Discussion:
[PREFIX] Interix _ALL_SOURCE
(too old to reply)
Fabian Groffen
2008-07-01 14:49:13 UTC
Permalink
While compiling Mutt on Interix, I was confronted with flock + constants
not being defined. Further inspection of the header file (manpage said
it should exist in fcntl.h) revealed the need for _ALL_SOURCE. Many
ebuilds just include flag-o-matic to do append-flags -D_ALL_SOURCE on
Interix. Not being very amused by the beauty of that solution, and
looking at the insanity of the Interix headers, what's against just
adding -D_ALL_SOURCE in the Interix' profile.bashrc file (with
encompassing logic not to add it multiple times)? Wouldn't that be much
saner (and cleaner in the ebuilds)? Are there packages known to break
when _ALL_SOURCE is defined?
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-02 06:23:54 UTC
Permalink
Post by Fabian Groffen
While compiling Mutt on Interix, I was confronted with flock +
constants
not being defined. Further inspection of the header file (manpage said
it should exist in fcntl.h) revealed the need for _ALL_SOURCE. Many
ebuilds just include flag-o-matic to do append-flags -D_ALL_SOURCE on
Interix. Not being very amused by the beauty of that solution, and
looking at the insanity of the Interix headers, what's against just
adding -D_ALL_SOURCE in the Interix' profile.bashrc file (with
encompassing logic not to add it multiple times)? Wouldn't that be much
saner (and cleaner in the ebuilds)? Are there packages known to break
when _ALL_SOURCE is defined?
I guess there where packages refusing to build with _ALL_SOURCE, but I can't think of a single one. After all it shouldn't be too much work to get those to build with _ALL_SOURCE then I think.

So from my POV, enabling should be ok. BUT: that would mean that packages will possibly change behavior, which means that eventually nothing has to stay compatible to the current builds. This is not a blocker I think, but we have to remember that it would be better to re-bootstrap most of the things, after this change, to keep everything consistent.

Cheers, Markus
Post by Fabian Groffen
--
Fabian Groffen
Gentoo on a different level
--
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-02 07:02:22 UTC
Permalink
Post by Markus Duft
Post by Fabian Groffen
While compiling Mutt on Interix, I was confronted with flock + constants
not being defined. Further inspection of the header file (manpage said
it should exist in fcntl.h) revealed the need for _ALL_SOURCE. Many
ebuilds just include flag-o-matic to do append-flags -D_ALL_SOURCE on
Interix. Not being very amused by the beauty of that solution, and
looking at the insanity of the Interix headers, what's against just
adding -D_ALL_SOURCE in the Interix' profile.bashrc file (with
encompassing logic not to add it multiple times)? Wouldn't that be much
saner (and cleaner in the ebuilds)? Are there packages known to break
when _ALL_SOURCE is defined?
I guess there where packages refusing to build with _ALL_SOURCE, but I can't think of a single one. After all it shouldn't be too much work to get those to build with _ALL_SOURCE then I think.
The inverse (strip-flags -D_ALL_SOURCE) should be possible, but then
happening quite a lot less, I think...
Post by Markus Duft
So from my POV, enabling should be ok. BUT: that would mean that packages will possibly change behavior, which means that eventually nothing has to stay compatible to the current builds. This is not a blocker I think, but we have to remember that it would be better to re-bootstrap most of the things, after this change, to keep everything consistent.
What do you mean? I would just enable it, rebuild system (wait for +- 2
days) if that's ok, commit it, and then remove all append-flags
-D_ALL_SOURCE stuff. People having a prefix shouldn't notice that,
should they? New packages they install just should build, and I hope no
upstream is so stupid to put #ifdef _ALL_SOURCE in their headers like
Interix does. Quick grepping doesn't show anything like a changing
definition or something.
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-02 07:38:30 UTC
Permalink
<snip>
Post by Markus Duft
Post by Markus Duft
I guess there where packages refusing to build with _ALL_SOURCE, but
I can't think of a single one. After all it shouldn't be too much work
to get those to build with _ALL_SOURCE then I think.
The inverse (strip-flags -D_ALL_SOURCE) should be possible, but then
happening quite a lot less, I think...
Post by Markus Duft
So from my POV, enabling should be ok. BUT: that would mean that
packages will possibly change behavior, which means that eventually
nothing has to stay compatible to the current builds. This is not a
blocker I think, but we have to remember that it would be better to re-
bootstrap most of the things, after this change, to keep everything
consistent.
What do you mean? I would just enable it, rebuild system (wait for +- 2
days) if that's ok, commit it, and then remove all append-flags
-D_ALL_SOURCE stuff. People having a prefix shouldn't notice that,
should they? New packages they install just should build, and I hope no
upstream is so stupid to put #ifdef _ALL_SOURCE in their headers like
Interix does. Quick grepping doesn't show anything like a changing
definition or something.
I meant that changing the define makes various configure scripts find different things, and thus may enable (or disable) different code paths, which _could_ in some stupid cases confuse already merged packages, or the new ones, since it expects something from the merged packages they don't provide, since at build time those things weren't there.

Still I guess we should enable it, and just see what's happening.

Cheers, Markus
Post by Markus Duft
--
Fabian Groffen
Gentoo on a different level
--
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-02 07:47:43 UTC
Permalink
Post by Markus Duft
Post by Fabian Groffen
What do you mean? I would just enable it, rebuild system (wait for +- 2
days) if that's ok, commit it, and then remove all append-flags
-D_ALL_SOURCE stuff. People having a prefix shouldn't notice that,
should they? New packages they install just should build, and I hope no
upstream is so stupid to put #ifdef _ALL_SOURCE in their headers like
Interix does. Quick grepping doesn't show anything like a changing
definition or something.
I meant that changing the define makes various configure scripts find different things, and thus may enable (or disable) different code paths, which _could_ in some stupid cases confuse already merged packages, or the new ones, since it expects something from the merged packages they don't provide, since at build time those things weren't there.
Ok, good point. Then I guess it's either you or me doing it, and just
aligning it, that when it's pushed into the tree, we
a) make a new bootstrap snapshot
b) upload new installers using the newly compiled packages

I can do the compilation of system, but maybe that doesn't make much
sense since you'll have to build the installers/binpkgs anyway?
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Mathias Laurin
2008-07-02 08:28:20 UTC
Permalink
Hi,

I found that I need to add the -c option to NROFF (i.e., get rid of
the colours) on my system in $EPREFIX/etc/man.conf; without this
options, escape sequences appear in the man page instead of colours,
making them unreadable. I tried from zsh and bash, I also tried
getting rid of my old ~/.Xdefaults and the -c option is present in /
etc/man.conf, the original OS X man.conf so I think it is real.

Here is the diff:

17:25 /Users/laurin/Library/Gentoo/etc [0]1 % diff -u man.conf{.orig,}
--- man.conf.orig 2008-07-02 17:06:44 +0900
+++ man.conf 2008-07-02 17:25:37 +0900
@@ -93,8 +93,8 @@
# causes problems, add the -c option to TROFF, NROFF, JNROFF.
#
TROFF /Users/laurin/Library/Gentoo/usr/bin/groff -Tps -mandoc
-NROFF /Users/laurin/Library/Gentoo/usr/bin/nroff -mandoc
-JNROFF /Users/laurin/Library/Gentoo/usr/bin/groff -Tnippon -mandocj
+NROFF /Users/laurin/Library/Gentoo/usr/bin/nroff -mandoc -c
+JNROFF /Users/laurin/Library/Gentoo/usr/bin/groff -Tnippon -mandocj -c
EQN /Users/laurin/Library/Gentoo/usr/bin/geqn -Tps
NEQN /Users/laurin/Library/Gentoo/usr/bin/geqn -Tlatin1
JNEQN /Users/laurin/Library/Gentoo/usr/bin/geqn -Tnippon
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-02 08:40:29 UTC
Permalink
This sounds like bug http://bugs.gentoo.org/show_bug.cgi?id=102800
Hi,
I found that I need to add the -c option to NROFF (i.e., get rid of the
colours) on my system in $EPREFIX/etc/man.conf; without this options,
escape sequences appear in the man page instead of colours, making them
unreadable. I tried from zsh and bash, I also tried getting rid of my old
~/.Xdefaults and the -c option is present in /etc/man.conf, the original
OS X man.conf so I think it is real.
17:25 /Users/laurin/Library/Gentoo/etc [0]1 % diff -u man.conf{.orig,}
--- man.conf.orig 2008-07-02 17:06:44 +0900
+++ man.conf 2008-07-02 17:25:37 +0900
@@ -93,8 +93,8 @@
# causes problems, add the -c option to TROFF, NROFF, JNROFF.
#
TROFF /Users/laurin/Library/Gentoo/usr/bin/groff -Tps -mandoc
-NROFF /Users/laurin/Library/Gentoo/usr/bin/nroff -mandoc
-JNROFF /Users/laurin/Library/Gentoo/usr/bin/groff -Tnippon -mandocj
+NROFF /Users/laurin/Library/Gentoo/usr/bin/nroff -mandoc -c
+JNROFF /Users/laurin/Library/Gentoo/usr/bin/groff -Tnippon -mandocj -c
EQN /Users/laurin/Library/Gentoo/usr/bin/geqn -Tps
NEQN /Users/laurin/Library/Gentoo/usr/bin/geqn -Tlatin1
JNEQN /Users/laurin/Library/Gentoo/usr/bin/geqn -Tnippon
--
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Mathias Laurin
2008-07-20 03:16:50 UTC
Permalink
Yes exactly, I should have checked bugzie before posting...

Cheers,
Mathias
Post by Fabian Groffen
This sounds like bug http://bugs.gentoo.org/show_bug.cgi?id=102800
Hi,
I found that I need to add the -c option to NROFF (i.e., get rid of the
colours) on my system in $EPREFIX/etc/man.conf; without this options,
escape sequences appear in the man page instead of colours, making them
unreadable. I tried from zsh and bash, I also tried getting rid of my old
~/.Xdefaults and the -c option is present in /etc/man.conf, the original
OS X man.conf so I think it is real.
17:25 /Users/laurin/Library/Gentoo/etc [0]1 % diff -u man.conf{.orig,}
--- man.conf.orig 2008-07-02 17:06:44 +0900
+++ man.conf 2008-07-02 17:25:37 +0900
@@ -93,8 +93,8 @@
# causes problems, add the -c option to TROFF, NROFF, JNROFF.
#
TROFF /Users/laurin/Library/Gentoo/usr/bin/groff -Tps -mandoc
-NROFF /Users/laurin/Library/Gentoo/usr/bin/nroff -mandoc
-JNROFF /Users/laurin/Library/Gentoo/usr/bin/groff -Tnippon -mandocj
+NROFF /Users/laurin/Library/Gentoo/usr/bin/nroff -mandoc -c
+JNROFF /Users/laurin/Library/Gentoo/usr/bin/groff -Tnippon -mandocj -c
EQN /Users/laurin/Library/Gentoo/usr/bin/geqn -Tps
NEQN /Users/laurin/Library/Gentoo/usr/bin/geqn -Tlatin1
JNEQN /Users/laurin/Library/Gentoo/usr/bin/geqn -Tnippon
--
--
Fabian Groffen
Gentoo on a different level
--
Markus Duft
2008-07-02 08:24:57 UTC
Permalink
Post by Fabian Groffen
Post by Markus Duft
Post by Fabian Groffen
What do you mean? I would just enable it, rebuild system (wait for
+- 2
Post by Markus Duft
Post by Fabian Groffen
days) if that's ok, commit it, and then remove all append-flags
-D_ALL_SOURCE stuff. People having a prefix shouldn't notice that,
should they? New packages they install just should build, and I
hope no
Post by Markus Duft
Post by Fabian Groffen
upstream is so stupid to put #ifdef _ALL_SOURCE in their headers
like
Post by Markus Duft
Post by Fabian Groffen
Interix does. Quick grepping doesn't show anything like a changing
definition or something.
I meant that changing the define makes various configure scripts find
different things, and thus may enable (or disable) different code
paths, which _could_ in some stupid cases confuse already merged
packages, or the new ones, since it expects something from the merged
packages they don't provide, since at build time those things weren't
there.
Ok, good point. Then I guess it's either you or me doing it, and just
aligning it, that when it's pushed into the tree, we
a) make a new bootstrap snapshot
b) upload new installers using the newly compiled packages
I can do the compilation of system, but maybe that doesn't make much
sense since you'll have to build the installers/binpkgs anyway?
*sigh* :) okok... i'll do it then. Where can I put the _ALL_SOURCE definition to?
Post by Fabian Groffen
--
Fabian Groffen
Gentoo on a different level
--
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-02 08:37:51 UTC
Permalink
Post by Markus Duft
Post by Fabian Groffen
I can do the compilation of system, but maybe that doesn't make much
sense since you'll have to build the installers/binpkgs anyway?
*sigh* :) okok... i'll do it then. Where can I put the _ALL_SOURCE definition to?
:)
Ok, I'll create the patch for the profile :)
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-14 13:16:39 UTC
Permalink
Post by Markus Duft
Post by Markus Duft
Post by Fabian Groffen
I can do the compilation of system, but maybe that doesn't make
much
Post by Markus Duft
Post by Fabian Groffen
sense since you'll have to build the installers/binpkgs anyway?
*sigh* :) okok... i'll do it then. Where can I put the _ALL_SOURCE
definition to?
:)
Ok, I'll create the patch for the profile :)
Hi!

Just wanted to ask what's the state of the patch... Have you done anything yet?

Cheers, Markus
Post by Markus Duft
--
Fabian Groffen
Gentoo on a different level
--
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-14 13:33:15 UTC
Permalink
Post by Markus Duft
Post by Markus Duft
Post by Markus Duft
Post by Fabian Groffen
I can do the compilation of system, but maybe that doesn't make
much
Post by Markus Duft
Post by Fabian Groffen
sense since you'll have to build the installers/binpkgs anyway?
*sigh* :) okok... i'll do it then. Where can I put the _ALL_SOURCE
definition to?
:)
Ok, I'll create the patch for the profile :)
Hi!
Just wanted to ask what's the state of the patch... Have you done anything yet?
Uhm, no I fail. Probably because I got stuck when Portage broke on
Interix (well, it was already broken, but some new issues with poll()
were introduced, should be fixed now) and I haven't yet gotten the
courage to actually unbreak it.
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-14 13:40:21 UTC
Permalink
Post by Fabian Groffen
Post by Markus Duft
Post by Fabian Groffen
Ok, I'll create the patch for the profile :)
Just wanted to ask what's the state of the patch... Have you done anything yet?
Uhm, no I fail. Probably because I got stuck when Portage broke on
Interix (well, it was already broken, but some new issues with poll()
were introduced, should be fixed now) and I haven't yet gotten the
courage to actually unbreak it.
How about this patch?

Index: profile.bashrc
===================================================================
--- profile.bashrc (revision 27541)
+++ profile.bashrc (working copy)
@@ -1,3 +1,8 @@
+# One basically always wants -D_ALL_SOURCE, it's some stupid restriction
+# to hide useful stuff
+[[ ${CFLAGS} != *"-D_ALL_SOURCE"* ]] && CFLAGS="-D_ALL_SOURCE ${CFLAGS}"
+[[ ${CXXFLAGS} != *"-D_ALL_SOURCE"* ]] && CXXFLAGS="-D_ALL_SOURCE ${CXXFLAGS}"
+
# On interix, binary files (executables, shared libraries) in use
# cannot be replaced during merge.
# But it is possible to rename them and remove lateron when they are
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-14 14:05:53 UTC
Permalink
Post by Fabian Groffen
How about this patch?
Index: profile.bashrc
===================================================================
--- profile.bashrc (revision 27541)
+++ profile.bashrc (working copy)
@@ -1,3 +1,8 @@
+# One basically always wants -D_ALL_SOURCE, it's some stupid
restriction
+# to hide useful stuff
+[[ ${CFLAGS} != *"-D_ALL_SOURCE"* ]] && CFLAGS="-D_ALL_SOURCE
${CFLAGS}"
+[[ ${CXXFLAGS} != *"-D_ALL_SOURCE"* ]] && CXXFLAGS="-D_ALL_SOURCE ${CXXFLAGS}"
+
Haubi suggested this small enhancement:

[[ " ${CFLAGS} " != *" -D_ALL_SOURCE "* ]] && .......

Another thought from haubi:

What if we just patch gcc to define _ALL_SOURCE as built-in. if we require it to be undefined, we can still use -U on the command line, right?

Cheers, Markus
Post by Fabian Groffen
# On interix, binary files (executables, shared libraries) in use
# cannot be replaced during merge.
# But it is possible to rename them and remove lateron when they are
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-14 14:21:24 UTC
Permalink
Post by Markus Duft
[[ " ${CFLAGS} " != *" -D_ALL_SOURCE "* ]] && .......
ok with me, smarter actually ;)
Post by Markus Duft
What if we just patch gcc to define _ALL_SOURCE as built-in. if we
require it to be undefined, we can still use -U on the command line,
right?
That last thing is a requirement of course (hence I put the define first
thing, so it can be overridden).

Further thoughts include:
- messing with gcc in general is no fun, alternative is the wrapper,
but could do it in the gcc spec language too, I guess
- how about the bootstrap phase, is setting -D_ALL_SOURCE easy there (I
think for my method it is)
- patched gcc (invocation) is nice since we always did all the work to
have a working compiler setup in/with Prefix without having to do
anything special (think of library paths to Prefix' libs)

Looking at the random thoughts above, I guess it means haubi is going to
fix it ;)
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-15 05:55:16 UTC
Permalink
Post by Fabian Groffen
Post by Markus Duft
[[ " ${CFLAGS} " != *" -D_ALL_SOURCE "* ]] && .......
ok with me, smarter actually ;)
Hm. Thinking about it, i guess i like this one better than defining _ALL_SOURCE inside gcc or the wrapper. Using strip-flags in ebuilds seems more readable, than append-flags -U_ALL_SOURCE... I think this is the way we should go.

If you agree, I'll bootstrap a prefix on interix 5.2, and also patch this (and all ebuilds)...

Cheers, Markus
Post by Fabian Groffen
Post by Markus Duft
What if we just patch gcc to define _ALL_SOURCE as built-in. if we
require it to be undefined, we can still use -U on the command line,
right?
That last thing is a requirement of course (hence I put the define first
thing, so it can be overridden).
- messing with gcc in general is no fun, alternative is the wrapper,
but could do it in the gcc spec language too, I guess
- how about the bootstrap phase, is setting -D_ALL_SOURCE easy there (I
think for my method it is)
- patched gcc (invocation) is nice since we always did all the work to
have a working compiler setup in/with Prefix without having to do
anything special (think of library paths to Prefix' libs)
Looking at the random thoughts above, I guess it means haubi is going to
fix it ;)
--
Fabian Groffen
Gentoo on a different level
--
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-15 09:36:46 UTC
Permalink
Post by Markus Duft
Post by Fabian Groffen
Post by Markus Duft
[[ " ${CFLAGS} " != *" -D_ALL_SOURCE "* ]] && .......
ok with me, smarter actually ;)
Hm. Thinking about it, i guess i like this one better than defining _ALL_SOURCE inside gcc or the wrapper. Using strip-flags in ebuilds seems more readable, than append-flags -U_ALL_SOURCE... I think this is the way we should go.
If you agree, I'll bootstrap a prefix on interix 5.2, and also patch this (and all ebuilds)...
Hmmm... what do you think about the argument that a user of Prefix GCC
has to set it now for her* own compilation tasks?



* yes we like female geeks
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Michael Haubenwallner
2008-07-15 11:39:28 UTC
Permalink
Post by Fabian Groffen
Post by Markus Duft
Post by Fabian Groffen
Post by Markus Duft
[[ " ${CFLAGS} " != *" -D_ALL_SOURCE "* ]] && .......
ok with me, smarter actually ;)
Hm. Thinking about it, i guess i like this one better than defining _ALL_SOURCE inside gcc or the wrapper. Using strip-flags in ebuilds seems more readable, than append-flags -U_ALL_SOURCE... I think this is the way we should go.
If you agree, I'll bootstrap a prefix on interix 5.2, and also patch this (and all ebuilds)...
Hmmm... what do you think about the argument that a user of Prefix GCC
has to set it now for her* own compilation tasks?
(-;
Basically: its her fault to use Windows.
But: she's clever and uses Gentoo Prefix.
;-)

Compared to what Interopsystems provides (set -D_ALL_SOURCE in $CFLAGS
shell variable at login), we should have gcc to define _ALL_SOURCE.

