Re: Does "makeoptions WITH_CTF=yes" actually work?

From: Alexander Leidinger <netchild_at_FreeBSD.org>
Date: Thu, 22 Apr 2010 11:02:30 +0200
Quoting Navdeep Parhar <nparhar_at_gmail.com> (from Thu, 22 Apr 2010  
01:33:22 -0700):

> On Thu, Apr 22, 2010 at 09:44:47AM +0200, Alexander Leidinger wrote:
>> Quoting Navdeep Parhar <nparhar_at_gmail.com> (from Wed, 21 Apr 2010
>> 18:23:33 -0700):
>>
>> >Your patch works for me, thanks.  There is just one more problem  
>> with the CTF
>>
>> I found a case where it does not work (not kernel related), I have
>> another one which works better.
>>
>> >generation that needs to be fixed:
>> >
>> >http://lists.freebsd.org/pipermail/freebsd-hackers/2009-April/028244.html
>> >
>> >While you're here can you take a look at the patch in that email too?
>>
>> Looks good in concept, but the CTFMERGE line needs the same
>> treatment like all the other ones in the .mk files. I want to commit
>> a suitable change today.
>
> Does "same treatment" mean it should run silently too?  My personal

Yes.

> opinion is that all ctfconvert and ctfmerge commands should show up in
> the output of make iff they run.  I believe that used to be the case
> before r206082.

Correct, and I agree.

The problem is the inverse-logic construct for the check if it shall  
be run or not which is consistent with all places where this is done.  
There is no easy way to only display a part of the command which is  
executed. It was decided by ... (jhb and rwatson?) to not display at  
all while we still have the default to without ctf (without the _at_ we  
will even have some display of something with ctfconvert or ctfmerge  
in the name, when no ctf info is put into the files). They want to  
have the default to with ctf when it is ready/stable enough. I assume  
that at this point the commands get shown again, as the handling of  
the with/without CTF stuff can be simplified in this case. It is not  
as easy as all the other with/without stuff we have, due to the fact  
that parts of the ctf stuff is in sys.mk, which is read before every  
other file.

Currently I want to finish the edge cases we noticed in a *consistent*  
way, to have something which is giving us stable behavior. After that  
I will go out of the loop and anyone is free to try/optimize what he  
wants (as long as I can get a kernel compiled with CTF info without  
much hassle, I do not care much what is done and how).

HTH,
Alexander.

-- 
http://www.Leidinger.net  Alexander _at_ Leidinger.net: PGP ID = B0063FE7
http://www.FreeBSD.org     netchild _at_ FreeBSD.org  : PGP ID = 72077137
A hermit is a deserter from the army of humanity.
Received on Thu Apr 22 2010 - 07:02:40 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:02 UTC