zsh-workers
 help / color / mirror / code / Atom feed
* Speaking of 5.8.1.3-test ...
@ 2022-05-07  2:59 Bart Schaefer
  2022-05-07  3:50 ` dana
                   ` (4 more replies)
  0 siblings, 5 replies; 13+ messages in thread
From: Bart Schaefer @ 2022-05-07  2:59 UTC (permalink / raw)
  To: Zsh hackers list

[-- Attachment #1: Type: text/plain, Size: 54 bytes --]

Has anyone tried this at all?  No reports +/- so far.

[-- Attachment #2: Type: text/html, Size: 81 bytes --]

^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-07  2:59 Speaking of 5.8.1.3-test Bart Schaefer
@ 2022-05-07  3:50 ` dana
  2022-05-07  4:06 ` Daniel Shahaf
                   ` (3 subsequent siblings)
  4 siblings, 0 replies; 13+ messages in thread
From: dana @ 2022-05-07  3:50 UTC (permalink / raw)
  To: Zsh hackers list

On Fri 6 May 2022, at 21:59, Bart Schaefer wrote:
> Has anyone tried this at all?  No reports +/- so far.

I've been running it without issues on Catalina, though aside from
running `make check` when i posted it i haven't gone out of my way to
test anything in particular

dana


^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-07  2:59 Speaking of 5.8.1.3-test Bart Schaefer
  2022-05-07  3:50 ` dana
@ 2022-05-07  4:06 ` Daniel Shahaf
  2022-05-08 14:26   ` Axel Beckert
  2022-05-08  4:10 ` Matthew Martin
                   ` (2 subsequent siblings)
  4 siblings, 1 reply; 13+ messages in thread
From: Daniel Shahaf @ 2022-05-07  4:06 UTC (permalink / raw)
  To: zsh-workers

Bart Schaefer wrote on Sat, 07 May 2022 02:59 +00:00:
> Has anyone tried this at all?  No reports +/- so far.

I'm running 5.8.1.2-test.  Haven't gotten around to testing/installing .3.


^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-07  2:59 Speaking of 5.8.1.3-test Bart Schaefer
  2022-05-07  3:50 ` dana
  2022-05-07  4:06 ` Daniel Shahaf
@ 2022-05-08  4:10 ` Matthew Martin
  2022-05-09  6:52 ` Kamil Dudka
  2022-05-10  4:05 ` Jun T
  4 siblings, 0 replies; 13+ messages in thread
From: Matthew Martin @ 2022-05-08  4:10 UTC (permalink / raw)
  To: zsh-workers

On Fri, May 06, 2022 at 07:59:09PM -0700, Bart Schaefer wrote:
> Has anyone tried this at all?  No reports +/- so far.

Been running it on OpenBSD for a few days without issues.


^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-07  4:06 ` Daniel Shahaf
@ 2022-05-08 14:26   ` Axel Beckert
  2022-05-09  7:07     ` Luna Jernberg
  2022-05-09  9:58     ` Axel Beckert
  0 siblings, 2 replies; 13+ messages in thread
From: Axel Beckert @ 2022-05-08 14:26 UTC (permalink / raw)
  To: zsh-workers

Hi,

On Sat, May 07, 2022 at 04:06:34AM +0000, Daniel Shahaf wrote:
> Bart Schaefer wrote on Sat, 07 May 2022 02:59 +00:00:
> > Has anyone tried this at all?  No reports +/- so far.
> 
> I'm running 5.8.1.2-test.  Haven't gotten around to
> testing/installing .3.

Same here for the Debian package. Hope to get it done this evening.

		Kind regards, Axel
-- 
PGP: 2FF9CD59612616B5      /~\  Plain Text Ribbon Campaign, http://arc.pasp.de/
Mail: abe@deuxchevaux.org  \ /  Say No to HTML in E-Mail and Usenet
Mail+Jabber: abe@noone.org  X
https://axel.beckert.ch/   / \  I love long mails: https://email.is-not-s.ms/


^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-07  2:59 Speaking of 5.8.1.3-test Bart Schaefer
                   ` (2 preceding siblings ...)
  2022-05-08  4:10 ` Matthew Martin
@ 2022-05-09  6:52 ` Kamil Dudka
  2022-05-10  4:05 ` Jun T
  4 siblings, 0 replies; 13+ messages in thread
From: Kamil Dudka @ 2022-05-09  6:52 UTC (permalink / raw)
  To: Bart Schaefer; +Cc: Zsh hackers list

On Saturday, May 7, 2022 4:59:09 AM CEST Bart Schaefer wrote:
> Has anyone tried this at all?  No reports +/- so far.

I was able to successfully build RPM packages out of it and pass all tests
in Fedora and EPEL build roots:

    https://copr.fedorainfracloud.org/coprs/kdudka/zsh-next/build/4380208/

Thanks!

Kamil




^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-08 14:26   ` Axel Beckert
@ 2022-05-09  7:07     ` Luna Jernberg
  2022-05-09  9:58     ` Axel Beckert
  1 sibling, 0 replies; 13+ messages in thread
From: Luna Jernberg @ 2022-05-09  7:07 UTC (permalink / raw)
  To: Zsh hackers list, abe, Martin Jernberg

[-- Attachment #1: Type: text/plain, Size: 807 bytes --]

Hey!

Works fine on my Raspberry Pi with Raspberry Pi OS (Debian)

On Sun, May 8, 2022 at 6:38 PM Axel Beckert <abe@deuxchevaux.org> wrote:

> Hi,
>
> On Sat, May 07, 2022 at 04:06:34AM +0000, Daniel Shahaf wrote:
> > Bart Schaefer wrote on Sat, 07 May 2022 02:59 +00:00:
> > > Has anyone tried this at all?  No reports +/- so far.
> >
> > I'm running 5.8.1.2-test.  Haven't gotten around to
> > testing/installing .3.
>
> Same here for the Debian package. Hope to get it done this evening.
>
>                 Kind regards, Axel
> --
> PGP: 2FF9CD59612616B5      /~\  Plain Text Ribbon Campaign,
> http://arc.pasp.de/
> Mail: abe@deuxchevaux.org  \ /  Say No to HTML in E-Mail and Usenet
> Mail+Jabber: abe@noone.org  X
> https://axel.beckert.ch/   / \  I love long mails:
> https://email.is-not-s.ms/
>
>

[-- Attachment #2: Type: text/html, Size: 1546 bytes --]

^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-08 14:26   ` Axel Beckert
  2022-05-09  7:07     ` Luna Jernberg
@ 2022-05-09  9:58     ` Axel Beckert
  2022-05-10 17:28       ` Axel Beckert
  1 sibling, 1 reply; 13+ messages in thread
From: Axel Beckert @ 2022-05-09  9:58 UTC (permalink / raw)
  To: zsh-workers

Hi again,

On Sun, May 08, 2022 at 04:26:21PM +0200, Axel Beckert wrote:
> On Sat, May 07, 2022 at 04:06:34AM +0000, Daniel Shahaf wrote:
> > Bart Schaefer wrote on Sat, 07 May 2022 02:59 +00:00:
> > > Has anyone tried this at all?  No reports +/- so far.
> > 
> > I'm running 5.8.1.2-test.  Haven't gotten around to
> > testing/installing .3.
> 
> Same here for the Debian package. Hope to get it done this evening.

Looks good so far: All release architectures build successfully again
thanks to Jun T's patches against 5.8.1.3-test:

https://buildd.debian.org/status/package.php?p=zsh&suite=experimental

Four (rather exotic or vintage) non-release architectures still
haven't been built yet, but all which have been built, are fine.

One of those who haven't built 5.8.1.3-test is alpha. And alpha had
two test suite failures with 5.8.1.2-test (.2-test, not .3-test). Back
then it was only build quite some days after the upload, so I didn't
notice the test suite failures until I looked on that page when
waiting for the current upload being picked up by the build daemons.

So here are the test suite failures of 5.8.1.2-test on the alpha
architecture:

Running test: delete all word on other characters
--- /tmp/zsh.ztst.19000/ztst.out	2022-04-19 05:05:01.494426862 +0000
+++ /tmp/zsh.ztst.19000/ztst.tout	2022-04-19 05:05:01.547161643 +0000
@@ -1,2 +1,2 @@
 BUFFER:     wordword
-CURSOR: 4
+URSOR: 4
Test ../../Test/X02zlevi.ztst failed: output differs from expected as shown above for:
  zletest $'    ----word----    ----word----    ----    \e38|daw30|daw22|daw14|daw6|daw'
Was testing: delete all word on other characters
../../Test/X02zlevi.ztst: test failed.
[…]
Running test:  menu select=long-list starts menu selection for list widgets
--- /tmp/zsh.ztst.19709/ztst.out	2022-04-19 05:05:20.416447321 +0000
+++ /tmp/zsh.ztst.19709/ztst.tout	2022-04-19 05:05:21.132273136 +0000
@@ -17,7 +17,6 @@
 NO:{p}
 NO:{q}
 NO:{r}
-NO:{s}
 NO:{t}
 NO:{u}
 NO:{v}
Test ../../Test/Y01completion.ztst failed: output differs from expected as shown above for:
  comptesteval '_tst() { local disp=( {a..z} ); compadd -ld disp $disp[@]; comppostfuncs=( _pst ) }'
  comptesteval '_pst() { local disp=( "<INSERT>$compstate[insert]</INSERT>" ); compadd -Qld disp $disp }'
  comptesteval "zstyle ':completion:*' menu select=long-list"
  # This test is sensitive to sorting differences across platforms
  comptesteval 'export LC_ALL=C'
  comptest $'tst \C-d'
  comptesteval "export LC_ALL=${(q)ZSH_TEST_LANG}"
Was testing:  menu select=long-list starts menu selection for list widgets
../../Test/Y01completion.ztst: test failed.
[…]
../../Test/Z03run-help.ztst: all tests successful.
**************************************
60 successful test scripts, 2 failures, 2 skipped
**************************************

		Kind regards, Axel
-- 
PGP: 2FF9CD59612616B5      /~\  Plain Text Ribbon Campaign, http://arc.pasp.de/
Mail: abe@deuxchevaux.org  \ /  Say No to HTML in E-Mail and Usenet
Mail+Jabber: abe@noone.org  X
https://axel.beckert.ch/   / \  I love long mails: https://email.is-not-s.ms/


^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-07  2:59 Speaking of 5.8.1.3-test Bart Schaefer
                   ` (3 preceding siblings ...)
  2022-05-09  6:52 ` Kamil Dudka
@ 2022-05-10  4:05 ` Jun T
  2022-05-10 16:41   ` Bart Schaefer
  4 siblings, 1 reply; 13+ messages in thread
From: Jun T @ 2022-05-10  4:05 UTC (permalink / raw)
  To: zsh-workers


> 2022/05/07 11:59, Bart Schaefer <schaefer@brasslantern.com> wrote:
> 
> Has anyone tried this at all?  No reports +/- so far.

If there is anyone familiar with NetBSD please correct me if I'm wrong.

If I run P01privileged as a root on NetBSD-9.2, the following four test
chunks fail:
  EUID set to RUID after disabling PRIVILEGED
  not possible to regain EUID when unprivileged after disabling PRIVILEGED
  not possible to regain EGID when unprivileged after disabling PRIVILEGED
  getpwuid() fails with non-existent RUID and 0 EUID

In all of these cases, error output includes:
zsh:unsetopt:1: PRIVILEGED: can't drop privileges; was able to restore the e[ug]id

"unsetopt privileged" calls
  setresuid(getuid(), getuid(), getuid())
at line 853 in options.c. NetBSD does not have setresuid(), and the wrapper
in openssh_bsd_setres_id.c is used, which calls
  setreuid(ruid, ruid)
But man setreuid(2) (on NetBSD) says:
  "If the real user ID is changed, the saved user ID is
    changed to the new value of the effective user ID."
In the above setreuid(ruid, ruid), the real uid ruid=getuid() does not change,
so the saved uid is not changed (and euid can be restored to it later).

The same manpage also says that setreuid() "is made obsolete by the saved ID
functionality in setuid(2) and seteuid(2)". And man setuid(2) says:
  "The setuid() function sets the real and effective user IDs and the saved
   set-user-ID of the current process to the specified value."
So just calling setuid(getuid()) is enough for dropping the privilege.

A simple workaround would be the following.
Or we can define BROKEN_SETRE{U,G}ID in configure.ac if $host_os is netbsd.

PS
P01privileged passes on FreeBSD-13, Dragonfly-6 and OpenBSD-7.


diff --git a/Src/openssh_bsd_setres_id.c b/Src/openssh_bsd_setres_id.c
index 217a6d074..26c7d3958 100644
--- a/Src/openssh_bsd_setres_id.c
+++ b/Src/openssh_bsd_setres_id.c
@@ -55,6 +55,16 @@
 #include <unistd.h>
 #include <string.h>
 
+#ifdef __NetBSD__
+/*
+ * On NetBSD, setreuid() does not reset the saved uid if the real uid
+ * is not modified. Better to use setuid() that resets all of real,
+ * effective and saved uids to the specified value. Same for setregid().
+ */
+#define BROKEN_SETREUID
+#define BROKEN_SETREGID
+#endif
+
 #if defined(ZSH_IMPLEMENT_SETRESGID) || defined(BROKEN_SETRESGID)
 int
 setresgid(gid_t rgid, gid_t egid, gid_t sgid)





^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-10  4:05 ` Jun T
@ 2022-05-10 16:41   ` Bart Schaefer
  0 siblings, 0 replies; 13+ messages in thread
From: Bart Schaefer @ 2022-05-10 16:41 UTC (permalink / raw)
  To: Jun T; +Cc: Zsh hackers list

On Mon, May 9, 2022 at 9:08 PM Jun T <takimoto-j@kba.biglobe.ne.jp> wrote:
>
> So just calling setuid(getuid()) is enough for dropping the privilege.
>
> A simple workaround would be the following.
> Or we can define BROKEN_SETRE{U,G}ID in configure.ac if $host_os is netbsd.

I'd vote for the simpler approach (your included diff) given proximity
to release.  Perhaps we can come up with a better configure.ac test
later.


^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-09  9:58     ` Axel Beckert
@ 2022-05-10 17:28       ` Axel Beckert
  2022-05-10 17:55         ` Bart Schaefer
  0 siblings, 1 reply; 13+ messages in thread
From: Axel Beckert @ 2022-05-10 17:28 UTC (permalink / raw)
  To: zsh-workers

Hi again,

On Mon, May 09, 2022 at 11:58:31AM +0200, Axel Beckert wrote:
> So here are the test suite failures of 5.8.1.2-test on the alpha
> architecture:

5.8.1.3-test failed on alpha again, but this time differently:

[…]
../../Test/W03jobparameters.ztst: all tests successful.
../../Test/X02zlevi.ztst: starting.
Running test: change last character in buffer
Test successful.
Running test: change forward to matching bracket
Test successful.
Running test: delete backwards to matching bracket
Test successful.
Running test: delete empty line
Test successful.
Running test: yank and paste blank line
Test successful.
Running test: swap case on a blank line
Test successful.
Running test: replace character when at end of buffer or newline
Test successful.
Running test: replace acts like insert at newline
Test successful.
Running test: join lines with line count
Test successful.
Running test: join more lines than possible
Test successful.
Running test: don't indent blank lines
Test successful.
Running test: additional indentation
Test successful.
Running test: indent with one character on final line
Test successful.
Running test: yank left moves the cursor
Test successful.
Running test: yank right leaves the cursor
Test successful.
Running test: yank up line moves cursor up but not to buffer start
Test successful.
Running test: yank down line leaves the cursor
Test successful.
Running test: yank up clears lastcol
Test successful.
Running test: yank up honours lastcol
Test successful.
Running test: delete up clears lastcol
Test successful.
Running test: repeat replace chars at the end of the line consumes the replace char
Test successful.
Running test: numeric argument to repeat replaces change count
Test successful.
Running test: numeric args to both action and movement are multiplied (and saved for any repeat)
Test successful.
Running test: paste register 1 to get last deletion
Test successful.
Running test: paste register 0 to get last yank
E: Build killed with signal TERM after 150 minutes of inactivity

So it seems to have hung at "Running test: paste register 0 to get
last yank" for some reason.

Full log at https://buildd.debian.org/status/fetch.php?pkg=zsh&arch=alpha&ver=5.8.1.3-test-1&stamp=1652169697&raw=0

		Kind regards, Axel
-- 
PGP: 2FF9CD59612616B5      /~\  Plain Text Ribbon Campaign, http://arc.pasp.de/
Mail: abe@deuxchevaux.org  \ /  Say No to HTML in E-Mail and Usenet
Mail+Jabber: abe@noone.org  X
https://axel.beckert.ch/   / \  I love long mails: https://email.is-not-s.ms/


^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-10 17:28       ` Axel Beckert
@ 2022-05-10 17:55         ` Bart Schaefer
  2022-05-11 10:06           ` Axel Beckert
  0 siblings, 1 reply; 13+ messages in thread
From: Bart Schaefer @ 2022-05-10 17:55 UTC (permalink / raw)
  To: Zsh hackers list

On Tue, May 10, 2022 at 10:39 AM Axel Beckert <abe@deuxchevaux.org> wrote:
>
> 5.8.1.3-test failed on alpha again, but this time differently:

Looks like a zpty problem.  I'm not sure how much effort we want to
put into that, because it's likely to be a problem only within the
test suite.

Axel, you may be the only person with sufficient access to alpha to do
any deeper debugging.


^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: Speaking of 5.8.1.3-test ...
  2022-05-10 17:55         ` Bart Schaefer
@ 2022-05-11 10:06           ` Axel Beckert
  0 siblings, 0 replies; 13+ messages in thread
From: Axel Beckert @ 2022-05-11 10:06 UTC (permalink / raw)
  To: zsh-workers

Hi,

On Tue, May 10, 2022 at 10:55:40AM -0700, Bart Schaefer wrote:
> On Tue, May 10, 2022 at 10:39 AM Axel Beckert <abe@deuxchevaux.org> wrote:
> >
> > 5.8.1.3-test failed on alpha again, but this time differently:
> 
> Looks like a zpty problem.  I'm not sure how much effort we want to
> put into that, because it's likely to be a problem only within the
> test suite.

And it only happens on rather exotic/vintage architecture which had
other arch-specific issues beforehand.

So yes, I fully agree. And I would even say that this can be ignored
for the upcoming release.

But even though I don't consider these issues being relevant for the
release, I also didn't want to sweep them under the carpet, so I
mentioned them.

> Axel, you may be the only person with sufficient access to alpha to do
> any deeper debugging.

Probably yes. :-)

[The remainder of this mail is just reminiscing about vintage hardware.]

And likely even more than you might expect: On the one hand, I have
access to Debian's porter boxes. (Can't share that access, of couse.)

But on the other hand I also do have an HP AlphaServer DS15(*) sitting
on my desk and catching dust. Haven't powered it up for a few years,
so I might need to apply quite some updates first. But I expect it to
run as it did when I had it with me on a vintage computing festival a
few years ago showing of old, exotic workstations running a current
Debian Unstable.

Anyway, that's kinda motivation to get that box dedusted and running
again. No promises on a time frame or if it will work at all, though.
:-)

Depending on how well-known a person is to me, I might even give out
(IPv6-only) SSH accounts on that box — once I've got it up and running
again. PGP-signed mails and a PGP-key, to which I can find a trust
chain, are necessary for that, though.

Footnotes:

(*) The DS15 has been built between 2003 and 2007 according to
    https://en.wikipedia.org/wiki/AlphaServer and seems to have been
    one of the models produced until HP phased out the Alpha
    architecture in favour of the also already dead Itanium/IA64
    architecture. It's actually quite performant for being "vintage
    hardware" due to its 1 GHz (single core) CPU and 4 GB of RAM. :-)

		Kind regards, Axel
-- 
PGP: 2FF9CD59612616B5      /~\  Plain Text Ribbon Campaign, http://arc.pasp.de/
Mail: abe@deuxchevaux.org  \ /  Say No to HTML in E-Mail and Usenet
Mail+Jabber: abe@noone.org  X
https://axel.beckert.ch/   / \  I love long mails: https://email.is-not-s.ms/


^ permalink raw reply	[flat|nested] 13+ messages in thread

end of thread, other threads:[~2022-05-11 10:07 UTC | newest]

Thread overview: 13+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-05-07  2:59 Speaking of 5.8.1.3-test Bart Schaefer
2022-05-07  3:50 ` dana
2022-05-07  4:06 ` Daniel Shahaf
2022-05-08 14:26   ` Axel Beckert
2022-05-09  7:07     ` Luna Jernberg
2022-05-09  9:58     ` Axel Beckert
2022-05-10 17:28       ` Axel Beckert
2022-05-10 17:55         ` Bart Schaefer
2022-05-11 10:06           ` Axel Beckert
2022-05-08  4:10 ` Matthew Martin
2022-05-09  6:52 ` Kamil Dudka
2022-05-10  4:05 ` Jun T
2022-05-10 16:41   ` Bart Schaefer

Code repositories for project(s) associated with this public inbox

	https://git.vuxu.org/mirror/zsh/

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).