bug#39575: guix time-machine fails when a tarball was modified in-place

bug#39575: guix time-machine fails when a tarball was modified in-place

[Jan Nieuwenhuizen]

Hi,

Trying to travel back to Sun Apr 7 22:07:14 2019 +0200 (commit
56e95d54d209c2428f970d65d9b27ae4168449ad) to re-create mcrl2-minimal by
doing

--8<---------------cut here---------------start------------->8---
guix time-machine --commit=56e95d54d209c2428f970d65d9b27ae4168449ad -- environment --ad-hoc mcrl2-minimal
--8<---------------cut here---------------end--------------->8---

fails with

--8<---------------cut here---------------start------------->8---
building /gnu/store/cjim33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv...
downloading from https://www.freedesktop.org/software/harfbuzz/release/harfbuzz-2.4.0.tar.bz2...
|offloading build of /gnu/store/6fgg1irkcvqyb4f9f8n0nzi5gknyqhfn-gcc-mesboot1-4.7.4.drv to 'kluit.dezyne.org'
- 'build' phasesha256 hash mismatch for /gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2:
  expected hash: 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch
  actual hash:   0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
hash mismatch for store item '/gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2'
build of /gnu/store/cjim33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv failed
View build log at '/var/log/guix/drvs/cj/im33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv.bz2'.
cannot build derivation `/gnu/store/p6gfcdacjcqf2br0zwsyzx1chfvg9gxi-harfbuzz-2.4.0.drv': 1 dependencies couldn't be built
killing process 5083
--8<---------------cut here---------------end--------------->8---

The recipe for harfbuzz has a sha256 that used to be valid in April, but
hasn't been valid anymore since May, as this fix

--8<---------------cut here---------------start------------->8---
commit a8bb8fccd82a10a46f127b2235675b4f6cbaaf98
Author: Marius Bakke <mbakke@fastmail.com>
Date:   Sat May 4 18:01:12 2019 +0200

    gnu: harfbuzz: Update source hash.
    
    The previous tarball was modified in-place; see
    <https://github.com/harfbuzz/harfbuzz/issues/1641>.
    
    * gnu/packages/gtk.scm (harfbuzz)[source](sha256): Update.
--8<---------------cut here---------------end--------------->8---

shows.  Thoughts?

Greetings,
janneke


-- 
Jan Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar® http://AvatarAcademy.com

bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

Hi,


On Wed, 12 Feb 2020 at 14:44, Jan Nieuwenhuizen <janneke@gnu.org> wrote:

> Trying to travel back to Sun Apr 7 22:07:14 2019 +0200 (commit
> 56e95d54d209c2428f970d65d9b27ae4168449ad) to re-create mcrl2-minimal by
> doing
>
> --8<---------------cut here---------------start------------->8---
> guix time-machine --commit=56e95d54d209c2428f970d65d9b27ae4168449ad -- environment --ad-hoc mcrl2-minimal
> --8<---------------cut here---------------end--------------->8---

Even the simple:

--8<---------------cut here---------------start------------->8---
guix time-machine --commit=56e95d54d209c2428f970d65d9b27ae4168449ad -- help
--8<---------------cut here---------------end--------------->8---

> fails with
>
> --8<---------------cut here---------------start------------->8---
> building /gnu/store/cjim33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv...
> downloading from https://www.freedesktop.org/software/harfbuzz/release/harfbuzz-2.4.0.tar.bz2...
> |offloading build of /gnu/store/6fgg1irkcvqyb4f9f8n0nzi5gknyqhfn-gcc-mesboot1-4.7.4.drv to 'kluit.dezyne.org'
> - 'build' phasesha256 hash mismatch for /gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2:
>   expected hash: 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch
>   actual hash:   0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
> hash mismatch for store item '/gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2'
> build of /gnu/store/cjim33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv failed
> View build log at '/var/log/guix/drvs/cj/im33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv.bz2'.
> cannot build derivation `/gnu/store/p6gfcdacjcqf2br0zwsyzx1chfvg9gxi-harfbuzz-2.4.0.drv': 1 dependencies couldn't be built
> killing process 5083
> --8<---------------cut here---------------end--------------->8---

same for e.g., this commit:

--8<---------------cut here---------------start------------->8---
 guix time-machine --commit=ae528aaf19f3828d3d7d204b15570800e1bbf100 -- help
--8<---------------cut here---------------end--------------->8---


> The recipe for harfbuzz has a sha256 that used to be valid in April, but
> hasn't been valid anymore since May, as this fix
>
> --8<---------------cut here---------------start------------->8---
> commit a8bb8fccd82a10a46f127b2235675b4f6cbaaf98
> Author: Marius Bakke <mbakke@fastmail.com>
> Date:   Sat May 4 18:01:12 2019 +0200
>
>     gnu: harfbuzz: Update source hash.
>
>     The previous tarball was modified in-place; see
>     <https://github.com/harfbuzz/harfbuzz/issues/1641>.
>
>     * gnu/packages/gtk.scm (harfbuzz)[source](sha256): Update.
> --8<---------------cut here---------------end--------------->8---
>
> shows.  Thoughts?

Therefore, all the commits between the introduction of harfbuzz with
the old sha256 (commit 2da9b81837fd1e6f08a10952784d3358be982855) and
the commit updating to the new sha256 should be broken.
Roughly speaking, all the commits between April, 7th and the May, 4th;
i.e., 1100+ commits, isn't it?


Well, this ask an interesting question: how Guix can fallback when
upstream is doing wrong?


Considering this 'harbuzz' issue, is it possible to rebuild the old
tarball and push it to SoftWare Heritage? Then when a sha mismatch
happens, fallback and try to fetch it from SWH?


WDYT?

Cheers,
simon

bug#39575: guix time-machine fails when a tarball was modified in-place

[Ludovic Courtès]

Hi,

Jan Nieuwenhuizen <janneke@gnu.org> skribis:

> building /gnu/store/cjim33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv...
> downloading from https://www.freedesktop.org/software/harfbuzz/release/harfbuzz-2.4.0.tar.bz2...
> |offloading build of /gnu/store/6fgg1irkcvqyb4f9f8n0nzi5gknyqhfn-gcc-mesboot1-4.7.4.drv to 'kluit.dezyne.org'
> - 'build' phasesha256 hash mismatch for /gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2:
>   expected hash: 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch
>   actual hash:   0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
> hash mismatch for store item '/gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2'

The problem here is really that we fall back to content-addressed
mirrors instead of using them directly:

  https://issues.guix.gnu.org/issue/28659

The file itself is still available on our machines though, and you can
get it with:

  guix download -o harfbuzz-2.4.0.tar.bz2 \
  https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l

  guix download file://$PWD/harfbuzz-2.4.0.tar.bz2

After that, re-running ‘guix time-machine’ should work.

Using ci.guix.gnu.org for substitutes should have the same effect.

HTH,
Ludo’.

bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

Hi Ludo,

On Thu, 13 Feb 2020 at 22:34, Ludovic Courtès <ludo@gnu.org> wrote:

> The problem here is really that we fall back to content-addressed
> mirrors instead of using them directly:
>
>   https://issues.guix.gnu.org/issue/28659

Thank you for the pointer. Good to see that the problem is almost addressed.
I will try to understand the discussion and see what is the status of
the proposed patch.


> The file itself is still available on our machines though, and you can
> get it with:

It is an half cooked solution because the Guix project cannot archive
all; for example in term of store resources.

The content-addressed mirror should be SWH, IMHO.

Well, once sources.json will be up, it should be almost done for the future.
But we still need to push all the correct sources that are on
ci.guix.gnu.org; at least all the url based source of package.

Do you have suggestion for a plan?


>   guix download -o harfbuzz-2.4.0.tar.bz2 \
>   https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
>
>   guix download file://$PWD/harfbuzz-2.4.0.tar.bz2
>
> After that, re-running ‘guix time-machine’ should work.

Thank you. This should fix the harfbuzz mismatch issue. Cool! :-)


> Using ci.guix.gnu.org for substitutes should have the same effect.

Hum? I thought that I used ci.guix.gnu.org as substitutes... soI need to check.


Cheers,
simon

bug#39575: guix time-machine fails when a tarball was modified in-place

[Giovanni Biscuolo]

[-- Attachment #1: Type: text/plain, Size: 1102 bytes --]

Hello Ludo'

Ludovic Courtès <ludo@gnu.org> writes:

[...]

> The problem here is really that we fall back to content-addressed
> mirrors instead of using them directly:
>
>   https://issues.guix.gnu.org/issue/28659

Given the natute (AFAIU) of this issue is the very same of the bug you
mention, shouldn't this bug be merged (forcemerged?) whith #28659?

If you agree and have no time, I can try doing this

> The file itself is still available on our machines though, and you can
> get it with:
>
>   guix download -o harfbuzz-2.4.0.tar.bz2 \
>   https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
>
>   guix download file://$PWD/harfbuzz-2.4.0.tar.bz2
>
> After that, re-running ‘guix time-machine’ should work.

Does make it sense to make this workaround more general and add this as
a Cookbook or blog entry? If the patch you propose in #28659 solve any
issue and is ready to be merget of course it's not worth the effort

[...]

Thanks! Gio'

-- 
Giovanni Biscuolo

Xelera IT Infrastructures

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

Hi Giovanni,

On Fri, 14 Feb 2020 at 11:07, Giovanni Biscuolo <g@xelera.eu> wrote:
> Ludovic Courtès <ludo@gnu.org> writes:

> > The problem here is really that we fall back to content-addressed
> > mirrors instead of using them directly:
> >
> >   https://issues.guix.gnu.org/issue/28659
>
> Given the natute (AFAIU) of this issue is the very same of the bug you
> mention, shouldn't this bug be merged (forcemerged?) whith #28659?

AFAIU, there is 2 issues:
 1. how to do with the particular case of harfbuzz -- bug#39575
 2. how to solve the general case of unreliable sources -- bug#28659

So instead of merging the bugs (case 2.), I would like to solve 1.,
mark it as done and report to bug#28659. It will ease to follow
because the thread in bug#28659 is already heavy. :-)



All the best,
simon

bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

Hi Ludo,

On Thu, 13 Feb 2020 at 22:34, Ludovic Courtès <ludo@gnu.org> wrote:
>
> Hi,
>
> Jan Nieuwenhuizen <janneke@gnu.org> skribis:
>
> > building /gnu/store/cjim33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv...
> > downloading from https://www.freedesktop.org/software/harfbuzz/release/harfbuzz-2.4.0.tar.bz2...
> > |offloading build of /gnu/store/6fgg1irkcvqyb4f9f8n0nzi5gknyqhfn-gcc-mesboot1-4.7.4.drv to 'kluit.dezyne.org'
> > - 'build' phasesha256 hash mismatch for /gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2:
> >   expected hash: 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch
> >   actual hash:   0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
> > hash mismatch for store item '/gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2'
>
> The file itself is still available on our machines though, and you can
> get it with:
>
>   guix download -o harfbuzz-2.4.0.tar.bz2 \
>   https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l

Maybe I miss a point, but the file we need is the old one, not the new
one, i.e., the one with the expected hash
1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch. And I should do
wrong but ci.guix.gnu.org does not have this file -- otherwise it will
find it because of substitutes mechanism.

--8<---------------cut here---------------start------------->8---
 $ guix download -o /tmp/harfbuzz-old.tar.bz2 \
 https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch

Starting download of /tmp/harfbuzz-old.tar.bz2
From https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch...
download failed
"https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch"
404 "Not Found"
failed to download "/tmp/harfbuzz-old.tar.bz2" from
"https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch"
guix download: error: open-file: No such file or directory:
"/tmp/harfbuzz-old.tar.bz2"
--8<---------------cut here---------------end--------------->8---



Cheers,
simon

bug#39575: guix time-machine fails when a tarball was modified in-place

[Ludovic Courtès]

Hi,

zimoun <zimon.toutoune@gmail.com> skribis:

> On Thu, 13 Feb 2020 at 22:34, Ludovic Courtès <ludo@gnu.org> wrote:
>>
>> Hi,
>>
>> Jan Nieuwenhuizen <janneke@gnu.org> skribis:
>>
>> > building /gnu/store/cjim33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv...
>> > downloading from https://www.freedesktop.org/software/harfbuzz/release/harfbuzz-2.4.0.tar.bz2...
>> > |offloading build of /gnu/store/6fgg1irkcvqyb4f9f8n0nzi5gknyqhfn-gcc-mesboot1-4.7.4.drv to 'kluit.dezyne.org'
>> > - 'build' phasesha256 hash mismatch for /gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2:
>> >   expected hash: 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch
>> >   actual hash:   0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
>> > hash mismatch for store item '/gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2'
>>
>> The file itself is still available on our machines though, and you can
>> get it with:
>>
>>   guix download -o harfbuzz-2.4.0.tar.bz2 \
>>   https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
>
> Maybe I miss a point, but the file we need is the old one, not the new
> one, i.e., the one with the expected hash
> 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch.

Oops, my bad.

> And I should do wrong but ci.guix.gnu.org does not have this file --
> otherwise it will find it because of substitutes mechanism.
>
>  $ guix download -o /tmp/harfbuzz-old.tar.bz2 \
>  https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch

I checked on a bunch of machines and couldn’t find it.

Everyone, please check whether you have
/gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2 and
so share!

Ludo’.

bug#39575: guix time-machine fails when a tarball was modified in-place

[Jan Nieuwenhuizen]

Ludovic Courtès writes:

> Hi,
>
> zimoun <zimon.toutoune@gmail.com> skribis:
>
>> On Thu, 13 Feb 2020 at 22:34, Ludovic Courtès <ludo@gnu.org> wrote:
>>>
>>> Hi,
>>>
>>> Jan Nieuwenhuizen <janneke@gnu.org> skribis:
>>>
>>> > building /gnu/store/cjim33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv...
>>> > downloading from https://www.freedesktop.org/software/harfbuzz/release/harfbuzz-2.4.0.tar.bz2...
>>> > |offloading build of /gnu/store/6fgg1irkcvqyb4f9f8n0nzi5gknyqhfn-gcc-mesboot1-4.7.4.drv to 'kluit.dezyne.org'
>>> > - 'build' phasesha256 hash mismatch for /gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2:
>>> >   expected hash: 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch
>>> >   actual hash:   0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
>>> > hash mismatch for store item '/gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2'
>>>
>>> The file itself is still available on our machines though, and you can
>>> get it with:
>>>
>>>   guix download -o harfbuzz-2.4.0.tar.bz2 \
>>>   https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
>>
>> Maybe I miss a point, but the file we need is the old one, not the new
>> one, i.e., the one with the expected hash
>> 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch.
>
> Oops, my bad.
>
>> And I should do wrong but ci.guix.gnu.org does not have this file --
>> otherwise it will find it because of substitutes mechanism.
>>
>>  $ guix download -o /tmp/harfbuzz-old.tar.bz2 \
>>  https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch
>
> I checked on a bunch of machines and couldn’t find it.
>
> Everyone, please check whether you have
> /gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2 and
> so share!

What about

    https://snapshot.debian.org/archive/debian/20190406T212022Z/pool/main/h/harfbuzz/harfbuzz_2.4.0.orig.tar.bz2

(The strange thing being here, that snapshot.debian.org does not provide
a copy of the the in-place rewritten upstream tarball, either on
2019-05-06 or later.)

So, this now becomes the recipe

    wget -O harfbuzz-2.4.0.tar.bz2 https://snapshot.debian.org/archive/debian/20190406T212022Z/pool/main/h/harfbuzz/harfbuzz_2.4.0.orig.tar.bz2
    guix download $PWD/harfbuzz-2.4.0.tar.bz2
    guix time-machine --commit=56e95d54d209c2428f970d65d9b27ae4168449ad --no-offload -- help

that i'm trying now, and for now it looks fine (lots of stuff to build,
i'll report success or failure when it's done).

It seems, however, that for offload builds to work the guix download
needs to be repeated on the offload build farm machines too?

janneke

-- 
Jan Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar® http://AvatarAcademy.com

bug#39575: guix time-machine fails when a tarball was modified in-place

[Ludovic Courtès]

Hi,

Jan Nieuwenhuizen <janneke@gnu.org> skribis:

> What about
>
>     https://snapshot.debian.org/archive/debian/20190406T212022Z/pool/main/h/harfbuzz/harfbuzz_2.4.0.orig.tar.bz2

Good idea.

> So, this now becomes the recipe
>
>     wget -O harfbuzz-2.4.0.tar.bz2 https://snapshot.debian.org/archive/debian/20190406T212022Z/pool/main/h/harfbuzz/harfbuzz_2.4.0.orig.tar.bz2
>     guix download $PWD/harfbuzz-2.4.0.tar.bz2
>     guix time-machine --commit=56e95d54d209c2428f970d65d9b27ae4168449ad --no-offload -- help
>
> that i'm trying now, and for now it looks fine (lots of stuff to build,
> i'll report success or failure when it's done).

OK!

> It seems, however, that for offload builds to work the guix download
> needs to be repeated on the offload build farm machines too?

No, I don’t think so, because the head node copies all the inputs to
build machines before it actually offloads the build.

Ludo’.

bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

Hi,

On Fri, 14 Feb 2020 at 14:51, Ludovic Courtès <ludo@gnu.org> wrote:
> Jan Nieuwenhuizen <janneke@gnu.org> skribis:

> > What about
> >
> >     https://snapshot.debian.org/archive/debian/20190406T212022Z/pool/main/h/harfbuzz/harfbuzz_2.4.0.orig.tar.bz2
>
> Good idea.

Cool!
But how do you determine the "date", i.e., this reference '20190406T212022Z' ?

Could it be automated?


Cheers,
simon

bug#39575: guix time-machine fails when a tarball was modified in-place

[Tobias Geerinckx-Rice via Bug reports for GNU Guix]

[-- Attachment #1: Type: text/plain, Size: 1056 bytes --]

Jan, Simon,

Janneke 写道：
> https://snapshot.debian.org/archive/debian/20190406T212022Z/pool/main/h/harfbuzz/harfbuzz_2.4.0.orig.tar.bz2

This is a wonderful resource!  Thank you, Janneke (and Debian)!

zimoun 写道：
> Cool!
> But how do you determine the "date", i.e., this reference 
> '20190406T212022Z' ?

You'd take the timestamp immediately preceding your desired (Guix) 
commit's date, or something like that.  The fact that git commit 
dates aren't linear shouldn't hurt here.

> Could it be automated?

Not without parsing HTML to get the valid timestamps: 
<https://snapshot.debian.org/archive/debian/?year=2020&month=2>.

Also, this doesn't seem to be a supported service yet[0]:

  “This is an implementation for a possible snapshot.debian.org 
  service.
   It's not yet finished, it's more a prototype/proof of concept 
   to show
   and learn what we want and can provide.  So far it seems to 
   actually work.”

Still really cool,

T G-R

[0]: https://salsa.debian.org/snapshot-team/snapshot

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

bug#39575: guix time-machine fails when a tarball was modified in-place

[Bengt Richter]

On +2020-02-15 21:01:36 +0100, Tobias Geerinckx-Rice via Bug reports for GNU Guix wrote:
> Jan, Simon,
> 
> Janneke 写道：
> > https://snapshot.debian.org/archive/debian/20190406T212022Z/pool/main/h/harfbuzz/harfbuzz_2.4.0.orig.tar.bz2
> 
> This is a wonderful resource!  Thank you, Janneke (and Debian)!
> 
> zimoun 写道：
> > Cool!
> > But how do you determine the "date", i.e., this reference
> > '20190406T212022Z' ?
> 
> You'd take the timestamp immediately preceding your desired (Guix) commit's
> date, or something like that.  The fact that git commit dates aren't linear
> shouldn't hurt here.
> 
> > Could it be automated?
> 
> Not without parsing HTML to get the valid timestamps:
> <https://snapshot.debian.org/archive/debian/?year=2020&month=2>.
>

You may not need to parse the html fully if the part you need is
isolatable into delimited scopes that you can successively narrow.

For example, I while back I wanted a command I could type to get
the url of the latest linux kernel at kernel.org:

stable-kernel.scm -h 
--8<---------------cut here---------------start------------->8---
Usage: stable-kernel-scm [ -h ]
       -h for this message
          (without args):
       go to https://www.kernel.org/ to wget page,
       extract URL of latest stable release tarball
       and write that URL to stdout.
--8<---------------cut here---------------end--------------->8---
(oops, I see I din't use $0 in the usage text -- should be .scm, not -scm)

I offer it below [1], with the thought that you could probably
modify (not to mention improve :-) it to get the timestamps you want.
Especially if you could get them to make the narrow context unique enough
that it's delimiters can delimit it in one shot.

The page at kernel.org is apparently stable enough that this still works,
but YMMV until the snapshot page is similarly stable. (You could ask
them to make it easy :)

> Also, this doesn't seem to be a supported service yet[0]:
> 
>  “This is an implementation for a possible snapshot.debian.org  service.
>   It's not yet finished, it's more a prototype/proof of concept   to show
>   and learn what we want and can provide.  So far it seems to   actually
> work.”
> 
> Still really cool,
> 
> T G-R
> 
> [0]: https://salsa.debian.org/snapshot-team/snapshot

HTH or is useful some way.
-- 
Regards,
Bengt Richter

[1]
--8<---------------cut here---------------start------------->8---
#!/usr/bin/bash
exec guile -e main -s "$0" "$@"
!#
;;;; stable-kernel.scm
;;;; goes to https://www.kernel.org/ to wget page, then
;;;; extracts name of latest stable release tarball to stdout

;;;;
(define (usage)
  (format (current-error-port)
	  (string-join
	   '(
	    "Usage: stable-kernel-scm [ -h ]"
	    "       -h for this message"
	    "          (without args):"
            "       go to https://www.kernel.org/ to wget page,"
	    "       extract URL of latest stable release tarball"
	    "       and write that URL to stdout."
	    "")
	   "\n")))

(use-modules (ice-9 format))
(use-modules (ice-9 rdelim))
(use-modules (ice-9 popen))
(use-modules (ice-9 textual-ports))
(use-modules (ice-9 and-let-star))
(use-modules (ice-9 regex))

(define (extract-delimited str s-beg s-end)
  (and-let*
   ((ix-beg (string-contains str s-beg))
    (ix-post-beg (+ ix-beg (string-length s-beg)))
    (ix-end   (string-contains str s-end ix-post-beg)))
   (substring str ix-post-beg  ix-end)))

(define kernel-url "https://www.kernel.org/")

(define (get-kern-name)
  (let*((cmd-kern (string-append "wget -q -O - " kernel-url))
	(p-inp (open-input-pipe cmd-kern))
	(wgot-pinp-str (get-string-all p-inp))
	(extracted-table-releases
	 (extract-delimited wgot-pinp-str
			    "<table id=\"releases\">"
			    "</table>"))
	(extracted-stable-tarball-anchor
	 (extract-delimited extracted-table-releases
			    "<td>stable:</td>"
			    ">tarball<"))
	(extracted-stable-href
	 (extract-delimited extracted-stable-tarball-anchor
			    "<a href=\""
			    "\"")))
   (begin
    extracted-stable-href)))

(define (main args)
  (begin
    (set! args (cdr args)) ;; always dump callee arg
    (if (not (pair? args))
	(set! args '("-do-default") ))

    ;; simple opts...
    (cond
     ((string-prefix? "-h" (car args))
      (begin (usage)
	     (exit)))
     
     ((string-prefix? "-do-default" (car args))
      (let ((kern-name (get-kern-name)))
	(display kern-name)(newline)))

     (#t
      (begin
	(format (current-error-port) "Error: bad opt: '~a'\n" (car args))
	(usage)
	(exit))))))
--8<---------------cut here---------------end--------------->8---

bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

Hi,

On Sat, 15 Feb 2020 at 21:01, Tobias Geerinckx-Rice <me@tobias.gr> wrote:

> Janneke 写道：
> > https://snapshot.debian.org/archive/debian/20190406T212022Z/pool/main/h/harfbuzz/harfbuzz_2.4.0.orig.tar.bz2
>
> This is a wonderful resource!  Thank you, Janneke (and Debian)!
>
> zimoun 写道：
> > Cool!
> > But how do you determine the "date", i.e., this reference
> > '20190406T212022Z' ?
>
> You'd take the timestamp immediately preceding your desired (Guix)
> commit's date, or something like that.  The fact that git commit
> dates aren't linear shouldn't hurt here.

You assume that Debian packs packages as fast as Guix, I mean on the
same schedule which is a strong assumption IMHO.
For example, if it was the contrary and the "new" release of harfbuzz
2.4.0 were missing, then would Debian be helpful?


> Also, this doesn't seem to be a supported service yet[0]:
>
>   “This is an implementation for a possible snapshot.debian.org
>   service.
>    It's not yet finished, it's more a prototype/proof of concept
>    to show
>    and learn what we want and can provide.  So far it seems to
>    actually work.”
>
> Still really cool,

Yes, still cool! :-)


Thanks,
simon

bug#39575: guix time-machine fails when a tarball was modified in-place

[Efraim Flashner]

[-- Attachment #1: Type: text/plain, Size: 1845 bytes --]

On Mon, Feb 17, 2020 at 09:47:41AM +0100, zimoun wrote:
> Hi,
> 
> On Sat, 15 Feb 2020 at 21:01, Tobias Geerinckx-Rice <me@tobias.gr> wrote:
> 
> > Janneke 写道：
> > > https://snapshot.debian.org/archive/debian/20190406T212022Z/pool/main/h/harfbuzz/harfbuzz_2.4.0.orig.tar.bz2
> >
> > This is a wonderful resource!  Thank you, Janneke (and Debian)!
> >
> > zimoun 写道：
> > > Cool!
> > > But how do you determine the "date", i.e., this reference
> > > '20190406T212022Z' ?
> >
> > You'd take the timestamp immediately preceding your desired (Guix)
> > commit's date, or something like that.  The fact that git commit
> > dates aren't linear shouldn't hurt here.
> 
> You assume that Debian packs packages as fast as Guix, I mean on the
> same schedule which is a strong assumption IMHO.
> For example, if it was the contrary and the "new" release of harfbuzz
> 2.4.0 were missing, then would Debian be helpful?
> 
> 

We could first try
mirror://debian/pool/main/harfbuzz/harfbuzz_2.4.0.orig.tar.bz2

and then scrape https://snapshot.debian.org/package/harfbuzz/ for
2.4.0-1 and then parse the website for harfbuzz_2.4.0.orig.tar.bz2. Or
for just 'orig.tar'

> > Also, this doesn't seem to be a supported service yet[0]:
> >
> >   “This is an implementation for a possible snapshot.debian.org
> >   service.
> >    It's not yet finished, it's more a prototype/proof of concept
> >    to show
> >    and learn what we want and can provide.  So far it seems to
> >    actually work.”
> >
> > Still really cool,
> 
> Yes, still cool! :-)
> 
> 
> Thanks,
> simon
> 
> 
> 

-- 
Efraim Flashner   <efraim@flashner.co.il>   אפרים פלשנר
GPG key = A28B F40C 3E55 1372 662D  14F7 41AA E7DC CA3D 8351
Confidentiality cannot be guaranteed on emails sent or received unencrypted

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

bug#39575: guix time-machine fails when a tarball was modified in-place

[Tobias Geerinckx-Rice via Bug reports for GNU Guix]

[-- Attachment #1: Type: text/plain, Size: 162 bytes --]

zimoun 写道：
> You assume that Debian packs packages as fast as Guix

Indeed I do!  :-D

Efraim's solution sounds reasonable.

Kind regards,

T G-R

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

HI Jan,

On Fri, 14 Feb 2020 at 14:24, Jan Nieuwenhuizen <janneke@gnu.org> wrote:

This command

> >>  $ guix download -o /tmp/harfbuzz-old.tar.bz2 \
> >>  https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch

now works.


However, this command

$ guix time-machine \
     --commit=56e95d54d209c2428f970d65d9b27ae4168449ad -- help

still fails for me with the message:

--8<---------------cut here---------------start------------->8---
[...]
building /gnu/store/gglbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv...
- 'check' phasebuilder for
`/gnu/store/gglbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv' failed
with exit code 1
build of /gnu/store/gglbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv failed
View build log at
'/var/log/guix/drvs/gg/lbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv.bz2'.
cannot build derivation
`/gnu/store/yqpxm07zm0mirrdvl2c4qvf8biyzg468-guix-56e95d54d.drv': 1
dependencies couldn't be built
cannot build derivation
`/gnu/store/7z7p0m7abi246gzigw8as2q3w33k1n31-profile.drv': 1
dependencies couldn't be built
guix time-machine: error: build of
`/gnu/store/7z7p0m7abi246gzigw8as2q3w33k1n31-profile.drv' failed
--8<---------------cut here---------------end--------------->8---

The log /var/log/guix/drvs/gg/lbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv.bz2
is not meaningful for me... but I can report it here.


> that i'm trying now, and for now it looks fine (lots of stuff to build,
> i'll report success or failure when it's done).

Well, is it a success or a failure for you?


Cheers,
simon

bug#39575: guix time-machine fails when a tarball was modified in-place

[Jan Nieuwenhuizen]

zimoun writes:

Hi Simon,

> On Fri, 14 Feb 2020 at 14:24, Jan Nieuwenhuizen <janneke@gnu.org> wrote:
>
> This command
>
>> >>  $ guix download -o /tmp/harfbuzz-old.tar.bz2 \
>> >>  https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch
>
> now works.
>
>
> However, this command
>
> $ guix time-machine \
>      --commit=56e95d54d209c2428f970d65d9b27ae4168449ad -- help
>
> still fails for me with the message:
>
> [...]
> building /gnu/store/gglbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv...
> - 'check' phasebuilder for
> `/gnu/store/gglbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv' failed
> with exit code 1
> build of /gnu/store/gglbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv failed
> View build log at
> '/var/log/guix/drvs/gg/lbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv.bz2'.
> cannot build derivation
> `/gnu/store/yqpxm07zm0mirrdvl2c4qvf8biyzg468-guix-56e95d54d.drv': 1
> dependencies couldn't be built
> cannot build derivation
> `/gnu/store/7z7p0m7abi246gzigw8as2q3w33k1n31-profile.drv': 1
> dependencies couldn't be built
> guix time-machine: error: build of
> `/gnu/store/7z7p0m7abi246gzigw8as2q3w33k1n31-profile.drv' failed
>
> The log /var/log/guix/drvs/gg/lbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv.bz2
> is not meaningful for me... but I can report it here.
>
>
>> that i'm trying now, and for now it looks fine (lots of stuff to build,
>> i'll report success or failure when it's done).
>
> Well, is it a success or a failure for you?

For me, pythohn-minimal fails to build

    build-started /gnu/store/s0lw23myd3hvpw28sffkhz8b30x1hcz0-python-minimal-3.7.3.drv - x86_64-linux /var/log/guix/drvs/s0//lw23myd3hvpw28sffkhz8b30x1hcz0-python-minimal-3.7.3.drv.bz2 20827

    ======================================================================
    FAIL: test_register_chain (test.test_faulthandler.FaultHandlerTests)
    ----------------------------------------------------------------------
    Traceback (most recent call last):
      File "/tmp/guix-build-python-minimal-3.7.3.drv-0/Python-3.7.3/Lib/test/test_faulthandler.py", line 724, in test_register_chain
        self.check_register(chain=True)
      File "/tmp/guix-build-python-minimal-3.7.3.drv-0/Python-3.7.3/Lib/test/test_faulthandler.py", line 702, in check_register
        self.assertEqual(exitcode, 0)
    AssertionError: -11 != 0

    ----------------------------------------------------------------------

    Ran 42 tests in 18.782s

    FAILED (failures=1, skipped=4)
    1 test failed again:
        test_faulthandler

    == Tests result: FAILURE then FAILURE ==

    382 tests OK.

    1 test failed:
        test_faulthandler

Not sure what to do here.  Could this be a (harmless) coincident?

Greetings
janneke

-- 
Jan Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar® http://AvatarAcademy.com

bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

Hi,

Some follow ups.


For me, now, the command

> > $ guix time-machine \
> >      --commit=56e95d54d209c2428f970d65d9b27ae4168449ad -- help

does not fail anymore at the Guile step:

--8<---------------cut here---------------start------------->8---
  building /gnu/store/gglbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv...
  - 'check' phasebuilder for
  `/gnu/store/gglbrs8j0iq8ygh55inwfvpwb5z2x254-guile-2.2.4.drv' failed
--8<---------------cut here---------------end--------------->8---

because it was about "FAIL: test-out-of-memory".


Well, now, I am failing at the Python 3.7.3 step too:

--8<---------------cut here---------------start------------->8---
  /gnu/store/s0lw23myd3hvpw28sffkhz8b30x1hcz0-python-minimal-3.7.3.drv failed
--8<---------------cut here---------------end--------------->8---


However, the python error seems about TLS:

--8<---------------cut here---------------start------------->8---
test.test_asyncio.test_windows_utils (unittest.loader.ModuleSkipped)
... test test_asyncio failed
skipped 'Windows only'

======================================================================
ERROR: test_start_tls_server_1
(test.test_asyncio.test_sslproto.SelectorStartTLSTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/tmp/guix-build-python-minimal-3.7.3.drv-0/Python-3.7.3/Lib/test/test_asyncio/test_sslproto.py",
line 507, in test_start_tls_server_1
    self.loop.run_until_complete(run_main())
  File "/tmp/guix-build-python-minimal-3.7.3.drv-0/Python-3.7.3/Lib/asyncio/base_events.py",
line 584, in run_until_complete
    return future.result()
  File "/tmp/guix-build-python-minimal-3.7.3.drv-0/Python-3.7.3/Lib/test/test_asyncio/test_sslproto.py",
line 502, in run_main
    loop=self.loop, timeout=self.TIMEOUT)
  File "/tmp/guix-build-python-minimal-3.7.3.drv-0/Python-3.7.3/Lib/asyncio/tasks.py",
line 423, in wait_for
    raise futures.TimeoutError()
concurrent.futures._base.TimeoutError
--8<---------------cut here---------------end--------------->8---


> Not sure what to do here.  Could this be a (harmless) coincident?

Me neither.


Cheers,
simon

bug#39575: guix time-machine fails when a tarball was modified in-place

[Ludovic Courtès]

Hi,

zimoun <zimon.toutoune@gmail.com> skribis:

> Well, now, I am failing at the Python 3.7.3 step too:
>
>   /gnu/store/s0lw23myd3hvpw28sffkhz8b30x1hcz0-python-minimal-3.7.3.drv failed
>
>
>
> However, the python error seems about TLS:
>
> test.test_asyncio.test_windows_utils (unittest.loader.ModuleSkipped)
> ... test test_asyncio failed
> skipped 'Windows only'
>
> ======================================================================
> ERROR: test_start_tls_server_1
> (test.test_asyncio.test_sslproto.SelectorStartTLSTests)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
>   File "/tmp/guix-build-python-minimal-3.7.3.drv-0/Python-3.7.3/Lib/test/test_asyncio/test_sslproto.py",
> line 507, in test_start_tls_server_1
>     self.loop.run_until_complete(run_main())
>   File "/tmp/guix-build-python-minimal-3.7.3.drv-0/Python-3.7.3/Lib/asyncio/base_events.py",
> line 584, in run_until_complete
>     return future.result()
>   File "/tmp/guix-build-python-minimal-3.7.3.drv-0/Python-3.7.3/Lib/test/test_asyncio/test_sslproto.py",
> line 502, in run_main
>     loop=self.loop, timeout=self.TIMEOUT)
>   File "/tmp/guix-build-python-minimal-3.7.3.drv-0/Python-3.7.3/Lib/asyncio/tasks.py",
> line 423, in wait_for
>     raise futures.TimeoutError()
> concurrent.futures._base.TimeoutError

It seems to be timing-sensitive, is it deterministic?

One lesson here is that we should keep substitutes for a longer amount
of time—we actually have a lot of storage space on berlin so we should
investigate what happened.  (The NixOS folks currently keep substitutes
forever but they found it’s starting to be expensive…)

Thanks,
Ludo’.

bug#39575: guix time-machine fails when a tarball was modified in-place

[Tobias Geerinckx-Rice via Bug reports for GNU Guix]

[-- Attachment #1: Type: text/plain, Size: 608 bytes --]

zimoun 写道：
> Maybe I miss a point, but the file we need is the old one, not 
> the new
> one, i.e., the one with the expected hash
> 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch. And I 
> should do

~ λ guix download 
https://www.tobias.gr/guix/harfbuzz-2.4.0.tar.bz2
Starting download of /tmp/guix-file.JSWxOl
From https://www.tobias.gr/guix/harfbuzz-2.4.0.tar.bz2...
 ...4.0.tar.bz2  17.1MiB 
 6.5MiB/s 00:03 [##################] 100.0%
/gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2
1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch

Enjoy,

T G-R

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

bug#39575: guix time-machine fails when a tarball was modified in-place

[Ludovic Courtès]

Hi,

Tobias Geerinckx-Rice <me@tobias.gr> skribis:

> zimoun 写道：
>> Maybe I miss a point, but the file we need is the old one, not the
>> new
>> one, i.e., the one with the expected hash
>> 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch. And I should
>> do
>
> ~ λ guix download https://www.tobias.gr/guix/harfbuzz-2.4.0.tar.bz2

Thanks, you saved us!

Now we have it here:

  https://ci.guix.gnu.org/file/harfbuzz-2.4.0.tar.bz2/sha256/1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch

I’ve also registered a GC root.

Anyway, everything will be so much better when SWH archives tarballs!

Ludo’.

bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

Hi,

On Fri, 14 Feb 2020 at 14:14, Ludovic Courtès <ludo@gnu.org> wrote:
> Tobias Geerinckx-Rice <me@tobias.gr> skribis:

> > ~ λ guix download https://www.tobias.gr/guix/harfbuzz-2.4.0.tar.bz2
>
> Thanks, you saved us!

Thank you! :-)


> Anyway, everything will be so much better when SWH archives tarballs!

The future will be better. :-)
Even some details need to be discussed: frequency of source.json
generation, frequency of the SWH crawler ingest it, etc.

How could all the archives living in ci.guix.gnu.org be sent to SWH?
Because IMHO the of Berlin (or other) is to archive the world but to
build it. :-)


Cheers,
simon

bug#39575: guix time-machine fails when a tarball was modified in-place

[Jan Nieuwenhuizen]

Ludovic Courtès writes:

> Jan Nieuwenhuizen <janneke@gnu.org> skribis:
>
>> building /gnu/store/cjim33x0q1bv1ppkv3qijvr1pvsn4y0q-harfbuzz-2.4.0.tar.bz2.drv...
>> downloading from https://www.freedesktop.org/software/harfbuzz/release/harfbuzz-2.4.0.tar.bz2...
>> |offloading build of /gnu/store/6fgg1irkcvqyb4f9f8n0nzi5gknyqhfn-gcc-mesboot1-4.7.4.drv to 'kluit.dezyne.org'
>> - 'build' phasesha256 hash mismatch for /gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2:
>>   expected hash: 1mpah6kwqid1kxsj4rwqsniivqbrx231j65v51yncx6s0dch0dch
>>   actual hash:   0vrkvdlmihdg62a4c6h5kx27khc33xmb95l50zgnwnavvpwyyw5l
>> hash mismatch for store item '/gnu/store/b4cdp9sp44848348lrpzbfafhmjqf8nr-harfbuzz-2.4.0.tar.bz2'
>
> The problem here is really that we fall back to content-addressed
> mirrors instead of using them directly:
>
>   https://issues.guix.gnu.org/issue/28659

Wait, what happened here; you finally proposed a patch two years ago and
nothing happened/we all forgot to follow up?

I cannot determine if possibly we were hoping to "wait" for the guile
build daemon?

janneke

-- 
Jan Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar® http://AvatarAcademy.com

bug#39575: guix time-machine fails when a tarball was modified in-place

[Ludovic Courtès]

Jan Nieuwenhuizen <janneke@gnu.org> skribis:

> Ludovic Courtès writes:

[...]

>> The problem here is really that we fall back to content-addressed
>> mirrors instead of using them directly:
>>
>>   https://issues.guix.gnu.org/issue/28659
>
> Wait, what happened here; you finally proposed a patch two years ago and
> nothing happened/we all forgot to follow up?

I think we forgot, indeed.

One thing I don’t quite like about the patch is the fact that ‘guix
substitutes’ connects to the daemon in ‘content-addressed-item?’.

Also, one could argue that we’d steer users towards downloading from our
server, which could be a privacy concern (probably not a strong argument
since one can easily change the substitute URLs.)

Thoughts?

Ludo’.

bug#28659: bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

Hi,

On Fri, 14 Feb 2020 at 22:34, Ludovic Courtès <ludo@gnu.org> wrote:

> Also, one could argue that we’d steer users towards downloading from our
> server, which could be a privacy concern (probably not a strong argument
> since one can easily change the substitute URLs.)

I am not following the privacy concern.
What do you mean?

Cheers,
simon

bug#28659: bug#39575: guix time-machine fails when a tarball was modified in-place

[Ludovic Courtès]

Hi!

zimoun <zimon.toutoune@gmail.com> skribis:

> On Fri, 14 Feb 2020 at 22:34, Ludovic Courtès <ludo@gnu.org> wrote:
>
>> Also, one could argue that we’d steer users towards downloading from our
>> server, which could be a privacy concern (probably not a strong argument
>> since one can easily change the substitute URLs.)
>
> I am not following the privacy concern.
> What do you mean?

I mean that by default, someone who’s disabled substitutes (presumably
out of security or privacy concerns) would find themself downloading
source code from ci.guix.gnu.org instead of various upstream sites.

Ludo’.

bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

Hi Ludo,

On Sun, 16 Feb 2020 at 11:59, Ludovic Courtès <ludo@gnu.org> wrote:
> zimoun <zimon.toutoune@gmail.com> skribis:
> > On Fri, 14 Feb 2020 at 22:34, Ludovic Courtès <ludo@gnu.org> wrote:

> >> Also, one could argue that we’d steer users towards downloading from our
> >> server, which could be a privacy concern (probably not a strong argument
> >> since one can easily change the substitute URLs.)
> >
> > I am not following the privacy concern.
> > What do you mean?
>
> I mean that by default, someone who’s disabled substitutes (presumably
> out of security or privacy concerns) would find themself downloading
> source code from ci.guix.gnu.org instead of various upstream sites.

I do not see the difference between mirroring and traveling back in
time with missing upstream sources.
And because it is content-addressed, it seems even more secure than
downloading from a upstream URL, IMHO.
If one trusts Guix, then an attacker needs to corrupt in the same time
the Guix history and Berlin (and/or any other farm).
If one does not trust Guix, why does they use the recipe coming from
Guix? To be precise, this person has to check all the recipes of all
the dependencies.

Well, I do not see a security concern because we are talking about
serving the sources.
It is another story when the substitutes serve the results of the
build (binaries); because one does not have any strong guarantee that
the substitute serves the expected binaries.

By privacy concern, do you mean that Guix could collect who downloads
what; in a central fashion? Which is not the case when one downloads
from several distributed upstream sources. Right?
Well, I am not convinced because the case of missing upstream source
is rare. And it is easy to protect against such collecting data
process.
In paranoid mode, traveling back in time is becoming difficult because
of the reliability of the sources; I mean if the sources were
reliable, SWH would not exist. ;-) The solution should be an IPFS /
GNUnet / full distributed archive... which is not ready... yet! :-)


Well, maybe for the TODO list of the time-machine: add an option to
allow substitutes *only* for the sources (substitutes meaning
ci.guix.gnu.org and/or SWH). If this option does not exist yet. ;-)


Cheers,
simon

bug#28659: bug#39575: guix time-machine fails when a tarball was modified in-place

[Ludovic Courtès]

Hi,

zimoun <zimon.toutoune@gmail.com> skribis:

> On Sun, 16 Feb 2020 at 11:59, Ludovic Courtès <ludo@gnu.org> wrote:
>> zimoun <zimon.toutoune@gmail.com> skribis:
>> > On Fri, 14 Feb 2020 at 22:34, Ludovic Courtès <ludo@gnu.org> wrote:
>
>> >> Also, one could argue that we’d steer users towards downloading from our
>> >> server, which could be a privacy concern (probably not a strong argument
>> >> since one can easily change the substitute URLs.)
>> >
>> > I am not following the privacy concern.
>> > What do you mean?
>>
>> I mean that by default, someone who’s disabled substitutes (presumably
>> out of security or privacy concerns) would find themself downloading
>> source code from ci.guix.gnu.org instead of various upstream sites.

[...]

> By privacy concern, do you mean that Guix could collect who downloads
> what; in a central fashion? Which is not the case when one downloads
> from several distributed upstream sources. Right?

Exactly.  But like I wrote above, I don’t think it’s a strong argument.

What remains is the issue with ‘content-addressed-item?’, then.

Ludo’.

bug#28659: bug#39575: guix time-machine fails when a tarball was modified in-place

[zimoun]

On Mon, 17 Feb 2020 at 15:40, Ludovic Courtès <ludo@gnu.org> wrote:

> Exactly.  But like I wrote above, I don’t think it’s a strong argument.

I agree and the big picture depends on the audience.
Scientific communities would be fine with centralized archives such as
SWH. And only centralized archives IMHO can provide a reliable "long
term" support which is the point for that communities. (Quote because
not clearly defined what it is. :-))
Other communities would prefer distributed archive such as IPFS or
GNUnet but 1. it still needs some work and 2. the "long term" is not
guarantee by nature, IMHO. But it is probably not an issue for that
communities.


> What remains is the issue with ‘content-addressed-item?’, then.

I agree.
The bridge with SWH is in good shape, IMHO.
And the pending IPFS patch would deserve more love. :-) Maybe soon...



Cheers,
simon

bug#28659: Content-addressed mirror is not used upon invalid hash

[zimoun]

Hi,

On Fri, 14 Feb 2020 at 22:34, Ludovic Courtès <ludo@gnu.org> wrote:

> One thing I don’t quite like about the patch is the fact that ‘guix
> substitutes’ connects to the daemon in ‘content-addressed-item?’.

What is the status of this patch [1] following the recent discussion about
tar “disarchive” and SWH?

Related:
 - http://issues.guix.gnu.org/issue/39575
 - http://issues.guix.gnu.org/42162
 - https://git.ngyro.com/disarchive/
 
All the best,
simon

[1] http://issues.guix.gnu.org/issue/28659#26

bug#28659: Content-addressed mirror is not used upon invalid hash

[Ludovic Courtès]

Hello,

zimoun <zimon.toutoune@gmail.com> skribis:

> On Fri, 14 Feb 2020 at 22:34, Ludovic Courtès <ludo@gnu.org> wrote:
>
>> One thing I don’t quite like about the patch is the fact that ‘guix
>> substitutes’ connects to the daemon in ‘content-addressed-item?’.
>
> What is the status of this patch [1] following the recent discussion about
> tar “disarchive” and SWH?
>
> Related:
>  - http://issues.guix.gnu.org/issue/39575
>  - http://issues.guix.gnu.org/42162
>  - https://git.ngyro.com/disarchive/

Thanks for the reminder.  I don’t think Timothy’s work changes anything
wrt. to this issue: it would still need to be addressed.

Ludo’.