shadow: update to 4.16.0. #48813

dataCobra · 2024-02-18T12:12:10Z

Testing the changes

I tested the changes in this PR: NOT YET

Local build testing

I built this PR locally for my native architecture, (x86_64-glibc)
I built this PR locally for these architectures (if supported. mark crossbuilds):
- x86_64-musl
- i686

I welcome everyone to test this version. Maybe also on a new installation.

dataCobra · 2024-02-18T13:34:04Z

The file /etc/default/useradd is no longer created by default. Instead now the patched useradd binary is aware of the defaults that we provided before with the old version in /etc/default/useradd.

srcpkgs/shadow/patches/xstrdup.patch

srcpkgs/shadow/patches/useradd-defaults.patch

dkwo · 2024-02-19T21:27:14Z

Some distros are also carrying this patch https://git.alpinelinux.org/aports/tree/community/shadow/fix-undefined-reference.patch Maybe it's useful?

dkwo · 2024-02-19T22:08:23Z

As a reference, /usr/bin/lastlog and its manpage are now gone, and there are new /usr/bin/getsubids and its manpage

dkwo · 2024-02-19T22:10:55Z

also the file login.defs seems outdated.
distros like arch and chimera patch it instead of replacing it.

dataCobra · 2024-02-20T10:05:57Z

Some distros are also carrying this patch https://git.alpinelinux.org/aports/tree/community/shadow/fix-undefined-reference.patch Maybe it's useful?

I've added the patch and everything still builds fine.

As a reference, /usr/bin/lastlog and its manpage are now gone, and there are new /usr/bin/getsubids and its manpage

Forgot to add the new configuration argument to add lastlog. Fixed with the new push.

also the file login.defs seems outdated. distros like arch and chimera patch it instead of replacing it.

I've checked what Arch did and added and modified patch 2 and 3 so they fit our needs. The file will now be installed from shadow itself and patched beforehand. I removed the fixed file we provided.

dkwo · 2024-02-20T14:23:24Z

about the patches from arch linux: they may need more adaptating to our needs. for example, yescrypt needs a build option if i remember right.

dataCobra · 2024-02-20T15:13:07Z

I've added yescrypt to be build for shadow to make sure we got a secure password hash.

dataCobra · 2024-02-20T15:14:18Z

CC: @Gottox

Could you as a maintainer please check this as well? 🙂

dkwo · 2024-02-20T16:18:48Z

what is the reason for removing xstrdup.patch instead of updating it?

dataCobra · 2024-02-22T10:01:19Z

what is the reason for removing xstrdup.patch instead of updating it?

What do you want to update?

The function does no longer exist and the file is also removed.

In lib/alloc.c a comment says:

/* Replacements for malloc and strdup with error checking.  Too trivial
   to be worth copyrighting :-).  I did that because a lot of code used
   malloc and strdup without checking for NULL pointer, and I like some
   message better than a core dump...  --marekm

   Yeh, but.  Remember that bailing out might leave the system in some
   bizarre state.  You really want to put in error checking, then add
   some back-out failure recovery code. -- jfh */

As I understand the patch is no longer needed.

dkwo · 2024-02-22T13:44:27Z

see https://github.com/chimera-linux/cports/blob/master/main/shadow/patches/xstrdup.patch

dataCobra · 2024-02-22T15:19:13Z

see https://github.com/chimera-linux/cports/blob/master/main/shadow/patches/xstrdup.patch

Thank you for the link.

dkwo · 2024-02-24T20:40:49Z

does it make sense to disable RUSEROK for all libc through a patch, instead of selectively in pre_configure?
same for groups(1): instead of in do_build, can this be done in a patch?
is the use of a license file still needed?
i think we should have only one patch for login.defs, which can be void-specific

dkwo · 2024-02-24T20:42:59Z

e.g. see https://gitlab.archlinux.org/archlinux/packaging/packages/shadow/-/blob/main/0001-Disable-replaced-tools-their-man-pages-and-PAM-integ.patch?ref_type=heads and https://github.com/chimera-linux/cports/blob/master/main/shadow/patches/disable-ruserok.patch

dataCobra · 2024-02-25T14:31:02Z

does it make sense to disable RUSEROK for all libc through a patch, instead of selectively in pre_configure?

I've added the patch you recommended from chimera linux.

same for groups(1): instead of in do_build, can this be done in a patch?

We could do that in a patch, but I feel like the way in the template is more convenient and easier to update.

is the use of a license file still needed?

I'll check that.

i think we should have only one patch for login.defs, which can be void-specific

Agree. For the moment I've only modified the patches a bit. But refactoring and cleaning up the files in the patches folder after we've finished the decision which patches to include must be done. Otherwise it might be hard to update the package in the future.

dkwo · 2024-02-25T16:00:13Z

We could do that in a patch, but I feel like the way in the template is more convenient and easier to update.

alternatively, could it be moved to either post_configure or pre_build?

dkwo · 2024-03-06T16:23:02Z

btw, 4.14.6 is out, soon maybe 4.15

alejandro-colomar · 2024-03-09T12:44:50Z

Some distros are also carrying this patch https://git.alpinelinux.org/aports/tree/community/shadow/fix-undefined-reference.patch Maybe it's useful?

That patch shouldn't be necesary. Please don't apply it unless you find a reason to. And if you do (find a build error that requires the patch), please report it as an upstream bug.

Would you mind asking the Alpine maintainer if they can comment on that patch? I'm interested in fixing upstream if there's something broken, but I'd like to learn what's broken, because that patch looks like a red herring.

srcpkgs/shadow/patches/fix-undefined-reference.patch

alejandro-colomar · 2024-03-09T13:01:12Z

srcpkgs/shadow/patches/xstrdup.patch

+ inline char *
+ xstrdup(const char *str)
 {
-+	if(str == NULL) return NULL;
- 	return strcpy (xmalloc (strlen (str) + 1), str);
+	if (str == NULL) return NULL;
+ 	return strcpy(XMALLOC(strlen(str) + 1, char), str);


These xfoo() functions are designed to exit/abort on errors reported by the corresponding foo() function.

Thus, xstrdup() is strdup(3) plus a check of an output NULL from strdup(3) to error out. The precise reason to have these functions is so that they can never return NULL.

Adding null checks to a NULL input is tangential to the purpose of these functions, and defeats their purpose. Now the function can return NULL, which needs to be checked by the caller.

So, if you really want to check for programmer errors, you'd need to exit on a NULL input, not just pass it through.

But you probably shouldn't do that at all. strdup(3) doesn't handle NULL input, and I don't see a reason why xstrdup() should.

This patch goes back a long time (introduced in Void in 2012, shadow 4.1.5), so many things may have changed. If you say it's not needed, I tend to trust you. Maybe @q66 can comment as to why Chimera still keeps it?

Hmmm, I'd prefer to know why it originated, before confirming that it should be removed, but most likely. Let's hear @q66 . :)

dkwo · 2024-03-09T18:00:22Z

@alejandro-colomar Thanks a lot for taking a look. Do you think it would be possible to make stuff that conflicts with coreutils and util-linux (e.g. groups) optional via a configure? right now most distros have to patch it (and their man) out (see the patch taken from arch). Ditto for respecting usr/bin and usr/sbin config options.

For the Alpine patch, maybe can you take a look at this https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/50121 pull request? the first thread has a discussion about why this was done.

alejandro-colomar · 2024-03-09T18:26:33Z

Hi!

@alejandro-colomar Thanks a lot for taking a look.

=)

Do you think it would be possible to make stuff that conflicts with coreutils and util-linux (e.g. groups) optional via a configure?

I'm neutral to that, but others seem to not like the idea. (Now I see you're the same one that reported the issue shadow-maint/shadow#842.) How about opening an issue, not asking to make it conditional, but rather reporting the conflict with other projects? Maybe that's more convincing. You could document which distros use shadow's groups, and which distros use others' groups. Maybe we could merge the efforts from those other projects and shadow into a single groups implementation. While some competition is good, it might be good to merge at some point.

right now most distros have to patch it (and their man) out (see the patch taken from arch). Ditto for respecting usr/bin and usr/sbin config options.

You know what? I would wipe out the entire autotools-based build system, which has been more problematic than anything else. I would write a hand-written GNUmakefile that allows more flexibility. But some distro maintainers (cough, Gentoo, cough) opposed strongly.

Please, please, report a bug in shadow. That will add up to the current issues with the build system. :)

For the Alpine patch, maybe can you take a look at this https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/50121 pull request? the first thread has a discussion about why this was done.

Yup, I found that MR a moment ago, and sent an email https://lists.sr.ht/~hallyn/shadow/%3CZeyg8ClVMNeRifua%40debian%3E, https://lists.alpinelinux.org/~alpine/aports/%3CZeyg8ClVMNeRifua%40debian%3E.

dataCobra · 2024-03-10T15:39:51Z

Thank you both for all the input and information.

I'm currently working on an update to 4.15.0.

dkwo · 2024-03-22T22:52:09Z

it may also be possible to drop the ruserok patch

alejandro-colomar · 2024-03-22T23:11:59Z

it may also be possible to drop the ruserok patch

If you investigate that ruserok thing, you may want to have fun upstream with the "quick hack":

https://github.com/shadow-maint/shadow/blob/ead55e9ba8958504e23e29545f90c4dd925c7462/configure.ac#L162
shadow-maint/shadow@428a207
shadow-maint/shadow@45c6603#diff-bb21aa33a3f69ccb36c68b220f40ad08f29b9cd2c05dfedae7b9e3d5d4d08f6bR196

:)

dataCobra · 2024-03-24T19:40:57Z

@Gottox are you able to help as the maintainer of the package?

If you have some more knowledge/information.

dkwo · 2024-05-20T21:18:50Z

btw, 4.15.1 is out

dataCobra · 2024-06-10T07:45:18Z

Hey, I'm currently not having the time to work on this on my own.

Help from the Maintainer @Gottox would be nice or another person who is capable of helping me out.

dataCobra · 2024-07-14T16:00:02Z

Updated the PR to shadow 4.16.0. But didn't had the time to check the package.

Just updating to make sure everyone who wants to can have a look.
Help and discussion is appreciated.

github-actions · 2024-10-17T01:59:33Z

Pull Requests become stale 90 days after last activity and are closed 14 days after that. If this pull request is still relevant bump it or assign it.

dataCobra force-pushed the shadow branch 3 times, most recently from a1b07c7 to 2701019 Compare February 18, 2024 13:00

dkwo reviewed Feb 19, 2024

View reviewed changes

srcpkgs/shadow/patches/xstrdup.patch Show resolved Hide resolved

dkwo reviewed Feb 19, 2024

View reviewed changes

srcpkgs/shadow/patches/useradd-defaults.patch Show resolved Hide resolved

dataCobra force-pushed the shadow branch from 2701019 to 5b62251 Compare February 20, 2024 10:03

dataCobra force-pushed the shadow branch from 5b62251 to 0eb6856 Compare February 20, 2024 15:11

dataCobra force-pushed the shadow branch from 0eb6856 to 028b253 Compare February 22, 2024 12:02

dataCobra force-pushed the shadow branch from 028b253 to 2d48622 Compare February 24, 2024 17:38

dataCobra requested review from sgn and dkwo February 24, 2024 17:52

dataCobra force-pushed the shadow branch from 2d48622 to d7101d7 Compare February 25, 2024 14:26

alejandro-colomar reviewed Mar 9, 2024

View reviewed changes

srcpkgs/shadow/patches/fix-undefined-reference.patch Outdated Show resolved Hide resolved

alejandro-colomar reviewed Mar 9, 2024

View reviewed changes

dataCobra marked this pull request as draft March 20, 2024 06:31

dataCobra force-pushed the shadow branch 3 times, most recently from a05dcf8 to d3597ad Compare March 22, 2024 19:09

dataCobra changed the title ~~shadow: update to 4.14.5.~~ shadow: update to 4.15.0. Mar 22, 2024

alejandro-colomar mentioned this pull request May 17, 2024

Remote login issue and questions shadow-maint/shadow#992

Closed

dataCobra force-pushed the shadow branch from d3597ad to aec780c Compare July 14, 2024 15:57

dataCobra changed the title ~~shadow: update to 4.15.0.~~ shadow: update to 4.16.0. Jul 14, 2024

This comment was marked as off-topic.

Sign in to view

alejandro-colomar mentioned this pull request Jul 14, 2024

Package request: liba2i #51261

Closed

github-actions bot added Stale and removed Stale labels Oct 17, 2024

shadow: update to 4.16.0.

1aecd2f

dataCobra force-pushed the shadow branch from aec780c to 1aecd2f Compare October 18, 2024 08:31

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

shadow: update to 4.16.0. #48813

shadow: update to 4.16.0. #48813

dataCobra commented Feb 18, 2024 •

edited

Loading

dataCobra commented Feb 18, 2024

dkwo commented Feb 19, 2024

dkwo commented Feb 19, 2024

dkwo commented Feb 19, 2024

dataCobra commented Feb 20, 2024 •

edited

Loading

dkwo commented Feb 20, 2024

dataCobra commented Feb 20, 2024

dataCobra commented Feb 20, 2024

dkwo commented Feb 20, 2024

dataCobra commented Feb 22, 2024

dkwo commented Feb 22, 2024

dataCobra commented Feb 22, 2024

dkwo commented Feb 24, 2024

dkwo commented Feb 24, 2024

dataCobra commented Feb 25, 2024 •

edited

Loading

dkwo commented Feb 25, 2024

dkwo commented Mar 6, 2024

alejandro-colomar commented Mar 9, 2024

alejandro-colomar Mar 9, 2024

dkwo Mar 9, 2024

alejandro-colomar Mar 9, 2024

dkwo commented Mar 9, 2024

alejandro-colomar commented Mar 9, 2024 •

edited

Loading

dataCobra commented Mar 10, 2024

dkwo commented Mar 22, 2024

alejandro-colomar commented Mar 22, 2024 •

edited

Loading

dataCobra commented Mar 24, 2024

dkwo commented May 20, 2024

dataCobra commented Jun 10, 2024

dataCobra commented Jul 14, 2024

This comment was marked as off-topic.

github-actions bot commented Oct 17, 2024

shadow: update to 4.16.0. #48813

Are you sure you want to change the base?

shadow: update to 4.16.0. #48813

Conversation

dataCobra commented Feb 18, 2024 • edited Loading

Testing the changes

Local build testing

dataCobra commented Feb 18, 2024

dkwo commented Feb 19, 2024

dkwo commented Feb 19, 2024

dkwo commented Feb 19, 2024

dataCobra commented Feb 20, 2024 • edited Loading

dkwo commented Feb 20, 2024

dataCobra commented Feb 20, 2024

dataCobra commented Feb 20, 2024

dkwo commented Feb 20, 2024

dataCobra commented Feb 22, 2024

dkwo commented Feb 22, 2024

dataCobra commented Feb 22, 2024

dkwo commented Feb 24, 2024

dkwo commented Feb 24, 2024

dataCobra commented Feb 25, 2024 • edited Loading

dkwo commented Feb 25, 2024

dkwo commented Mar 6, 2024

alejandro-colomar commented Mar 9, 2024

alejandro-colomar Mar 9, 2024

Choose a reason for hiding this comment

dkwo Mar 9, 2024

Choose a reason for hiding this comment

alejandro-colomar Mar 9, 2024

Choose a reason for hiding this comment

dkwo commented Mar 9, 2024

alejandro-colomar commented Mar 9, 2024 • edited Loading

dataCobra commented Mar 10, 2024

dkwo commented Mar 22, 2024

alejandro-colomar commented Mar 22, 2024 • edited Loading

dataCobra commented Mar 24, 2024

dkwo commented May 20, 2024

dataCobra commented Jun 10, 2024

dataCobra commented Jul 14, 2024

This comment was marked as off-topic.

github-actions bot commented Oct 17, 2024

dataCobra commented Feb 18, 2024 •

edited

Loading

dataCobra commented Feb 20, 2024 •

edited

Loading

dataCobra commented Feb 25, 2024 •

edited

Loading

alejandro-colomar commented Mar 9, 2024 •

edited

Loading

alejandro-colomar commented Mar 22, 2024 •

edited

Loading