> Warner Losh wrote:
>
> > In message <[EMAIL PROTECTED]> "Dan Langille"
writes:
> > : http://www.freebsd.org/handbook/kernelconfig-building.html
> >
> > This change should do the trick if I'm reading things right.
>
> It's still KERNEL= in -stable though I think. I think I'll just add a
> wh
Warner Losh wrote:
> In message <[EMAIL PROTECTED]> "Dan Langille" writes:
> : http://www.freebsd.org/handbook/kernelconfig-building.html
>
> This change should do the trick if I'm reading things right.
It's still KERNEL= in -stable though I think. I think I'll just add a
which says "in -curr
In message <[EMAIL PROTECTED]> void writes:
: On Tue, Jan 23, 2001 at 02:52:54PM -0700, Warner Losh wrote:
: >
: > I don't know how to generate a warning that isn't a false positive in
: > most cases.
:
: Hmm ... how about something like,
:
: "make buildkernel failed, please check /usr/src/UPDA
On Tue, Jan 23, 2001 at 02:52:54PM -0700, Warner Losh wrote:
>
> I don't know how to generate a warning that isn't a false positive in
> most cases.
Hmm ... how about something like,
"make buildkernel failed, please check /usr/src/UPDATING for relevant
information"
?
--
Ben
"I told Paddy
In message <[EMAIL PROTECTED]> void writes:
: I would suggest not trying to detect the condition but simply mentioning
: that it is a possible cause of failure and letting the user figure it out
: from there.
I don't know how to generate a warning that isn't a false positive in
most cases.
Warne
On Tue, Jan 23, 2001 at 02:33:52PM -0700, Warner Losh wrote:
> In message <[EMAIL PROTECTED]> void writes:
> : Is it possible/desirable to have make print a message upon failure of
> : the buildkernel target, suggesting that the user do a buildworld if
> : they haven't done so? I know that sounds
In message <[EMAIL PROTECTED]> void writes:
: Is it possible/desirable to have make print a message upon failure of
: the buildkernel target, suggesting that the user do a buildworld if
: they haven't done so? I know that sounds a bit more hand-holdish than
: FreeBSD tends to be, but I think it w
On Mon, Jan 22, 2001 at 01:11:15AM -0700, Warner Losh wrote:
>
> I've committed this change, as threatened late last week, since no one
> said not to.
Thank you.
> buildworld is still required acorss major releases, when binutils
> change, and when config's version changes. if buildkernel fai
Thanks. I'll use that for starters. I still think that part of the handbook
needs to be modified, along the lines of my original patch but excluding the
different build patterns, just to stop the confusion. I won't be able to
submit a PR for a few days. So if anyone wants to jump in, my old PR
In message <[EMAIL PROTECTED]> "Dan Langille" writes:
: http://www.freebsd.org/handbook/kernelconfig-building.html
This change should do the trick if I'm reading things right.
Warner
Index: chapter.sgml
===
RCS file:
/home/imp/Fre
On 22 Jan 2001, at 1:11, Warner Losh wrote:
> In message <[EMAIL PROTECTED]> Neil Blakey-Milner writes:
> : "make buildkernel" currently fails if a "make buildworld" has not
>
> I've committed this change, as threatened late last week, since no one
> said not to.
>
> buildworld is still require
In message <[EMAIL PROTECTED]> Neil Blakey-Milner writes:
: "make buildkernel" currently fails if a "make buildworld" has not
I've committed this change, as threatened late last week, since no one
said not to.
buildworld is still required acorss major releases, when binutils
change, and when con
> Like lots of people who use FreeBSD rather than tinker with it,
> I have never done "make any-kind-of-world" and never expect to.
> I create a kernel config with my stuff in it, and do config,
> make, make install. I trust this is not going to be broken?
This is not broken, and will not be bro
In message <[EMAIL PROTECTED]> Neil Blakey-Milner writes:
: I've kept on forgetting to apply a patch similar to this one.
I'm annoncing my intention to commit this patch this weekend unless
someone objects. It tilts the balance a little far in the dangerous
direction for the upgrade customers, b
On Thu, Jan 18, 2001 at 08:13:57PM -0700, Warner Losh wrote:
>
> I thought that crunchgen dealt with the .o's that were created in the
> buildworld process. It looks like it rebuilds them itself in a
> whacked out way to make the whole thing work.
No - it looks at the original Makefile to deter
On Fri, Jan 19, 2001 at 07:43:50PM +1300, Dan Langille wrote:
> > answered one too many questions that boil down to "just rebuild the
> > world and stop arguing with me, things will start to work".
>
> Either we change the handbook to make it clear, as I hope my patch[1]
> does, or we fix buildk
On Fri, Jan 19, 2001 at 07:43:50PM +1300, Dan Langille wrote:
> On 18 Jan 2001, at 20:13, Warner Losh wrote:
>
> Either we change the handbook to make it clear, as I hope my patch[1]
> does, or we fix buildkernel so it works as advertised in the handbook. I
> don't mind either solution.
>
Do
Dan Langille wrote:
> On 18 Jan 2001, at 20:13, Warner Losh wrote:
>
> > Still, I don't think it is too onerous a requirement that a buildworld
> > have happened first.
>
> I disagree. Unless you qualify the above, you're saying that if I install
> FreeBSD for the first time, in order to crea
On Thu 2001-01-18 (20:13), Warner Losh wrote:
> In message <[EMAIL PROTECTED]> Josef Karthauser writes:
> : On Thu, Jan 18, 2001 at 05:36:30PM -0700, Warner Losh wrote:
> : > In message <[EMAIL PROTECTED]> Josef Karthauser writes:
> : > : Hear hear. We had to back out the 'make buildkernel' withi
On 18 Jan 2001, at 19:33, Dima Dorfman wrote:
> Personally, I think that if there's no technical reason why
> buildkernel can't work without a prior buildworld assuming that the
> kernel you're building is of the same version that's currently
> running, trying to explain two different methods to
On 18 Jan 2001, at 20:13, Warner Losh wrote:
> Still, I don't think it is too onerous a requirement that a buildworld
> have happened first.
I disagree. Unless you qualify the above, you're saying that if I install
FreeBSD for the first time, in order to create a custom kernel, I need to
make
> The other reason to encourage it strongly is that there are too many
> binary incompatibilities with the kernel interface for some programs,
> even in -stable, so we'd want to encourage people to build and install
> both at the same time. I'd imagine that the same sort of argument
> would apply
In message <[EMAIL PROTECTED]> Josef Karthauser writes:
: On Thu, Jan 18, 2001 at 05:36:30PM -0700, Warner Losh wrote:
: > In message <[EMAIL PROTECTED]> Josef Karthauser writes:
: > : Hear hear. We had to back out the 'make buildkernel' within PicoBSD
: > : because there was no guarentee that th
On Thu, Jan 18, 2001 at 05:36:30PM -0700, Warner Losh wrote:
> In message <[EMAIL PROTECTED]> Josef Karthauser writes:
> : Hear hear. We had to back out the 'make buildkernel' within PicoBSD
> : because there was no guarentee that the user had ever done a make
> : buildworld. Additionally if you
In message <[EMAIL PROTECTED]> Neil Blakey-Milner writes:
: I've kept on forgetting to apply a patch similar to this one.
This patch looks good to me. However, it will fail when you are
trying to build accross an upgrade to gcc and/or binutils. Since this
is an infrequent event, and since it us
In message <[EMAIL PROTECTED]> Josef Karthauser writes:
: Hear hear. We had to back out the 'make buildkernel' within PicoBSD
: because there was no guarentee that the user had ever done a make
: buildworld. Additionally if you do an
: env MAKEOBJDIRPREFIX=/usr/myobj make buildkernel
: you
On Thu, Jan 18, 2001 at 10:13:15AM +0200, Neil Blakey-Milner wrote:
> Hi,
>
> I've kept on forgetting to apply a patch similar to this one.
>
> "make buildkernel" currently fails if a "make buildworld" has not
> previously been done on the machine (and still has the populated object
> environmen
Hi,
I've kept on forgetting to apply a patch similar to this one.
"make buildkernel" currently fails if a "make buildworld" has not
previously been done on the machine (and still has the populated object
environment) because OBJFORMAT_PATH is explicitly set to only use
${WORLDTMP}/usr/libexec.
28 matches
Mail list logo
