Re: CVS commit: pkgsrc/devel/readline

[John Marino <netbsd%marino.st@localhost>]

On 11/25/2016 17:38, David Holland wrote:
> On Fri, Nov 25, 2016 at 01:02:50PM -0600, John Marino wrote:
>  > > > Once the condition is required, then changing the include is in
>  > > > the noise.
>
> No, because if you absolutely must must have Dragonfly-specific logic
> to avoid use of termcap, the right place for it is in mk/termcap.bl3.mk
> so it applies to all termcap users.
>
> However, from what you're describing the thing you want is to put a
> PREFER statement in mk/platform/DragonFly.mk to avoid using the
> builtin termcap on Dragonfly because it's broken.

The BUILTIN/PREFER logic is also broken.
jperkin suggested a fix but I'm not in position to test it yet.  It 
requires creating a new user, checking out a pristine copy of pkgsrc and 
bootstrapping.  I'll probably try it but I'm not there yet.

> Spewing adhoc logic all over everywhere in place of making the proper
> changes is destructive in all but the very short term. One of the
> reasons you're getting pushback is that you have a history of doing
> this and shouting down people who object.

Well, I take issue with all three characterizations.
1) This was the only place I targeted. Nothing else seems to fail with 
the platform/DragonFly.mk patch I have locally.
2) Even when I was committing frequently, I don't believe I was hanging 
hard to a bad solution where there was a better one.
3) Frankly, most of the people that have been objecting have been doing 
so in an extremely aggressive and unfriendly manner and often providing 
"insight" without all the facts.  I'm the one getting attacked, not the 
other way around.

Right now your tone appropriate and so my intent is to match it.




>  > A) That implies believing that building it with termcap and falling back to
>  > ncurses is more correct than building with ncurses and falling to termcap.
>
> For programs that use the termcap API, of course it is.

I actually couldn't name an example of modern software that links with 
termcap exclusively and not ncurses.


>  > B) That requires changing several things, all out of scope of this, and
>  > requiring having to take point to fix all of them and get approval.
>
> I don't know about where you come from but around here we prefer
> fixing things correctly the first time.

well, I spoke too soon.  That comment is inappropriate "where I come from".


>  > The status of other platforms is irrelevant.  The fact is that as
>  > devel/readline was, it didn't build on DF.
>
> > From what you've described, this really sounds like Dragonfly's bug.

The detection mechanism on pkgsrc failed because they didn't consider 
all possibilities, made some assumptions that aren't guaranteed to be 
true, and failed to check for the presence of all required components. 
On top of that, the Ports framework did not have the same issues on the 
same system.  Does that really sound like a "DragonFly" bug to you?  It 
doesn't to me.

John



---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus