this now.

My current version of booststrap genereates a b2 executable which throws
a seg fault at startup. Of course this should never occur. It suggests
that b2 needs more tests before being released.

>
>> But when I do, there's almost always some sort of issue which requires
>> going to the list or slack/boost. Eventually I get this to work and it's
>> quite satisfactory. Basically, the development process for b2 is not
>> resulting in a reliable product.
>
> Do you have suggestions as to what to change in the development process?

Independent test suite for b2.
Don't couple b2 with current version of Boost
>
> Just because you don't superficially see the progress doesn't mean
> it's not there.

I can only rely upon what I see.

Robert Ramey


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk