[boost] [githelp] modular boost instructions?

36 views
Skip to first unread message

Richard

unread,
May 16, 2013, 10:17:21 PM5/16/13
to bo...@lists.boost.org
[Please do not mail me a copy of your followup]

OK, we've already established that the 'Experimenting with Modular Boost'
instructions are woefully stale.

Can someone please briefly describe what submodules I need to manually
add just to get something bootstrapped?

In the svn boost there is <root>/build and <root>/tools/build; neither
of these is present in boostorg/boost and while there is a build repo,
it seems to define a structure that creates <root>/build/build, so I'm
confused.
--
"The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline>
The Computer Graphics Museum <http://computergraphicsmuseum.org>
The Terminals Wiki <http://terminals.classiccmp.org>
Legalize Adulthood! (my blog) <http://legalizeadulthood.wordpress.com>


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Dave Abrahams

unread,
May 17, 2013, 5:17:09 AM5/17/13
to bo...@lists.boost.org

on Thu May 16 2013, legalize+jeeves-AT-mail.xmission.com (Richard) wrote:

> [Please do not mail me a copy of your followup]
>
> OK, we've already established that the 'Experimenting with Modular Boost'
> instructions are woefully stale.
>
> Can someone please briefly describe what submodules I need to manually
> add just to get something bootstrapped?

None; we (finally!) have a modularized history with submodules!
Please

git clone --recursive http://github.com/boostorg/boost



--
Dave Abrahams

Nasos Iliopoulos

unread,
May 17, 2013, 8:47:20 AM5/17/13
to bo...@lists.boost.org
Great, it seems that this is gonna be ready soon!

Some problems I have:
1.

$ git clone --recursive http://github.com/boostorg/boost
Cloning into 'boost'...
remote: Counting objects: 138912, done.
remote: Compressing objects: 100% (53319/53319), done.
remote: Total 138912 (delta 84911), reused 138380 (delta 84398)
Receiving objects: 100% (138912/138912), 81.34 MiB | 7.09 MiB/s, done.
Resolving deltas: 100% (84911/84911), done.
No submodule mapping found in .gitmodules for path 'libs/chrono'

2. The latter seems to break submodule init.

$ git submodule init
No submodule mapping found in .gitmodules for path 'libs/chrono'

3. .gitmodules comes with the math entry be default?:
[submodule "libs/math"]
path = libs/math
url = http://github.com/boostorg/math


3. At least ublas subdir is not created in numeric.

Best,
Nasos


On 05/17/2013 05:17 AM, Dave Abrahams wrote:
> on Thu May 16 2013, legalize+jeeves-AT-mail.xmission.com (Richard) wrote:
>
>> [Please do not mail me a copy of your followup]
>>
>> OK, we've already established that the 'Experimenting with Modular Boost'
>> instructions are woefully stale.
>>
>> Can someone please briefly describe what submodules I need to manually
>> add just to get something bootstrapped?
> None; we (finally!) have a modularized history with submodules!
> Please
>
> git clone --recursive http://github.com/boostorg/boost
>
>
>


Bjørn Roald

unread,
May 17, 2013, 9:00:17 AM5/17/13
to bo...@lists.boost.org
On Fri, 2013-05-17 at 03:17 -0600, Dave Abrahams wrote:
> on Thu May 16 2013, legalize+jeeves-AT-mail.xmission.com (Richard) wrote:
> > Can someone please briefly describe what submodules I need to manually
> > add just to get something bootstrapped?
>
> None; we (finally!) have a modularized history with submodules!
> Please
>
> git clone --recursive http://github.com/boostorg/boost

I don't know exactly what to expect, but I get:


> git clone --recursive http://github.com/boostorg/boost

Cloning into 'boost'...
remote: Counting objects: 138912, done.
remote: Compressing objects: 100% (53319/53319), done.
remote: Total 138912 (delta 84911), reused 138380 (delta 84398)

Receiving objects: 100% (138912/138912), 81.34 MiB | 3.98 MiB/s, done.


Resolving deltas: 100% (84911/84911), done.
No submodule mapping found in .gitmodules for path 'libs/chrono'


The .gitmodules file only have an entry for math:


[submodule "libs/math"]
path = libs/math
url = http://github.com/boostorg/math

and all folders in libs/* are empty.

--
Bjørn

Jonathan Wakely

unread,
May 17, 2013, 9:41:02 AM5/17/13
to bo...@lists.boost.org
On 17 May 2013 14:00, Bjørn Roald wrote:
> On Fri, 2013-05-17 at 03:17 -0600, Dave Abrahams wrote:
>> on Thu May 16 2013, legalize+jeeves-AT-mail.xmission.com (Richard) wrote:
>> > Can someone please briefly describe what submodules I need to manually
>> > add just to get something bootstrapped?
>>
>> None; we (finally!) have a modularized history with submodules!
>> Please
>>
>> git clone --recursive http://github.com/boostorg/boost
>
> I don't know exactly what to expect, but I get:
>
>
>> git clone --recursive http://github.com/boostorg/boost
> Cloning into 'boost'...
> remote: Counting objects: 138912, done.
> remote: Compressing objects: 100% (53319/53319), done.
> remote: Total 138912 (delta 84911), reused 138380 (delta 84398)
> Receiving objects: 100% (138912/138912), 81.34 MiB | 3.98 MiB/s, done.
> Resolving deltas: 100% (84911/84911), done.
> No submodule mapping found in .gitmodules for path 'libs/chrono'

It appears to have worked fine for me, with Git 1.7.12.4

Nasos Iliopoulos

unread,
May 17, 2013, 10:01:27 AM5/17/13
to bo...@lists.boost.org
Unfortunately 1.8.3.rc2 gave me the same error.

Furthermore it seems that .gitmodules in
5b5883234251dbe6616c696c4b3f7fdda7fde7d7
was in a better state, but it still didn't include tools/build and
libs/numeric/ublas.

Even if you manually add those submodules, bootstrap will fail because
the build repository is under a "build" folder and is cloned into
tools/build/build. I am not sure if this is this the intended behavior.

-Nasos

Richard

unread,
May 17, 2013, 10:50:10 AM5/17/13
to bo...@lists.boost.org
[Please do not mail me a copy of your followup]

bo...@lists.boost.org spake the secret code
<BLU0-SMTP20584AB11...@phx.gbl> thusly:

>$ git clone --recursive http://github.com/boostorg/boost
>Cloning into 'boost'...
>remote: Counting objects: 138912, done.
>remote: Compressing objects: 100% (53319/53319), done.
>remote: Total 138912 (delta 84911), reused 138380 (delta 84398)
>Receiving objects: 100% (138912/138912), 81.34 MiB | 7.09 MiB/s, done.
>Resolving deltas: 100% (84911/84911), done.
>No submodule mapping found in .gitmodules for path 'libs/chrono'

I get the same thing.

git version 1.7.9.5
--
"The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline>
The Computer Graphics Museum <http://computergraphicsmuseum.org>
The Terminals Wiki <http://terminals.classiccmp.org>
Legalize Adulthood! (my blog) <http://legalizeadulthood.wordpress.com>


Bjørn Roald

unread,
May 17, 2013, 11:23:31 AM5/17/13
to bo...@lists.boost.org
On Fri, 2013-05-17 at 10:01 -0400, Nasos Iliopoulos wrote:
> Unfortunately 1.8.3.rc2 gave me the same error.
>
> Furthermore it seems that .gitmodules in
> 5b5883234251dbe6616c696c4b3f7fdda7fde7d7
> was in a better state,

yes that is better, it seems git commit today

Author: Paul A. Bristow <pbri...@hetp.u-net.com> 2013-05-17 12:39:59
Committer: Paul A. Bristow <pbri...@hetp.u-net.com> 2013-05-17
12:39:59

svn path=; revision=84317


added reference to lib/math in .gitmodules, but removed the other
entries in the file. I don't know if this is from svn 84317 or caused
by something in the conversion scripts.

--
Bjørn

Richard

unread,
May 17, 2013, 12:32:25 PM5/17/13
to bo...@lists.boost.org
[Please do not mail me a copy of your followup]

bo...@lists.boost.org spake the secret code
<1368804211.11156.27.camel@frodo2> thusly:

>On Fri, 2013-05-17 at 10:01 -0400, Nasos Iliopoulos wrote:
>> Unfortunately 1.8.3.rc2 gave me the same error.
>>
>> Furthermore it seems that .gitmodules in
>> 5b5883234251dbe6616c696c4b3f7fdda7fde7d7
>> was in a better state,
>
>yes that is better, it seems git commit today

I saw that too, so I updated to that version and updated all
submodules. This gets me stuff in libs, but I still don't get
tools/build or build. Therefore, I conclude that I still don't have
the build environment and the bootstrap.sh won't work.
--
"The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline>
The Computer Graphics Museum <http://computergraphicsmuseum.org>
The Terminals Wiki <http://terminals.classiccmp.org>
Legalize Adulthood! (my blog) <http://legalizeadulthood.wordpress.com>


Paul A. Bristow

unread,
May 17, 2013, 1:17:15 PM5/17/13
to bo...@lists.boost.org

> -----Original Message-----
> From: Boost [mailto:boost-...@lists.boost.org] On Behalf Of Bjørn Roald
> Sent: Friday, May 17, 2013 4:24 PM
> To: bo...@lists.boost.org
> Subject: Re: [boost] [githelp] modular boost instructions?
>
> On Fri, 2013-05-17 at 10:01 -0400, Nasos Iliopoulos wrote:
> > Unfortunately 1.8.3.rc2 gave me the same error.
> >
> > Furthermore it seems that .gitmodules in
> > 5b5883234251dbe6616c696c4b3f7fdda7fde7d7
> > was in a better state,
>
> yes that is better, it seems git commit today
>
> Author: Paul A. Bristow <pbri...@hetp.u-net.com> 2013-05-17 12:39:59
> Committer: Paul A. Bristow <pbri...@hetp.u-net.com> 2013-05-17
> 12:39:59
>
> svn path=; revision=84317

Guilty!

Revision: 84317
Author: pbristow
Date: 17 May 2013 11:39:59
Message:
New def links added
----
Modified : /trunk/libs/math/doc/math.qbk

And there are a few more later by me and John Maddock updating Boost.Math docs.

HTH

Paul


---
Paul A. Bristow,
Prizet Farmhouse, Kendal LA8 8AB UK
+44 1539 561830 07714330204
pbri...@hetp.u-net.com

Daniel Pfeifer

unread,
May 17, 2013, 4:59:59 PM5/17/13
to Boost Developers List
2013/5/17 Paul A. Bristow <pbri...@hetp.u-net.com>

>
>
> > -----Original Message-----
> > From: Boost [mailto:boost-...@lists.boost.org] On Behalf Of Bjørn
> Roald
> > Sent: Friday, May 17, 2013 4:24 PM
> > To: bo...@lists.boost.org
> > Subject: Re: [boost] [githelp] modular boost instructions?
> >
> > On Fri, 2013-05-17 at 10:01 -0400, Nasos Iliopoulos wrote:
> > > Unfortunately 1.8.3.rc2 gave me the same error.
> > >
> > > Furthermore it seems that .gitmodules in
> > > 5b5883234251dbe6616c696c4b3f7fdda7fde7d7
> > > was in a better state,
> >
> > yes that is better, it seems git commit today
> >
> > Author: Paul A. Bristow <pbri...@hetp.u-net.com> 2013-05-17 12:39:59
> > Committer: Paul A. Bristow <pbri...@hetp.u-net.com> 2013-05-17
> > 12:39:59
> >
> > svn path=; revision=84317
>
> Guilty!
>

This is a bug. Noted: https://github.com/ryppl/Boost2Git/issues/21

Richard

unread,
May 18, 2013, 9:09:59 PM5/18/13
to bo...@lists.boost.org
[Please do not mail me a copy of your followup]

bo...@lists.boost.org spake the secret code
<m2mwruu...@boostpro.com> thusly:

>on Thu May 16 2013, legalize+jeeves-AT-mail.xmission.com (Richard) wrote:
>
>> Can someone please briefly describe what submodules I need to manually
>> add just to get something bootstrapped?
>
>None; we (finally!) have a modularized history with submodules!
>Please
>
> git clone --recursive http://github.com/boostorg/boost

...still wondering where build is supposed to come from. Anyone?
--
"The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline>
The Computer Graphics Museum <http://computergraphicsmuseum.org>
The Terminals Wiki <http://terminals.classiccmp.org>
Legalize Adulthood! (my blog) <http://legalizeadulthood.wordpress.com>


Nasos Iliopoulos

unread,
May 19, 2013, 8:26:57 AM5/19/13
to bo...@lists.boost.org
It seems that .gitmodules is being pushed regularly by individual's
rather than retaining a common one.
A few days ago .gitmodules was overwritten to include only boost.math,
now it references only boost.log.

Look at those commits for example:
(older to newer)

https://github.com/boostorg/boost/commit/ec263acde5756da86a6036b36eabc8dbe5720d22
https://github.com/boostorg/boost/commit/1862b15a2b24f81a497c7d5fd198529159812c34#.gitmodules
https://github.com/boostorg/boost/commit/57867018811951232b896c2808f3b1710c101fce#.gitmodules

If this doesn't stop and build doesn't change it directory structure it
won't work but only by manually moving things around.


-Nasos


On 05/18/2013 09:09 PM, Richard wrote:
> [Please do not mail me a copy of your followup]
>
> bo...@lists.boost.org spake the secret code
> <m2mwruu...@boostpro.com> thusly:
>
>> on Thu May 16 2013, legalize+jeeves-AT-mail.xmission.com (Richard) wrote:
>>
>>> Can someone please briefly describe what submodules I need to manually
>>> add just to get something bootstrapped?
>> None; we (finally!) have a modularized history with submodules!
>> Please
>>
>> git clone --recursive http://github.com/boostorg/boost
> ...still wondering where build is supposed to come from. Anyone?


Bjørn Roald

unread,
May 19, 2013, 1:02:14 PM5/19/13
to bo...@lists.boost.org
On Sun, 2013-05-19 at 08:26 -0400, Nasos Iliopoulos wrote:
> It seems that .gitmodules is being pushed regularly by individual's
> rather than retaining a common one.

The git commits are generated by the Boost2git conversion script and
they are actually misleading on this point as the .gitmodules filr is
generated by the scrippts and not in the personal commits in svn.

I think all this is caused by a bug in the Boost2git conversion script
for incremental conversion.

https://github.com/ryppl/Boost2Git/issues/21

since the git repositories are slaves of svn until the switch, I think
this must be fixed as suggested in the bug report and then do a full
conversion to fix the affected part of git history.

> If this doesn't stop and build doesn't change it directory structure it
> won't work but only by manually moving things around.

Agree - why can't the needed parts of tools/build be as in converted git
repository until the switch is complete. To move the files afterwards
in step lock with the required bootstrap.(sh|bat) and boost-build.jam is
trivial. Alternatively more tricky scripting could be used to detect
and support both environments for bootstrap and ./b2.

--
Bjørn

Dave Abrahams

unread,
May 20, 2013, 2:22:45 AM5/20/13
to bo...@lists.boost.org

on Fri May 17 2013, Nasos Iliopoulos <nasos_i-AT-hotmail.com> wrote:

> Great, it seems that this is gonna be ready soon!
>
> Some problems I have:
> 1.
>
> $ git clone --recursive http://github.com/boostorg/boost
> Cloning into 'boost'...
> remote: Counting objects: 138912, done.
> remote: Compressing objects: 100% (53319/53319), done.
> remote: Total 138912 (delta 84911), reused 138380 (delta 84398)
> Receiving objects: 100% (138912/138912), 81.34 MiB | 7.09 MiB/s, done.
> Resolving deltas: 100% (84911/84911), done.
> No submodule mapping found in .gitmodules for path 'libs/chrono'
>
> 2. The latter seems to break submodule init.
>
> $ git submodule init
> No submodule mapping found in .gitmodules for path 'libs/chrono'
>
> 3. .gitmodules comes with the math entry be default?:
> [submodule "libs/math"]
> path = libs/math
> url = http://github.com/boostorg/math
>
> 3. At least ublas subdir is not created in numeric.

This problem should be fixed as of
https://github.com/ryppl/Boost2Git/commit/e13e87a

--
Dave Abrahams

Nasos Iliopoulos

unread,
May 20, 2013, 8:36:35 AM5/20/13
to bo...@lists.boost.org
It seems much better now. I have found though that ublas is still not in
the submodules with the new fix.

Best regards,
-Nasos

On 05/20/2013 02:22 AM, Dave Abrahams wrote:
> on Fri May 17 2013, Nasos Iliopoulos <nasos_i-AT-hotmail.com> wrote:
>
>> Great, it seems that this is gonna be ready soon!
>>
>> Some problems I have:
>> 1.
>>
>> $ git clone --recursive http://github.com/boostorg/boost
>> Cloning into 'boost'...
>> remote: Counting objects: 138912, done.
>> remote: Compressing objects: 100% (53319/53319), done.
>> remote: Total 138912 (delta 84911), reused 138380 (delta 84398)
>> Receiving objects: 100% (138912/138912), 81.34 MiB | 7.09 MiB/s, done.
>> Resolving deltas: 100% (84911/84911), done.
>> No submodule mapping found in .gitmodules for path 'libs/chrono'
>>
>> 2. The latter seems to break submodule init.
>>
>> $ git submodule init
>> No submodule mapping found in .gitmodules for path 'libs/chrono'
>>
>> 3. .gitmodules comes with the math entry be default?:
>> [submodule "libs/math"]
>> path = libs/math
>> url = http://github.com/boostorg/math
>>
>> 3. At least ublas subdir is not created in numeric.
> This problem should be fixed as of
> https://github.com/ryppl/Boost2Git/commit/e13e87a
>


Richard

unread,
May 20, 2013, 12:10:43 PM5/20/13
to bo...@lists.boost.org
[Please do not mail me a copy of your followup]

bo...@lists.boost.org spake the secret code
<1368982934.11156.71.camel@frodo2> thusly:

>On Sun, 2013-05-19 at 08:26 -0400, Nasos Iliopoulos wrote:
>> It seems that .gitmodules is being pushed regularly by individual's
>> rather than retaining a common one.
>
>The git commits are generated by the Boost2git conversion script and
>they are actually misleading on this point as the .gitmodules filr is
>generated by the scrippts and not in the personal commits in svn.

So you're saying if I look at the history of .gitmodules, I will find
an older version that has the submodules for tools/build and build?
--
"The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline>
The Computer Graphics Museum <http://computergraphicsmuseum.org>
The Terminals Wiki <http://terminals.classiccmp.org>
Legalize Adulthood! (my blog) <http://legalizeadulthood.wordpress.com>


Bjørn Roald

unread,
May 20, 2013, 5:26:24 PM5/20/13
to bo...@lists.boost.org
On Mon, 2013-05-20 at 16:10 +0000, Richard wrote:

> So you're saying if I look at the history of .gitmodules, I will find
> an older version that has the submodules for tools/build and build?

No, I am not saying that. I think conversion of tools/build is not
activated or it is broken.

I was talking of all the other submodules that was missing
from .gitmodules due to a bug in the conversion scripts that seems to
have been fixed.

--
Bjørn

Richard

unread,
May 20, 2013, 6:28:29 PM5/20/13
to bo...@lists.boost.org
[Please do not mail me a copy of your followup]

bo...@lists.boost.org spake the secret code
<1369085184.11156.127.camel@frodo2> thusly:

>On Mon, 2013-05-20 at 16:10 +0000, Richard wrote:
>
>> So you're saying if I look at the history of .gitmodules, I will find
>> an older version that has the submodules for tools/build and build?
>
>No, I am not saying that. I think conversion of tools/build is not
>activated or it is broken.

OK, so still looking for an answer to this question:

What git repo do I clone in order to get the build tools?

>I was talking of all the other submodules that was missing
>from .gitmodules due to a bug in the conversion scripts that seems to
>have been fixed.

I noticed that and updated to an earlier verison; I was able to get
all the stuff under libs, but still lacked the build system.
--
"The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline>
The Computer Graphics Museum <http://computergraphicsmuseum.org>
The Terminals Wiki <http://terminals.classiccmp.org>
Legalize Adulthood! (my blog) <http://legalizeadulthood.wordpress.com>


Bjørn Roald

unread,
May 21, 2013, 1:46:04 AM5/21/13
to bo...@lists.boost.org
On Mon, 2013-05-20 at 22:28 +0000, Richard wrote:

> OK, so still looking for an answer to this question:
>
> What git repo do I clone in order to get the build tools?

I would wait until it is fixed proper or simply copy tools/build files
from latest boost release.

If you have to try out this with git, there is a repository you can
clone. But be warned, some of this is not trivial unless you know git
well. It is also a bit of a mess as the structure in tools/build is
changed in the conversion and the top level scripts are not updated
accordingly. It is still being discussed how to deal with this, and the
structure may change again or change back.

# clone boost


git clone --recursive http://github.com/boostorg/boost

cd boost

Now all should be in place, but as you noted tools/build is missing.
To add tools/build submodule manually in boost repository, you may
create a branch to work on add it.

git checkout -b feature/fix-broken-build
git submodule add http://github.com/boostorg/build tools/build

But as you will see, the cloned tools/build submodule suffers from the
restructure of tools/build into tools/build/build :-(

You could checkout a branch in the submodule and edit the submodule for
your own needs - whatever they are, e.g. setting structure back to
tools/build/v2, not tools/build/build/v2:

cd tools/build
git checkout -b feature/fix-broken-build
git mv build/* .
git commit -m "only tools/build, not tools/build/build"

Then in the boost meta repository you may add the changed build
submodule to index and commit.

cd ../..
git add tools/build
git commit -m "only tools/build, not tools/build/build"

Then later, when you get updates upstream you could rebase your changes
on top of upstream commits from Boost2Git conversion with something
like. Be warned, I suspect you may hit non-trivial possibly confusing
rebase if you try it on the boost meta repository as the conversion
scripts are still adjusted, thus history change all the time - that will
stabilize when the switch is done. So the rebase is simpler if you do
it only in your tools/build submodule to get your changes applied on top
of the latest upstream.

git fetch
git rebase origin/master

hope that helps,
--
Bjørn

Richard

unread,
May 21, 2013, 11:16:00 PM5/21/13
to bo...@lists.boost.org
[Please do not mail me a copy of your followup]

bo...@lists.boost.org spake the secret code
<1369115164.11156.166.camel@frodo2> thusly:

>hope that helps,

Yes, that helps a lot. Mainly, you confirmed everything that I suspected.

I'm happy to wait for the dust to settle a little bit more.

I thought things were farther along because there have been some
threads about a review for modular boost. It ain't ready for
primetime yet.

I wanted to try it out for changes to boost.test documentation, but
I'll wait longer for htat. I've got my own mercurial workflow that's
capturing my changes in revisions right now.
--
"The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline>
The Computer Graphics Museum <http://computergraphicsmuseum.org>
The Terminals Wiki <http://terminals.classiccmp.org>
Legalize Adulthood! (my blog) <http://legalizeadulthood.wordpress.com>


Dave Abrahams

unread,
May 22, 2013, 1:57:28 AM5/22/13
to bo...@lists.boost.org

on Tue May 21 2013, legalize+jeeves-AT-mail.xmission.com (Richard) wrote:

> [Please do not mail me a copy of your followup]
>
> bo...@lists.boost.org spake the secret code
> <1369115164.11156.166.camel@frodo2> thusly:
>
>>hope that helps,
>
> Yes, that helps a lot. Mainly, you confirmed everything that I suspected.
>
> I'm happy to wait for the dust to settle a little bit more.
>
> I thought things were farther along because there have been some
> threads about a review for modular boost. It ain't ready for
> primetime yet.

Actually, what we actually wanted reviewed (distribution of SVN changes
to Git repositories, branches, and tags) is perfectly ready. Even
though it's a completely separate issue, I know that most people just
care about seeing submodules, though, so I'm working on a fix for the
problem.

--
Dave Abrahams

Dave Abrahams

unread,
May 22, 2013, 3:08:54 AM5/22/13
to bo...@lists.boost.org

on Tue May 21 2013, Dave Abrahams <dave-AT-boostpro.com> wrote:

> on Tue May 21 2013, legalize+jeeves-AT-mail.xmission.com (Richard) wrote:
>
>> [Please do not mail me a copy of your followup]
>>
>> bo...@lists.boost.org spake the secret code
>> <1369115164.11156.166.camel@frodo2> thusly:
>>
>>>hope that helps,
>>
>> Yes, that helps a lot. Mainly, you confirmed everything that I suspected.
>>
>> I'm happy to wait for the dust to settle a little bit more.
>>
>> I thought things were farther along because there have been some
>> threads about a review for modular boost. It ain't ready for
>> primetime yet.
>
> Actually, what we actually wanted reviewed (distribution of SVN changes
> to Git repositories, branches, and tags) is perfectly ready. Even
> though it's a completely separate issue, I know that most people just
> care about seeing submodules, though, so I'm working on a fix for the
> problem.

I think this is probably fixed now. We'll know for sure after
http://jenkins.boost.org/job/Boost2Git/1538/ completes.

Jürgen Hunold

unread,
May 22, 2013, 3:48:12 AM5/22/13
to bo...@lists.boost.org
Hi Dave,

On Wednesday, 22. May 2013 00:08:54 Dave Abrahams wrote:
> I think this is probably fixed now. We'll know for sure after
> http://jenkins.boost.org/job/Boost2Git/1538/ completes.

Aborted with.

Revision 44341
++ WARNING: refs/tags/tools/build/BOOST_BUILD_V1 is copying from branch
refs/tags/BOOST_BUILD_V1 but the latter doesn't exist. Continuing, assuming
the files exist in repository build

Revision 44923
++ ERROR: sandbox: fast-import process error

make[2]: *** [CMakeFiles/conversion] Error 255
make[1]: *** [CMakeFiles/conversion.dir/all] Error 2
make: *** [all] Error 2


Dave, can you please disable all history rewriting attemps for Boost.Build and
start again? It is much more important to get a working repository now as a
lot of people want to start contributing.

I'll address Volodya's concerns in the other thread.

Yours,

Jürgen
--
* Dipl.-Math. Jürgen Hunold !
* voice: ++49 4257 300 ! Fährstraße 1
* fax : ++49 4257 300 ! 31609 Balge/Sebbenhausen
* jhu...@gmx.eu ! Germany

Dave Abrahams

unread,
May 22, 2013, 1:55:18 PM5/22/13
to bo...@lists.boost.org
Nasos Iliopoulos <nasos_i <at> hotmail.com> writes:

>
> It seems much better now. I have found though that ublas is still not in
> the submodules with the new fix.

Try now.

There were several issues getting in the way. History should not only
contain all the submodules now, but be more complete in other ways as
well. Also I had to hack around GitHub's refusal to accept a push of the
boost super-module via SSH. Watch
http://jenkins.boost.org/job/Boost2Git/1555 for completion.

-Dave

Dave Abrahams

unread,
May 22, 2013, 1:57:32 PM5/22/13
to bo...@lists.boost.org
Jürgen Hunold <jhunold <at> gmx.eu> writes:

>
> Hi Dave,
>
> On Wednesday, 22. May 2013 00:08:54 Dave Abrahams wrote:
> > I think this is probably fixed now. We'll know for sure after
> > http://jenkins.boost.org/job/Boost2Git/1538/ completes.

Correction, http://jenkins.boost.org/job/Boost2Git/1555

> Dave, can you please disable all history rewriting attemps for Boost.Build and
> start again?

What do you consider a "history rewriting attempt" and how would I disable it?

> It is much more important to get a working repository now as a
> lot of people want to start contributing.

I'm sorry, I don't know what you're referring to here

Nasos Iliopoulos

unread,
May 22, 2013, 3:08:48 PM5/22/13
to bo...@lists.boost.org
It seems that it is fixed now.

Thank you for all the hard work!

-Nasos

Jürgen Hunold

unread,
May 22, 2013, 4:31:49 PM5/22/13
to bo...@lists.boost.org
HI Dave,

On Wednesday, 22. May 2013 17:57:32 Dave Abrahams wrote:
> > Dave, can you please disable all history rewriting attemps for Boost.Build
> > and start again?
>
> What do you consider a "history rewriting attempt" and how would I disable
> it?

Sorry, that should have read "path rewriting". That means the move of
"tools/build/v2/" to "tools/build" which now results in
"tools/build/build/v2".

> > It is much more important to get a working repository now as a
> > lot of people want to start contributing.
>
> I'm sorry, I don't know what you're referring to here

I thought that the conversion was hanging on Boost.Build

Sorry for the noise.

Yours,

Jürgen
--
* Dipl.-Math. Jürgen Hunold !
* voice: ++49 4257 300 ! Fährstraße 1
* fax : ++49 4257 300 ! 31609 Balge/Sebbenhausen
* jhu...@gmx.eu ! Germany

Dave Abrahams

unread,
May 23, 2013, 1:12:12 AM5/23/13
to bo...@lists.boost.org

on Wed May 22 2013, Nasos Iliopoulos <nasos_i-AT-hotmail.com> wrote:

> It seems that it is fixed now.
>
> Thank you for all the hard work!

You're welcome. It truly is a massive undertaking.

--
Dave Abrahams

Dave Abrahams

unread,
May 23, 2013, 1:11:40 AM5/23/13
to bo...@lists.boost.org

on Wed May 22 2013, Jürgen Hunold <jhunold-AT-gmx.eu> wrote:

> HI Dave,
>
> On Wednesday, 22. May 2013 17:57:32 Dave Abrahams wrote:
>> > Dave, can you please disable all history rewriting attemps for Boost.Build
>> > and start again?
>>
>> What do you consider a "history rewriting attempt" and how would I disable
>> it?
>
> Sorry, that should have read "path rewriting". That means the move of
> "tools/build/v2/" to "tools/build" which now results in
> "tools/build/build/v2".

It's not actually doing less path rewriting for Boost.Build than for
almost any other repository. Currently, we simply take the
subdirectories of tools/ that are part of Boost.Build (i.e. tools/jam
and tools/build) and put them at the top level of the Boost.Build Git
repo. It's like taking the tools directory but throwing out everything
that's not part of Boost.Build. That seemed to me to be the most
faithful representation of what was happening in SVN.

>> > It is much more important to get a working repository now as a
>> > lot of people want to start contributing.
>>
>> I'm sorry, I don't know what you're referring to here
>
> I thought that the conversion was hanging on Boost.Build

It sort of was, for a while... but not for the reasons you think.

--
Dave Abrahams

Dave Abrahams

unread,
May 23, 2013, 2:13:16 AM5/23/13
to bo...@lists.boost.org

on Wed May 22 2013, Dave Abrahams <dave-AT-boostpro.com> wrote:

> It's not actually doing less path rewriting for Boost.Build than for
> almost any other repository.

Err... s/not //

That is, it's doing less path rewriting for Boost.Build.

Beman Dawes

unread,
May 23, 2013, 11:06:13 AM5/23/13
to Boost Developers List
On Fri, May 17, 2013 at 5:17 AM, Dave Abrahams <da...@boostpro.com> wrote:
>
>
> ... we (finally!) have a modularized history with submodules!
> Please
>
> git clone --recursive http://github.com/boostorg/boost

I've updated the https://svn.boost.org/trac/boost/wiki/TryModBoost
instructions accordingly.

--Beman

Richard

unread,
May 23, 2013, 11:25:12 PM5/23/13
to bo...@lists.boost.org
[Please do not mail me a copy of your followup]

bo...@lists.boost.org spake the secret code
<m28v37r...@boostpro.com> thusly:

>Actually, what we actually wanted reviewed (distribution of SVN changes
>to Git repositories, branches, and tags) is perfectly ready.

Great!

I'm late to the party, so I'm doing my best to follow what's going on,
but the really stale pages on the wiki have kept me out of the loop
unless I want to go do some mailing list archaeology. I'm content to
wait for things to evolve further.
--
"The Direct3D Graphics Pipeline" free book <http://tinyurl.com/d3d-pipeline>
The Computer Graphics Museum <http://computergraphicsmuseum.org>
The Terminals Wiki <http://terminals.classiccmp.org>
Legalize Adulthood! (my blog) <http://legalizeadulthood.wordpress.com>


Reply all
Reply to author
Forward
0 new messages