On 08/04/21 at 09:06 +0200, Mattia Rizzolo wrote:

- source format 1.0 with direct changes in .diff.gz (no patch system)

Also dpatch. And also 1.0+quilt (ugh). !

So the breakdown for testing is:

27 1.0, dpatch
166 1.0, quilt
374 1.0, no changes
395 1.0, direct changes
660 3.0 (native)
29379 3.0 (quilt)

Indeed we could probably only allow 1.0 without changes to upstream
sources, and 3.0 formats.

Lucas

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Thu, Apr 08, 2021 at 02:58:14PM +0200, Lucas Nussbaum wrote:

On 08/04/21 at 09:06 +0200, Mattia Rizzolo wrote:

Also dpatch. And also 1.0+quilt (ugh). !

So the breakdown for testing is:

27 1.0, dpatch

I'll take upon myself to get rid of this set RSN.

166 1.0, quilt

And also think something about this, though I'd have to see some data
about them, but I'm interested in also moving this set forward.

374 1.0, no changes

I think it would be interesting to see data about "freshness" of these
sources, maintainers, etc. I'd expect that all the non-native of this
set could likely be made 3.0+quilt, but who knows...

395 1.0, direct changes

And, as said in the other thread, I'd leave these alone for now (maybe I
should query and see whether some are orphaned, unmaintained, etc, but generally leave them alone).

Indeed we could probably only allow 1.0 without changes to upstream
sources, and 3.0 formats.

1.0 without changes still has details that I don't like. For example
the fact that the upstream tarball can only be .gz. And the passerby contributor trying to add a patch will surely be confused with how
different the behaviour is compared to huge majority of all the other
packages, so I'd still try to move them over to 3.0, if the maintainer
doesn't push back.

--
regards,
Mattia Rizzolo

GPG Key: 66AE 2B4A FCCF 3F52 DA18 4D18 4B04 3FCD B944 4540 .''`.
More about me: https://mapreri.org : :' : Launchpad user: https://launchpad.net/~mapreri `. `'`
Debian QA page: https://qa.debian.org/developer.php?login=mattia `-

-----BEGIN PGP SIGNATURE-----

iQIzBAEBCgAdFiEEi3hoeGwz5cZMTQpICBa54Yx2K60FAmBvF9YACgkQCBa54Yx2 K62D2hAAnvAyDICuM6yLMWdYU/IVhSWJudYfSnYJ2bQQKo5Q233W5eBXpwCrdzAi Rr6IZRx88wYnG/0o0IoqitogtE5GZEQSruyK2jyxeEyxETafYgtOxCls1wVcnHCO NizN7zBUYvE71ZoGtkkixvLT1mQ3kEHVeRtja9kff/vvMXIxKEODT19I/v1fciV6 cVbwgjVvsUfoLKjF9rUcQehqQi6bgJXU5BdliYdeisdaKtpaMP0Y8/57ZfVpIe4z 4zlYptap2AUbe6bdwvviyIaHkmTqeZNgsARpBOQawkHwwUOJCBoesmN32PoTn5rd UXp/wRI4vrTwMZp4S8yvtFz0fPD409SCPBtOQozuFkbX1oJlVulH7sB2wClxr8H6 Iko5vhRRngILY8R/hOKZcFBqtLW389LYtk0v9SwdEHxkG/SwcGDmJ6p5cVihC+sX V3ESWusDdYfw5gC61TLKd3/lSrZu7FQwkG2dK2+vOqMO+UzfEX76UPpv3Coek3PT 3y5zhoybQUBeB7wM723xt6GeWb1MrlRrAVUx5cHemQWKiP+PTvo

On Thu, Apr 08, 2021 at 02:58:14PM +0200, Lucas Nussbaum wrote:

166 1.0, quilt

I don't see what's wrong with these.


--
cheers,
Holger

⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ holger@(debian|reproducible-builds|layer-acht).org
⢿⡄⠘⠷⠚⠋⠀ PGP fingerprint: B8BF 5413 7B09 D35C F026 FE9D 091A B856 069A AA1C
⠈⠳⣄

We live in a world where teenagers get more and more desperate trying to convince adults to behave like grown ups.

-----BEGIN PGP SIGNATURE-----

iQIzBAABCgAdFiEEuL9UE3sJ01zwJv6dCRq4VgaaqhwFAmBvJuIACgkQCRq4Vgaa qhzJ5w/+PQ44rM4Ie4clEyqWfny8K6nqRqicq1XMd42h5uTUqLUxV8MlXZtfqmth It1tf697icxb3Zp5ekoExXdshteFm0livvdP0rUeb5bFEcUy1dwLbZpUyjQiO09z sK11PBrhfwR+lQEUtMI2dC0d8hyq9MsiKYnVTya9mfIoPVYA5OUBW8tbUdmkS8UK 8A3guD5dABwqfTxcjhCGHadM2WX4slIMmBqL0G6/VxwSjETmPDWV3rvrATEm8+y8 n4feEpYUe4bveCidZvryPwMvC9GQZR4VyWdxVMTPq2fvDBLumzyJ3oA/glqg+p1d ECjKw90MFiA7Fi3Ibb48Odb1CVO98OYgcjMq978DV5jHNO8Bzoi/atO/HIMza2VD BCQGxPCcS+ZwJURXWRVwMN3qATZPbdHYZ4oDA8P6tp64I7mvn9ZmTWa+RzI/PsO0 AIK5qUqxAvdIZflvMG0kdBgCVw0C+w1PY7qJhDKF0uMVEZ6AfE1wLvp6gW9rDsqK kQlp8Lbqop2qYBtyQcvmJOVYx8wx+4ptYUA/gh81NSt

On Thu, Apr 08, 2021 at 03:53:06PM +0000, Holger Levsen wrote:

On Thu, Apr 08, 2021 at 02:58:14PM +0200, Lucas Nussbaum wrote:

166 1.0, quilt

I don't see what's wrong with these.

Nothing *wrong* as the hard meaning of that word.
But:
* They carry the usual set of downsides of 1.0 vs 3.0, like:
- no support for .tar.(bz2|xz|…)
- no support for multi tarballs
* possibility of bugs due to the implementations of the patch/unpatch
routines in d/rules
* also similar to the above, you can't assume the state of the
unpacked source (patched or unpatched?)
* they are different from no good reason (0.5% vs 94.8%), and
consistency in complex setup carry some good points by itself

And at the same time I can't really think of any good point of keeping
them 1.0.


IMHO, they aren't "wrong" or "inherently bad", but I believe keeping
them that way is more of technical debt than anything else.

--
regards,
Mattia Rizzolo

GPG Key: 66AE 2B4A FCCF 3F52 DA18 4D18 4B04 3FCD B944 4540 .''`.
More about me: https://mapreri.org : :' : Launchpad user: https://launchpad.net/~mapreri `. `'`
Debian QA page: https://qa.debian.org/developer.php?login=mattia `-

-----BEGIN PGP SIGNATURE-----

iQIzBAEBCgAdFiEEi3hoeGwz5cZMTQpICBa54Yx2K60FAmBvKRgACgkQCBa54Yx2 K621+A//ZmVoxbwUgue8DS8JnUn/zkTIyeyXZM/xtFh1k35Z/KHnSjdReJJYUrUb snPa/Qrh180r6smcqaoW6X5a7JRPWaK+Sk00zoZU6eONZY9Ow9rRsoM5xiTY/Kxh FOmS+9QfTu6HYPNJ/Anv9KrRAUK3MHmOHDrrswHfr9zvfYX9IGV8cxYFHdZc2+wI exGti+YMDB0pkKkuY9sPC9K13gpk/oPMLD+QGmOHz7RnqxQ07JKMRF/LEn8gdtYg RNov7ZPiAbL5mDSmh1oRSFG+wbUIfW3hHpgNaLXJgHRG8V7RF2IxpLNj1gDKhXPR dyNmul63SNiLHK4qGbm6WqqFpXR7/2zNsxnSy299HGbbmUX2RVCiHWRd5zRdmBZr 7BNG0MKMHgsyFA5SuJFszfSPjKDvLA+3nTQF6+SqdA+P61HpR+G+40zGqAaeZ7eh upxaWwXe7l0RsRXmKdXBTnYvHSmIaKWu1vYcDbmGZ/hgTtSXWXWWo3ejwhythehr fHpGx7np+tRamJlsf0uGGn+9Gb1C8GEKIRFbh43mE7KbEGzW7wx

On Thu, Apr 08, 2021 at 04:15:45PM +0000, Holger Levsen wrote:

IMHO, they aren't "wrong" or "inherently bad", but I believe keeping
them that way is more of technical debt than anything else.

right, so the severity of these bugs should be wishlist or maybe normal,
but I don't think important would be justified, and serious seriously not.

Yes, totally. I don't think anybody ever talked about the severity of
any such bugs, did they? :) Please don't overthink, within these
threads nobody has yet to "threaten" maintainers or anything else.

I also disagree with the nuance of "decide that we no longer want to
accept some packaging practices" in general in the scope of the "smell"
and "trends", as all of them really are only about best practices and consolidation from my point of view. I just would love to get rid of
special cases as much as possible when they are not needed.

Because things are the way they are, things will not stay the way they are. (Bertolt Brecht)

See, your signature agrees ↑↑↑ ;P

--
regards,
Mattia Rizzolo

GPG Key: 66AE 2B4A FCCF 3F52 DA18 4D18 4B04 3FCD B944 4540 .''`.
More about me: https://mapreri.org : :' : Launchpad user: https://launchpad.net/~mapreri `. `'`
Debian QA page: https://qa.debian.org/developer.php?login=mattia `-

-----BEGIN PGP SIGNATURE-----

iQIzBAEBCgAdFiEEi3hoeGwz5cZMTQpICBa54Yx2K60FAmBvNPYACgkQCBa54Yx2 K62Q0BAAvU9YiwSB4RT6tfZ01zy7m9ApgpM/pjq7ycD/fTca36ZhRDtV2SM9+fvY RP8ti9hxtydnJSh5sGVcan3j1S94wAUMC5hKVCXo130G/H/gfXHFf8F1daEV3dRb sHDALFoHYcvTlE+lL/iqryY/juazM+DeyMKOGO0H1N76l8fuhU9tUhnRdxW0HJoJ WNPM/LmUUrfkphp9aZRXI79y/QpYD82hbQPkb9p9YKvASbqITM0yHCh3mYoT/V5Q u+VBR5JrpP3nszHIZjjnhBZ9xYinW9nsKQ/0T93NSpUUrxuejY1iOhdzNIkLRzau wSa/hAWIChyRMAKNlzDZ2c+H+lHk2E5Bjs8yBqiz6r1+u9eyAtpzpTnFmqkWrp09 GUG0ywpzEbkL0eL+4YCiazWUQn+ApFIkDyMhg8vm/IUzouFmYpaXvBTx5Kpp6+rZ Cs/pp2X2FJg5KqEvlihMDQtY3xqtTfgMW0v9SRiUD8wBbihSK0+Me/dea1P+m0ps sw73qZq45KCilOCqoQWPbidkyaccCQePtSLqShgOUwIgN9YsMxy

On Thu, Apr 08, 2021 at 06:02:35PM +0200, Mattia Rizzolo wrote:

Nothing *wrong* as the hard meaning of that word.
But:
* They carry the usual set of downsides of 1.0 vs 3.0, like:
- no support for .tar.(bz2|xz|…)
- no support for multi tarballs

those don't seem to be relevant for those packages.

* possibility of bugs due to the implementations of the patch/unpatch
routines in d/rules
* also similar to the above, you can't assume the state of the
unpacked source (patched or unpatched?)
* they are different from no good reason (0.5% vs 94.8%), and
consistency in complex setup carry some good points by itself

And at the same time I can't really think of any good point of keeping
them 1.0.

change also has a costs.

IMHO, they aren't "wrong" or "inherently bad", but I believe keeping
them that way is more of technical debt than anything else.

right, so the severity of these bugs should be wishlist or maybe normal,
but I don't think important would be justified, and serious seriously not.


--
cheers,
Holger

⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ holger@(debian|reproducible-builds|layer-acht).org
⢿⡄⠘⠷⠚⠋⠀ PGP fingerprint: B8BF 5413 7B09 D35C F026 FE9D 091A B856 069A AA1C
⠈⠳⣄

Because things are the way they are, things will not stay the way they are. (Bertolt Brecht)

-----BEGIN PGP SIGNATURE-----

iQIzBAABCgAdFiEEuL9UE3sJ01zwJv6dCRq4VgaaqhwFAmBvLDEACgkQCRq4Vgaa qhzm2w//VlwKnAgQtF9PkeI9UJasXFDJhPlfYUich9HGl3kIqB7mmK6QwwYZnIFR nMvhoA3EmRE2lxvXwnGKJQfgm3LqfkM3y0Chw6wW3RYeKgmBLxrkkyAN4LUWMcO1 2VfZq0EKPpzlJJ1mwutBpFuxL4GaXRsc4/i3mIiqcs0ieRo40O9lzVwu3wxelYMk Kz1lRJN96TTm+ZNkRj8Yl3LlPPKqiwrHjY9YViLjpqPrSsfewBihDmIbYGm/ozs7 wEXGLvjA+H5FSt7caKnL5Q48XuHQCtOXXVoXTWrT6tNsEnTXbSMse1uflIblwWeC aPuAr4LgCWSyuOT6BIX9zszUYKInYkZf/bjxWh53oMcYBTC6XHIs+CejvER/CDVA 4XZBefvWO5k/ARUo7xVhxOQxft4AIf/Q2E33d3McjtAr2DpGv1AD4t/66qG/1/Ci o5UjTp6N2ASHEP8SZKBW7xOYhea437rIcfjP1tbEgUbEeKayZdJHJdROGwwJRk85 1aw10Tl0k/cy7duMV8qBunFsJocNIICkSooS5yc9pwq5mwNnLqI+0qD62zBw+js7

On 2021-04-08 18:02 +0200, Mattia Rizzolo wrote:

On Thu, Apr 08, 2021 at 03:53:06PM +0000, Holger Levsen wrote:

On Thu, Apr 08, 2021 at 02:58:14PM +0200, Lucas Nussbaum wrote:

166 1.0, quilt

I don't see what's wrong with these.

Nothing *wrong* as the hard meaning of that word.
But:
* They carry the usual set of downsides of 1.0 vs 3.0, like:
- no support for .tar.(bz2|xz|…)
- no support for multi tarballs
* possibility of bugs due to the implementations of the patch/unpatch
routines in d/rules
* also similar to the above, you can't assume the state of the
unpacked source (patched or unpatched?)
* they are different from no good reason (0.5% vs 94.8%), and
consistency in complex setup carry some good points by itself

And at the same time I can't really think of any good point of keeping
them 1.0.

IMHO, they aren't "wrong" or "inherently bad", but I believe keeping
them that way is more of technical debt than anything else.

The X Strike Force is still sticking to format 1.0, with one of the main reasons being that it makes it easier to cherry-pick one or several
upstream commits. In the 3.0 format you have to create a separate patch
and later remove it when merging in the next upstream version.

Whether this outweighs the disadvantages of the 1.0 format is debatable,
but I think it would be best to start a discussion on the debian-x list
or the #debian-x IRC channel before filing individual bugs.

Cheers,
Sven

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Thu, Apr 08, 2021 at 06:53:12PM +0200, Mattia Rizzolo wrote:

right, so the severity of these bugs should be wishlist or maybe normal, but I don't think important would be justified, and serious seriously not.

Yes, totally. I don't think anybody ever talked about the severity of
any such bugs, did they?
[...] I just would love to get rid of
special cases as much as possible when they are not needed.

well, 'getting rid of as much as possible' sounds to me like filing RC bugs
and let autoremovals do the work...

Because things are the way they are, things will not stay the way they are. (Bertolt Brecht)

See, your signature agrees ↑↑↑ ;P

:) my sig is not always serious, see below :) and 1.0 packages are slowly
going away - and I dont think we need to accellerate this. There are enough areas where we should do that OTOH :)


--
cheers,
Holger

⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ holger@(debian|reproducible-builds|layer-acht).org
⢿⡄⠘⠷⠚⠋⠀ PGP fingerprint: B8BF 5413 7B09 D35C F026 FE9D 091A B856 069A AA1C
⠈⠳⣄

This is the year of gpg on the desktop! (Gunnar Wolf)

-----BEGIN PGP SIGNATURE-----

iQIzBAABCgAdFiEEuL9UE3sJ01zwJv6dCRq4VgaaqhwFAmBwmHAACgkQCRq4Vgaa qhyO/Q/+Mnf7beSCi5dZFnO8wD5CnaPBXOi88ri2uzNv7N0zZkIH3yZoFZJ6eDu4 iQRjHIu1xcv0ep9wMxaGEPPQM5UvCQ2KieF7uvTgiT/HVRCnudMauibNhabfSwXK 6ZKYl3i8Y4aTM12Ml3c37xhcFTqirzuCJjbUAxnywcE10w5iQqzWakraX3y3kmHw 5DX5LUuT/B0ObrXKu58GyunlMiziTn+xeKLL4Tb0eycmXLGDuePPWrbNirmLr4Mg qxHSqI415DsqhsKNB9qtqArsJSfhM3iZMI/r6vf3jcLbjCJGd72mroY7mOF+9fBO nvx0tVuuY98iYTFy4XGJCpxM6/fR5j+TEhvpRd21M+nIpSTolm5OXB2IvcWl1Cyz iW1F1e/d5HJqW5+T1mITdzrnEmQuD3Td6ML1jw1+CuDfzmxsrIoA/im7/2NzOmn3 Vx9dSPI/1RGMmXcuJaqMdbmJIMo+BPUB6Dny1ryJzEKaA8jGSeQN14PEhXOkMZVV MylTmUiRhw/HGkhH9rb1rVfWN/+CC426b2N8qbRH4EJ8F5HbOV1jeKt/JTJDlHsM 4aKJeMQldYmcIdm/b6/bzHi7UDmvod614aJp/sR

On 09/04/21 at 19:49 +0200, Sven Joachim wrote:

The X Strike Force is still sticking to format 1.0, with one of the main reasons being that it makes it easier to cherry-pick one or several
upstream commits. In the 3.0 format you have to create a separate patch
and later remove it when merging in the next upstream version.

Would that be true also with single-debian-patch?

Also, while I was aware of this workflow for the X team, it seems that
there's at least one outlier: https://salsa.debian.org/xorg-team/app/xterm/-/tree/debian-unstable/debian/patches

Whether this outweighs the disadvantages of the 1.0 format is debatable,
but I think it would be best to start a discussion on the debian-x list
or the #debian-x IRC channel before filing individual bugs.

Sure, I take this discussion as a preliminary discussion before a
potential discussion, later, about a mass bug filing. It's not clear if
we will get there.

Lucas

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 2021-04-09 21:02 +0200, Lucas Nussbaum wrote:

On 09/04/21 at 19:49 +0200, Sven Joachim wrote:

The X Strike Force is still sticking to format 1.0, with one of the main
reasons being that it makes it easier to cherry-pick one or several
upstream commits. In the 3.0 format you have to create a separate patch
and later remove it when merging in the next upstream version.

Would that be true also with single-debian-patch?

Long lived patches are still managed with quilt, so single-debian-patch
is not an option. Thus, you often get a mixture of quilt patches and
direct changes.

Also, while I was aware of this workflow for the X team, it seems that there's at least one outlier: https://salsa.debian.org/xorg-team/app/xterm/-/tree/debian-unstable/debian/patches

More to the point, xterm actually uses format 3.0 now. That is because upstream does not use git, and because the only person in Uploaders is
an outlier in the X Strike Force. ;-)

Whether this outweighs the disadvantages of the 1.0 format is debatable,
but I think it would be best to start a discussion on the debian-x list
or the #debian-x IRC channel before filing individual bugs.

Sure, I take this discussion as a preliminary discussion before a
potential discussion, later, about a mass bug filing. It's not clear if
we will get there.

Cheers,
Sven

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)