Page MenuHomeFreeBSD

Fix buildworld when cross-building and MK_INFO == yes
ClosedPublic

Authored by bapt on Jan 2 2015, 5:03 AM.
Tags
None
Referenced Files
Unknown Object (File)
Thu, Apr 25, 7:51 PM
Unknown Object (File)
Thu, Apr 25, 7:48 PM
Unknown Object (File)
Mar 5 2024, 12:47 PM
Unknown Object (File)
Feb 23 2024, 9:12 AM
Unknown Object (File)
Feb 5 2024, 12:20 PM
Unknown Object (File)
Jan 8 2024, 7:44 PM
Unknown Object (File)
Dec 31 2023, 8:24 AM
Unknown Object (File)
Dec 31 2023, 8:24 AM
Subscribers

Details

Summary

Fix buildworld when cross-building and MK_INFO == yes

build-tools doesn't have consistent install logic, so to avoid disrupting that
I chose to move compiling texinfo from build-tools to bootstrap-tools. This is
a misnomer and I'd like to resolve that item sometime in the near future if
possible.

This commit adds a new knob, MK_INFO_SUPPORT, and builds texinfo depending on
whether or not that knob is set

Test Plan
% make tinderbox -DMAKE_JUST_WORLDS JFLAG=-j6 TARGETS="mips powerpc" \
               TARGET_ARCHES_mips=mipsel TARGET_ARCHES_powerpc=powerpc

Diff Detail

Repository
rS FreeBSD src repository - subversion
Lint
No Lint Coverage
Unit
No Test Coverage

Event Timeline

ngie retitled this revision from to Fix buildworld when cross-building and MK_INFO == yes.
ngie updated this object.
ngie edited the test plan for this revision. (Show Details)
ngie added a reviewer: dim.
ngie added subscribers: emaste, imp.
ngie edited edge metadata.
ngie edited the test plan for this revision. (Show Details)

Update the diff post-bapt's commits to remove texinfo from base on head

This diff doesn't make sense on head, but will make sense as a direct
commit to stable/9 and stable/10

Remove orphaned build knob from tools/build/options

OBE. You should abandon this, no?

Even in -stable, you need to define the _SUPPORT variables correctly....

Makefile.inc1
236

This is bogus. You need to define in the same file that MK_INFO is defined.

249–252

This isn't needed.

1304

This doesn't appear to be defined in src.opts.mk or bsd.opts.mk. Why not?

It seems like some revisions weren't backported to stable/10 for setting MK_INFO=no / -DNO_INFO, so I'm going to abandon/simplify this revision.

Makefile.inc1
236

Ok

1304

I was trying to keep the diff minimal, but I can add it to bsd.own.mk on stable/10.

May be it should not be named MK_INFO_SUPPORT as it is not a real option but something to workaround the fact that MK_INFO=no is part of NO_FUN

Rename MK_INFO_SUPPORT into an internal _BOOTSTRAP_MAKEINFO or something like this so it does not appear as an option but really looks like an internal variable

@imp what do you think?

In D1410#16, @bapt wrote:

May be it should not be named MK_INFO_SUPPORT as it is not a real option but something to workaround the fact that MK_INFO=no is part of NO_FUN

Rename MK_INFO_SUPPORT into an internal _BOOTSTRAP_MAKEINFO or something like this so it does not appear as an option but really looks like an internal variable

@imp what do you think?

I don't think it's necessary because bootstrap-tools isn't build with MK_INFO=no / -DNO_INFO on stable/10...

bapt reclaimed this revision.
bapt edited reviewers, added: ngie; removed: bapt.

Here is what I was talking about

From my testing MK_INFO is always set to know on stable/10

bapt removed a subscriber: imp.
Makefile.inc1
236

This isn't needed. Just put it in BMAKE.

251

Please capitalize _bootstrap_makeinfo for consistency.

1315

Why ?=?

1348

Please put _texinfo before that so it's easier to MFC changes and is in alphabetical order.

Makefile.inc1
236

Please ignore this comment. It's needed...

I'll update the patch

Makefile.inc1
236

This is needed otherwise _bootstrap_makeinfo is not defined while hitting the .if the first time resulting in a make(1) failure

Update according to @ngie comment

bapt added a reviewer: bapt.

@imp approved by irc

This revision is now accepted and ready to land.Jan 2 2015, 9:16 PM