So to find out how many packages are broken _with_ _ALL_SOURCE defined,
we can add -D_ALL_SOURCE to C{,XX}FLAGS in profile, and remerge -e
world.

If that's not too many (or less than the number of now-working ones),
add it to gcc (or its wrapper, as the first cmdline-arg), so -U still
works if necessary.

/haubi/
--
Michael Haubenwallner
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-15 11:44:44 UTC
Permalink
Post by Michael Haubenwallner
(-;
Basically: its her fault to use Windows.
But: she's clever and uses Gentoo Prefix.
;-)
(: Don't you love her madly? :)
Post by Michael Haubenwallner
Compared to what Interopsystems provides (set -D_ALL_SOURCE in $CFLAGS
shell variable at login), we should have gcc to define _ALL_SOURCE.
So to find out how many packages are broken _with_ _ALL_SOURCE defined,
we can add -D_ALL_SOURCE to C{,XX}FLAGS in profile, and remerge -e
world.
If that's not too many (or less than the number of now-working ones),
add it to gcc (or its wrapper, as the first cmdline-arg), so -U still
works if necessary.
Ok, but that means we have an agreement here. First we need to figure
out if we can do it (and we can start with that in any way now/today)
and then if that appears to work, make it permanent in GCC.

Given that Markus appears to have a machine that can recompile world on
Interix within 2 days, I'd suggest him to try it. (Compared to my
machine which a) has a broken Portage and b) needs over a week to do it.)
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-15 13:22:57 UTC
Permalink
Post by Fabian Groffen
Post by Michael Haubenwallner
(-;
Basically: its her fault to use Windows.
But: she's clever and uses Gentoo Prefix.
;-)
(: Don't you love her madly? :)
q: i do ... :p
Post by Fabian Groffen
Post by Michael Haubenwallner
Compared to what Interopsystems provides (set -D_ALL_SOURCE in
$CFLAGS
Post by Michael Haubenwallner
shell variable at login), we should have gcc to define _ALL_SOURCE.
So to find out how many packages are broken _with_ _ALL_SOURCE
defined,
Post by Michael Haubenwallner
we can add -D_ALL_SOURCE to C{,XX}FLAGS in profile, and remerge -e
world.
If that's not too many (or less than the number of now-working ones),
add it to gcc (or its wrapper, as the first cmdline-arg), so -U still
works if necessary.
Ok, but that means we have an agreement here. First we need to figure
out if we can do it (and we can start with that in any way now/today)
and then if that appears to work, make it permanent in GCC.
Agreed.
Post by Fabian Groffen
Given that Markus appears to have a machine that can recompile world on
Interix within 2 days, I'd suggest him to try it. (Compared to my
machine which a) has a broken Portage and b) needs over a week to do it.)
I'm doing this with one of the prefixes targeting the setup CDs, so this should be pretty much the most essential stuff :) CFLAGS now should contain _ALL_SOURCE.

(2) mduft gentoo $ emerge -ave world

These are the packages that would be merged, in order:

Calculating dependencies... done!

[ebuild R ] sys-apps/portage-2.2.00.10815 USE="-build -doc -epydoc (-selinux)" LINGUAS="-pl" 0 kB
*** Portage will stop merging at this point and reload itself,
then resume the merge.
[ebuild R ] dev-util/pkgconfig-0.23 USE="(-hardened)" 1,009 kB
[ebuild R ] sys-apps/sed-4.1.5-r1 USE="-nls -static" 781 kB
[ebuild R ] sys-libs/ncurses-5.6-r2 USE="-debug -doc (-gpm) -minimal -nocxx -profile -trace (-unicode) (-bootstrap%) (-build%)" 2,353 kB
[ebuild R ] sys-libs/zlib-1.2.3-r01.1 416 kB
[ebuild R ] sys-devel/gnuconfig-20080123 40 kB
[ebuild R ] sys-devel/automake-wrapper-3-r1 0 kB
[ebuild R ] app-arch/lzma-utils-4.32.6 USE="-nocxx" 0 kB
[ebuild R ] sys-devel/flex-2.5.33-r3 USE="-nls -static" 680 kB
[ebuild R ] sys-devel/autoconf-wrapper-5 0 kB
[ebuild R ] app-text/build-docbook-catalog-1.2 4 kB
[ebuild R ] dev-util/gtk-doc-am-1.10-r1 395 kB
[ebuild R ] sys-apps/gawk-3.1.6 USE="-nls" 1,818 kB
[ebuild R ] app-arch/bzip2-1.0.5-r1 USE="-static" 0 kB
[ebuild R ] sys-apps/findutils-4.4.0 USE="-nls (-selinux) -static" 1,983 kB
[ebuild R ] sys-apps/diffutils-2.8.7-r2 USE="-nls -static" 1,038 kB
[ebuild R ] app-arch/unzip-5.52-r2 1,114 kB
[ebuild R ] app-admin/python-updater-0.5 0 kB
[ebuild R ] app-arch/cabextract-1.2 190 kB
[ebuild R ] sys-apps/grep-2.5.1a-r1 USE="-nls -pcre -static" 516 kB
[ebuild R ] app-arch/gzip-1.3.12-r1 USE="-nls -pic -static" 452 kB
[ebuild R ] app-arch/tar-1.20 USE="-nls -static" 1,868 kB
[ebuild R ] sys-devel/patch-2.5.9-r1 USE="-static" 198 kB
[ebuild R ] app-arch/cpio-2.9-r2 USE="-nls" 0 kB
[ebuild R ] sys-devel/make-3.81 USE="-nls -static" 1,125 kB
[ebuild R ] sys-devel/m4-1.4.11 USE="-examples -nls" 684 kB
[ebuild R ] gnome-base/gnome-common-2.20.0 USE="-debug" 63 kB
[ebuild R ] sys-apps/less-418 USE="(-unicode)" 488 kB
[ebuild R ] x11-themes/hicolor-icon-theme-0.10 33 kB
[ebuild R ] media-libs/libart_lgpl-2.3.20 USE="-debug" 296 kB
[ebuild R ] app-editors/nano-2.1.2-r1 USE="ncurses -debug -justify -minimal -nls -slang -spell (-unicode)" 0 kB
[ebuild R ] dev-util/cvs-1.11.21 USE="-doc (-emacs)" 2,862 kB
[ebuild R ] sys-devel/bison-2.3 USE="-nls -static" 1,055 kB
[ebuild R ] virtual/editor-0 0 kB
[ebuild R ] virtual/pager-0 0 kB
[ebuild R ] sys-devel/libperl-5.8.8-r2 USE="-berkdb -debug -gdbm -ithreads" 9,887 kB
[ebuild R ] dev-lang/perl-5.8.8-r5 USE="-berkdb -build -debug -doc -gdbm -ithreads -perlsuid" 0 kB
[ebuild R ] sys-apps/help2man-1.36.4 USE="-nls" 84 kB
[ebuild U ] perl-core/Storable-2.18 [2.16] 174 kB
[ebuild U ] perl-core/Test-Simple-0.80 [0.70] 80 kB
[ebuild R ] perl-core/Test-Harness-3.10 173 kB
[ebuild R ] perl-core/PodParser-1.35 96 kB
[ebuild R ] dev-libs/libtasn1-1.4 USE="-doc" 1,516 kB
[ebuild R ] sys-apps/texinfo-4.12 USE="-nls -static" 1,353 kB
[ebuild U ] virtual/perl-Test-Simple-0.80 [0.70] 0 kB
[ebuild U ] virtual/perl-Storable-2.18 [2.16] 0 kB
[ebuild R ] sys-devel/autoconf-2.61-r2 USE="(-emacs)" 0 kB
[ebuild R ] sys-devel/autoconf-2.13 434 kB
[ebuild U ] sys-apps/ed-0.9 [0.8] 67 kB
[ebuild R ] sys-apps/which-2.19 130 kB
[ebuild R ] sys-devel/automake-1.10.1-r1 897 kB
[ebuild R ] sys-devel/automake-1.5 515 kB
[ebuild R ] sys-devel/automake-1.7.9-r1 565 kB
[ebuild R ] sys-devel/automake-1.9.6-r2 748 kB
[ebuild R ] sys-devel/automake-1.8.5-r3 648 kB
[ebuild R ] dev-util/cscope-15.6-r2 USE="(-emacs)" 383 kB
[ebuild R ] sys-devel/libtool-1.5.26 USE="-vanilla" 2,893 kB
[ebuild R ] x11-misc/util-macros-1.1.6-r1 48 kB
[ebuild R ] sys-apps/coreutils-6.12 USE="-acl -nls (-selinux) -static -vanilla -xattr" 0 kB
[ebuild R ] dev-libs/expat-2.0.1 436 kB
[ebuild R ] dev-libs/popt-1.13 USE="-nls" 667 kB
[ebuild R ] app-shells/bash-3.2_p39 USE="-afs -bashlogger -examples% -nls -plugins -vanilla" 2,582 kB
[ebuild R ] dev-libs/libiconv-1.12 0 kB
[ebuild R ] media-libs/libpng-1.2.27 520 kB
[ebuild R ] media-libs/jpeg-6b-r8 621 kB
[ebuild R ] media-libs/audiofile-0.2.6-r4 0 kB
[ebuild R ] app-text/sgml-common-0.6.3-r5 75 kB
[ebuild R ] dev-libs/libgpg-error-1.6 USE="-nls" 374 kB
[ebuild R ] sys-libs/readline-5.2_p12-r1 2,018 kB
[ebuild R ] virtual/libiconv-0 0 kB
[ebuild R ] dev-perl/XML-Parser-2.36 225 kB
[ebuild R ] net-misc/rsync-3.0.3 USE="-acl -iconv (-ipv6) -static -xattr -xinetd" 0 kB
[ebuild R ] app-admin/perl-cleaner-1.05 6 kB
[ebuild R ] sys-apps/debianutils-2.29 USE="-static" 0 kB
[ebuild R ] dev-libs/libgcrypt-1.4.1 USE="-bindist -idea -nls" 947 kB
[ebuild R ] dev-util/intltool-0.37.1 136 kB
[ebuild R ] sys-devel/gettext-0.16.1-r1 USE="-doc (-emacs) -nls -nocxx (-openmp)" 8,340 kB
[ebuild R ] sys-apps/baselayout-prefix-1.12.5-r5 216 kB
[ebuild R ] dev-libs/glib-2.16.3-r1 USE="-debug -doc -fam (-hardened) (-selinux) -xattr" 0 kB
[ebuild R ] gnome-base/gnome-mime-data-2.18.0 USE="-debug" 593 kB
[ebuild R ] dev-libs/atk-1.22.0 USE="-debug -doc" 693 kB
[ebuild R ] dev-libs/libIDL-0.8.10 USE="-debug" 364 kB
[ebuild R ] gnome-base/orbit-2.14.13 USE="-debug -doc" 0 kB
[ebuild R ] dev-libs/openssl-0.9.8h-r1 USE="zlib -bindist -gmp -kerberos (-sse2) -test" 0 kB
[ebuild R ] dev-lang/python-2.5.2-r5 USE="ncurses readline ssl threads -berkdb -bootstrap -build -doc -examples -gdbm (-ipv6) -sqlite -tk -ucs2 -wininst" 0 kB
[ebuild R ] app-misc/ca-certificates-20080514-r1 0 kB
[ebuild R ] net-misc/wget-1.11.3 USE="ssl -debug (-ipv6) -nls -socks5 -static" 0 kB
[ebuild R ] dev-libs/libxml2-2.6.32 USE="python readline -bootstrap -build -debug -doc -examples (-ipv6) -test" 4,612 kB
[ebuild R ] sys-apps/file-4.23 USE="python" 551 kB
[ebuild R ] dev-python/pygobject-2.14.2 USE="-debug -doc -examples -libffi" 359 kB
[ebuild R ] dev-python/numeric-24.2-r6 USE="-doc -lapack" 727 kB
[ebuild R ] app-admin/eselect-1.0.11-r01.1 USE="bash-completion -doc -vim-syntax" 150 kB
[ebuild R ] dev-perl/XML-NamespaceSupport-1.09 8 kB
[ebuild R ] dev-libs/libxslt-1.1.24 USE="python -crypt -debug -examples" 0 kB
[ebuild R ] x11-misc/shared-mime-info-0.40 0 kB
[ebuild R ] app-text/docbook-xsl-stylesheets-1.73.2 1,636 kB
[ebuild R ] app-text/scrollkeeper-dtd-1.0 0 kB
[ebuild R ] dev-perl/XML-LibXML-Common-0.13 13 kB
[ebuild R ] dev-perl/XML-SAX-0.16 59 kB
[ebuild R ] app-admin/eselect-vi-1.1.5 2 kB
[ebuild R ] app-text/docbook-xml-dtd-4.4-r1 94 kB
[ebuild R ] app-text/rarian-0.8.0-r1 USE="-debug" 331 kB
[ebuild R ] app-admin/eselect-esd-20060719 2 kB
[ebuild R ] app-admin/eselect-ctags-1.5 8 kB
[ebuild R ] dev-util/ctags-5.7-r1 USE="(-ada)" 281 kB
[ebuild R ] app-text/scrollkeeper-9999-r1 0 kB
[ebuild R ] app-text/gnome-doc-utils-0.12.2-r1 USE="-debug" 0 kB
[ebuild R ] media-sound/esound-0.2.38-r1 USE="(-alsa) -debug -doc (-ipv6) -tcpd" 385 kB
[ebuild U ] dev-perl/XML-LibXML-1.66 [1.65] 271 kB
[ebuild R ] dev-perl/XML-Simple-2.18 70 kB
[ebuild R ] x11-misc/icon-naming-utils-0.8.6 68 kB
[ebuild R ] x11-themes/gnome-icon-theme-2.22.0 USE="-debug" 3,581 kB
[ebuild R ] app-editors/vim-core-7.1.319 USE="bash-completion -acl -livecd -nls (-aqua%)" 0 kB
[ebuild R ] app-editors/vim-7.1.319 USE="bash-completion cscope python -acl (-gpm) -minimal -nls -perl -ruby -vim-pager -vim-with-x" 0 kB
[ebuild R ] app-vim/gentoo-syntax-20070506 USE="-ignore-glep31" 19 kB
[ebuild R ] virtual/xft-7.0 0 kB
[ebuild R ] sys-devel/gcc-config-1.4.0-r04.3 0 kB
[ebuild R ] sys-devel/binutils-config-1.9-r04.12 0 kB
[ebuild R ] sys-devel/binutils-2.18-r2 USE="-multislot -multitarget -nls -test -vanilla" 0 kB
[ebuild R ] x11-proto/xproto-7.0.13 0 kB
[ebuild R ] x11-proto/xextproto-7.0.3 0 kB
[ebuild R ] x11-libs/xtrans-1.2 USE="-debug" 105 kB
[ebuild R ] x11-proto/inputproto-1.4.3-r1 53 kB
[ebuild R ] x11-proto/renderproto-0.9.3 53 kB
[ebuild R ] x11-proto/kbproto-1.0.3 57 kB
[ebuild R ] x11-proto/damageproto-1.1.0 41 kB
[ebuild R ] x11-proto/xf86bigfontproto-1.1.2 37 kB
[ebuild R ] x11-proto/bigreqsproto-1.0.2 36 kB
[ebuild R ] x11-proto/xcmiscproto-1.1.2 36 kB
[ebuild R ] x11-proto/compositeproto-0.4 45 kB
[ebuild R ] x11-proto/randrproto-1.2.1 54 kB
[ebuild R ] x11-proto/fixesproto-4.0 38 kB
[ebuild R ] x11-libs/pixman-0.10.0 USE="-debug" 305 kB
[ebuild R ] sys-devel/gcc-4.2.4 USE="gtk (-altivec) -bootstrap -build -doc -fortran -gcj (-hardened) -ip28 -ip32r10k -libffi -mudflap (-multilib) -multislot (-n32) (-n64) -nls -nocxx -objc -objc++ -objc-gc (-openmp) -test -vanilla" 0 kB
[ebuild R ] x11-libs/libXau-1.0.3 USE="-debug" 225 kB
[ebuild R ] x11-libs/libICE-1.0.4 USE="-debug (-ipv6)" 247 kB
[ebuild R ] x11-libs/libXdmcp-1.0.2 USE="-debug" 216 kB
[ebuild R ] x11-libs/libfontenc-1.0.4 USE="-debug" 219 kB
[ebuild R ] x11-libs/libX11-1.1.4 USE="-debug (-ipv6) -xcb" 1,540 kB
[ebuild R ] x11-libs/libSM-1.0.3 USE="-debug (-ipv6)" 219 kB
[ebuild R ] x11-libs/libXext-1.0.4 USE="-debug" 259 kB
[ebuild R ] x11-libs/libXt-1.0.5 USE="-debug" 489 kB
[ebuild R ] x11-libs/libXrender-0.9.4 USE="-debug" 222 kB
[ebuild R ] media-libs/freetype-2.3.7 USE="X -bindist -debug -doc -utils" 0 kB
[ebuild R ] x11-libs/libXfixes-4.0.3 USE="-debug" 210 kB
[ebuild R ] x11-libs/libXmu-1.0.3 USE="-debug (-ipv6)" 290 kB
[ebuild R ] sys-apps/dbus-1.2.1 USE="X -debug -doc (-selinux)" 1,374 kB
[ebuild R ] x11-libs/startup-notification-0.9 221 kB
[ebuild R ] x11-libs/libXcursor-1.1.9 USE="-debug" 230 kB
[ebuild R ] x11-libs/libXi-1.1.3 USE="-debug" 243 kB
[ebuild R ] x11-libs/libXdamage-1.1.1 USE="-debug" 197 kB
[ebuild R ] x11-libs/libXrandr-1.2.2 USE="-debug" 220 kB
[ebuild R ] x11-libs/libXcomposite-0.4.0 USE="-debug" 213 kB
[ebuild R ] x11-libs/libXpm-3.5.7 USE="-debug" 350 kB
[ebuild U ] x11-apps/mkfontscale-1.0.5 [1.0.4] USE="-debug" 99 kB
[ebuild R ] x11-apps/xhost-1.0.2 USE="-debug (-ipv6)" 96 kB
[ebuild R ] x11-apps/xrdb-1.0.5 USE="-debug" 100 kB
[ebuild R ] x11-apps/mkfontdir-1.0.4 USE="-debug" 69 kB
[ebuild R ] x11-apps/xauth-1.0.3 USE="-debug (-ipv6)" 110 kB
[ebuild R ] dev-libs/dbus-glib-0.74-r1 USE="-debug -doc (-selinux)" 640 kB
[ebuild R ] x11-libs/libXaw-1.0.4 USE="-debug -xprint" 506 kB
[ebuild R ] gnome-base/libbonobo-2.22.0 USE="-debug -doc" 1,412 kB
[ebuild R ] sys-apps/groff-1.19.2-r3 USE="X -cjk" 2,836 kB
[ebuild R ] net-misc/openssh-4.7_p1-r20 USE="X -X509 -chroot -hpn -kerberos -ldap -libedit -pam (-selinux) -skey (-smartcard) -static -tcpd" 968 kB
[ebuild R ] sys-apps/man-1.6f-r2 USE="-nls" 249 kB
[ebuild R ] media-libs/fontconfig-2.6.0-r2 USE="-doc" 0 kB
[ebuild R ] x11-libs/libXft-2.1.12 USE="-debug" 255 kB
[ebuild R ] app-admin/eselect-fontconfig-1.0 0 kB
[ebuild R ] media-fonts/corefonts-1-r4 USE="X" 3,843 kB
[ebuild R ] media-fonts/freefonts-0.10-r3 USE="X" 2,367 kB
[ebuild R ] x11-libs/cairo-1.6.4 USE="X svg (-aqua) -debug (-directfb) -doc -glitz (-opengl) -test -xcb" 4,925 kB
[ebuild R ] x11-libs/pango-1.20.5 USE="X -debug -doc" 0 kB
[ebuild R ] dev-python/pycairo-1.4.12 USE="-examples" 479 kB
[ebuild R ] x11-libs/gtk+-2.12.10-r1 USE="X jpeg (-aqua) -cups -debug -doc (-tiff) -vim-syntax -xinerama" 0 kB
[ebuild R ] gnome-base/gconf-2.22.0 USE="-debug -doc -ldap" 1,384 kB
[ebuild R ] gnome-base/libglade-2.6.2 USE="-debug -doc" 344 kB
[ebuild R ] gnome-base/gail-1.22.1 USE="-debug -doc" 659 kB
[ebuild R ] x11-themes/gtk-engines-ubuntulooks-0.9.12-r2 346 kB
[ebuild R ] x11-themes/gtk-engines-2.14.3 USE="X -accessibility -debug -static" 0 kB
[ebuild R ] x11-themes/gtk-chtheme-0.3.1-r1 0 kB
[ebuild R ] gnome-base/gnome-vfs-2.22.0 USE="ssl -acl -avahi -debug -doc -fam -gnutls (-hal) (-ipv6) -kerberos -samba" 1,895 kB
[ebuild R ] gnome-base/libgnomecanvas-2.20.1.1 USE="X -debug -doc" 570 kB
[ebuild R ] dev-python/pygtk-2.12.1-r2 USE="X -doc -examples" 0 kB
[ebuild R ] gnome-base/gnome-keyring-2.22.3 USE="-debug -doc (-hal) -pam -test" 0 kB
[ebuild R ] gnome-base/libgnome-2.22.0 USE="esd -debug -doc" 1,375 kB
[ebuild R ] x11-libs/vte-0.16.14 USE="python -debug -doc -nowheelscroll (-opengl)" 0 kB
[ebuild R ] gnome-base/libbonoboui-2.22.0 USE="X -debug -doc" 983 kB
[ebuild R ] gnome-base/libgnomeui-2.22.1 USE="jpeg -debug -doc" 1,417 kB
[ebuild R ] x11-terms/gnome-terminal-2.22.3 USE="-debug" 0 kB

Total: 189 packages (7 upgrades, 182 reinstalls), Size of downloads: 110,099 kB

Would you like to merge these packages? [Yes/No]
Post by Fabian Groffen
--
Fabian Groffen
Gentoo on a different level
--
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-16 11:50:50 UTC
Permalink
Post by Markus Duft
I'm doing this with one of the prefixes targeting the setup CDs, so
this should be pretty much the most essential stuff :) CFLAGS now
should contain _ALL_SOURCE.
Ok, it's gone pretty much through (allthough i stopped it prematurely, since I accidently activated the _ALL_SOURCE gcc patch already :). Now I'll build all packages again, this time with the final solution (_ALL_SOURCE built-in to gcc). Also I took a quick look over the tree, and eliminated pretty much all append-flags for _ALL_SOURCE already. This is gonna be a _huge_ checkin if it works :)

The only thing that came to my mind that could be a problem: while bootstrapping as long as there is no gcc of our own, there could be problems building things. Should we just set C[XX]FLAGS during bootstrap? I can do it in eprefix-bootstrap - do you plan to create some manual bootstrap guide for interix too?

Cheers, Markus
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-16 11:59:00 UTC
Permalink
Post by Markus Duft
Post by Markus Duft
I'm doing this with one of the prefixes targeting the setup CDs, so
this should be pretty much the most essential stuff :) CFLAGS now
should contain _ALL_SOURCE.
Ok, it's gone pretty much through (allthough i stopped it prematurely, since I accidently activated the _ALL_SOURCE gcc patch already :). Now I'll build all packages again, this time with the final solution (_ALL_SOURCE built-in to gcc). Also I took a quick look over the tree, and eliminated pretty much all append-flags for _ALL_SOURCE already. This is gonna be a _huge_ checkin if it works :)
good!
Post by Markus Duft
The only thing that came to my mind that could be a problem: while bootstrapping as long as there is no gcc of our own, there could be problems building things. Should we just set C[XX]FLAGS during bootstrap? I can do it in eprefix-bootstrap - do you plan to create some manual bootstrap guide for interix too?
Yes. Like we need to set -L -R stuff, we just have to set the define,
until we emerged gcc.
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-17 06:52:54 UTC
Permalink
Post by Markus Duft
Post by Markus Duft
Post by Markus Duft
I'm doing this with one of the prefixes targeting the setup CDs, so
this should be pretty much the most essential stuff :) CFLAGS now
should contain _ALL_SOURCE.
Ok, it's gone pretty much through (allthough i stopped it
prematurely, since I accidently activated the _ALL_SOURCE gcc patch
already :). Now I'll build all packages again, this time with the final
solution (_ALL_SOURCE built-in to gcc). Also I took a quick look over
the tree, and eliminated pretty much all append-flags for _ALL_SOURCE
already. This is gonna be a _huge_ checkin if it works :)
good!
Yeeeha... it seems to work fine. I'll commit the gcc patch now, and the rest tomorrow? Or should I just commit everything? If things don't build anywhere, one would have to reinstall gcc. Should I rev-bump gcc?

Cheers, Markus
Post by Markus Duft
Post by Markus Duft
The only thing that came to my mind that could be a problem: while
bootstrapping as long as there is no gcc of our own, there could be
problems building things. Should we just set C[XX]FLAGS during
bootstrap? I can do it in eprefix-bootstrap - do you plan to create
some manual bootstrap guide for interix too?
Yes. Like we need to set -L -R stuff, we just have to set the define,
until we emerged gcc.
--
Fabian Groffen
Gentoo on a different level
--
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-17 08:20:41 UTC
Permalink
Post by Markus Duft
Yeeeha... it seems to work fine. I'll commit the gcc patch now, and the
rest tomorrow? Or should I just commit everything? If things don't
build anywhere, one would have to reinstall gcc. Should I rev-bump gcc?
Cheers, Markus
After discussing this with haubi, i bumped the gcc revision, and committed the patch. Now I'm about to commit all the ebuilds I changed.

Cheers, Markus
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-17 09:03:26 UTC
Permalink
Post by Markus Duft
Post by Markus Duft
Yeeeha... it seems to work fine. I'll commit the gcc patch now, and the
rest tomorrow? Or should I just commit everything? If things don't
build anywhere, one would have to reinstall gcc. Should I rev-bump gcc?
Cheers, Markus
After discussing this with haubi, i bumped the gcc revision, and committed the patch. Now I'm about to commit all the ebuilds I changed.
I'm just cleaning up after you. I masked your rev-bump, and unmasked on
Interix only. It's no fun to rebuild gcc for no particular reason on
certain platforms. I still have to recover the old ebuild to finish it.

While removing append-flags from ebuilds, please don't forget about the
flag-o-matic inherit (check it isn't necessary any more).
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-17 15:47:59 UTC
Permalink
Post by Markus Duft
Post by Markus Duft
Post by Markus Duft
Yeeeha... it seems to work fine. I'll commit the gcc patch now, and
the
Post by Markus Duft
Post by Markus Duft
rest tomorrow? Or should I just commit everything? If things don't
build anywhere, one would have to reinstall gcc. Should I rev-bump
gcc?
Post by Markus Duft
Post by Markus Duft
Cheers, Markus
After discussing this with haubi, i bumped the gcc revision, and
committed the patch. Now I'm about to commit all the ebuilds I changed.
I'm just cleaning up after you. I masked your rev-bump, and unmasked on
Interix only. It's no fun to rebuild gcc for no particular reason on
certain platforms. I still have to recover the old ebuild to finish it.
Doesn't work:

(2) mduft gentoo.binary.distribution-interix-6.0 $ emerge -av gcc

These are the packages that would be merged, in order:

Calculating dependencies... done!
[ebuild R ] sys-devel/gcc-4.2.4 USE="gtk (-altivec) -bootstrap -build -doc -fortran -gcj (-hardened) -ip28 -ip32r10k -libffi -mudflap (-multilib) -multislot (-n32) (-n64) -nls -nocxx -objc -objc++ -objc-gc (-openmp) -test -vanilla" 0 kB

Total: 1 package (1 reinstall), Size of downloads: 0 kB

Would you like to merge these packages? [Yes/No] n

Quitting.

(2) mduft gentoo.binary.distribution-interix-6.0 $
(2) mduft gentoo.binary.distribution-interix-6.0 $
(2) mduft gentoo.binary.distribution-interix-6.0 $ cat usr/portage/profiles/default-prefix/windows/interix/package.unmask


# Fabian Groffen <***@gentoo.org> (17 Jul 2008)
# This is a bump for Interix only
=sys-devel/gcc-4.2.4-r00.1
(2) mduft gentoo.binary.distribution-interix-6.0 $ emerge -av '=sys-devel/gcc-4.2.4-r00.1'

These are the packages that would be merged, in order:

Calculating dependencies... done!

!!! All ebuilds that could satisfy "=sys-devel/gcc-4.2.4-r00.1" have been masked.
!!! One of the following masked packages is required to complete your request:
- sys-devel/gcc-4.2.4-r00.1 (masked by: package.mask)
/opt/gentoo.system/usr/portage/profiles/default-prefix/package.mask:
# Fabian Groffen <***@gentoo.org> (17 Jul 2008)
# This is a bump for Interix only


For more information, see the MASKED PACKAGES section in the emerge
man page or refer to the Gentoo Handbook.

(2) mduft gentoo.binary.distribution-interix-6.0 $
Post by Markus Duft
While removing append-flags from ebuilds, please don't forget about the
flag-o-matic inherit (check it isn't necessary any more).
--
Fabian Groffen
Gentoo on a different level
--
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-17 16:07:19 UTC
Permalink
Post by Markus Duft
!!! All ebuilds that could satisfy "=sys-devel/gcc-4.2.4-r00.1" have been masked.
- sys-devel/gcc-4.2.4-r00.1 (masked by: package.mask)
# This is a bump for Interix only
Hmmmm, bah!
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-17 16:16:42 UTC
Permalink
Post by Fabian Groffen
Post by Markus Duft
!!! All ebuilds that could satisfy "=sys-devel/gcc-4.2.4-r00.1" have been masked.
- sys-devel/gcc-4.2.4-r00.1 (masked by: package.mask)
# This is a bump for Interix only
Hmmmm, bah!
can you svn up and try again?
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-18 07:36:53 UTC
Permalink
Post by Markus Duft
Post by Fabian Groffen
Post by Markus Duft
!!! All ebuilds that could satisfy "=sys-devel/gcc-4.2.4-r00.1"
have been masked.
Post by Fabian Groffen
Post by Markus Duft
!!! One of the following masked packages is required to complete
- sys-devel/gcc-4.2.4-r00.1 (masked by: package.mask)
/opt/gentoo.system/usr/portage/profiles/default-
# This is a bump for Interix only
Hmmmm, bah!
can you svn up and try again?
Works as expected now :) thanks a lot
Post by Markus Duft
--
Fabian Groffen
Gentoo on a different level
--
--
gentoo-***@lists.gentoo.org mailing list
Fabian Groffen
2008-07-18 08:20:44 UTC
Permalink
Post by Markus Duft
Post by Fabian Groffen
Post by Fabian Groffen
Hmmmm, bah!
can you svn up and try again?
Works as expected now :) thanks a lot
Sorry for the mess, though :(
--
Fabian Groffen
Gentoo on a different level
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-02 09:46:59 UTC
Permalink
Post by Fabian Groffen
Post by Markus Duft
Post by Fabian Groffen
What do you mean? I would just enable it, rebuild system (wait for
+- 2
Post by Markus Duft
Post by Fabian Groffen
days) if that's ok, commit it, and then remove all append-flags
-D_ALL_SOURCE stuff. People having a prefix shouldn't notice that,
should they? New packages they install just should build, and I
hope no
Post by Markus Duft
Post by Fabian Groffen
upstream is so stupid to put #ifdef _ALL_SOURCE in their headers
like
Post by Markus Duft
Post by Fabian Groffen
Interix does. Quick grepping doesn't show anything like a changing
definition or something.
I meant that changing the define makes various configure scripts find
different things, and thus may enable (or disable) different code
paths, which _could_ in some stupid cases confuse already merged
packages, or the new ones, since it expects something from the merged
packages they don't provide, since at build time those things weren't
there.
I thought about this again. Maybe we should try, and simply continue with live as if nothing happened after switching. I don't think that there would be too much packages having problems. We should simply keep this change in mind, in case something bad happens. Also I will try and bootstrap an extra prefix from zero to see whether this works with _ALL_SOURCE always defined. All other prefixes, I'll simply leave alone, and update as appropriate, just like always.

Cheers, Markus
Post by Fabian Groffen
Ok, good point. Then I guess it's either you or me doing it, and just
aligning it, that when it's pushed into the tree, we
a) make a new bootstrap snapshot
b) upload new installers using the newly compiled packages
I can do the compilation of system, but maybe that doesn't make much
sense since you'll have to build the installers/binpkgs anyway?
--
Fabian Groffen
Gentoo on a different level
--
--
gentoo-***@lists.gentoo.org mailing list
Markus Duft
2008-07-04 08:13:19 UTC
Permalink
<snip>
Post by Fabian Groffen
I can do the compilation of system, but maybe that doesn't make much
sense since you'll have to build the installers/binpkgs anyway?
I'll be unavailable (vacation) the next week. If you want, you can activate this, and give it a try. After I come back, I'll start building binary packages for 5.2 and 3.5. BTW, I've got approximately 270 binpkgs for 6.0 by now, still 312 to go... :)

Cheers, Markus
Post by Fabian Groffen
--
Fabian Groffen
Gentoo on a different level
--
--
--
gentoo-***@lists.gentoo.org mailing list
Loading...