From: David Dyck <david.dyck@gmail.com>
To: Ken Brown <kbrown@cornell.edu>
Cc: cygwin@cygwin.com
Subject: Re: ag 2 <(echo 2) gets assertion "p >= path" failed: .. /cygwin-3.1.7 ... /cygwin/path.cc", line 3065, function: int symlink_info::check
Date: Tue, 8 Sep 2020 20:48:28 -0700 [thread overview]
Message-ID: <CAJtSMgvAyB6PgJFsjzUX9fmYX7Tqm17OJgUOy+kFb5TSSt05rQ@mail.gmail.com> (raw)
In-Reply-To: <76a478c7-fabe-b8e9-581e-9ffcc08f53ed@cornell.edu>
Thank you Ken, Corinna, and list members
On Tue, Sep 8, 2020 at 12:32 PM Ken Brown <kbrown@cornell.edu> wrote:
> On 9/8/2020 3:26 PM, Ken Brown via Cygwin wrote:
> > On 9/7/2020 4:35 PM, Ken Brown via Cygwin wrote:
> >> On 9/6/2020 4:28 PM, Ken Brown via Cygwin wrote:
> >>> On 9/6/2020 3:47 PM, Ken Brown via Cygwin wrote:
> >>>> On 9/6/2020 2:43 PM, David Dyck via Cygwin wrote:
> >>>>> This command triggers an assertion failure
> >>>>> "ag" is from the_silver_searcher
> >>>>>
> >>>>> $ ag 2 <(echo 2)
> >>>>> assertion "p >= path" failed: file
> >>>>>
> "/home/corinna/src/cygwin/cygwin-3.1.7/cygwin-3.1.7-1.x86_64/src/newlib-cygwin/winsup/cygwin/path.cc",
>
> >>>>>
> >>>>> line 3065, function: int symlink_info::check(char*, const
> >>>>> suffix_info*, fs_info&, path_conv_handle&)
> >>>>> Aborted (core dumped)
> >>>>>
> >>>>> 3473k 2020/08/22 C:\cygwin64\bin\cygwin1.dll
> >>>>> Cygwin DLL version info:
> >>>>> DLL version: 3.1.7
> >>>>> bash
> 4.4.12-3 OK
> >>>>> the_silver_searcher
> 2.2.0-1 OK
> >>>> [...]
> >>>>> assertion "p >= path" failed: file
> >>>>>
> "/home/corinna/src/cygwin/cygwin-3.1.7/cygwin-3.1.7-1.x86_64/src/newlib-cygwin/winsup/cygwin/path.cc",
>
> >>>>>
> >>>>> line 3065, function: int symlink_info::check(char*, const
> >>>>> suffix_info*, fs_info&, path_conv_handle&)
> >>>>> Aborted (core dumped)
> >>>> [...]
> >>>>> I've reported this on github as an "ag" bug, but I think it is a bug
> in cygwin
> >>>>
> >>>> An assertion failure in Cygwin code is a Cygwin bug. I'll take a
> look.
> >>>
> >>> Running
> >>>
> >>> bash -c '/usr/bin/ag 2 <(echo 2)'
> >>>
> >>> under strace yields the following:
> >>>
> >>> 242 191767 [main] ag 33659 open: open(/dev/fd/63/.ignore, 0x0)
> >>> [...]
> >>> 30 192584 [main] ag 33659 mount_info::conv_to_win32_path:
> src_path
> >>> /proc/self/fd/63/.ignore, dst /proc/self/fd/63/.ignore, flags 0x0, rc 0
> >>> [...]
> >>> 31 193366 [main] ag 33659 mount_info::conv_to_win32_path:
> >>> conv_to_win32_path (pipe:[4295036184]/.ignore)
> >>> [...]
> >>> 31 193550 [main] ag 33659 mount_info::conv_to_win32_path:
> >>> conv_to_win32_path (pipe:[4295036184])
> >>> [...]
> >>> 34 193615 [main] ag 33659 symlink_info::check: 0xC0000034 =
> NtCreateFile
> >>> (\??\C:pipe:[4295036184])
> >>>
> >>> The assertion fails because the path 'C:pipe:[4295036184]' doesn't
> contain a
> >>> backslash. But probably we should never have allowed ourselves to get
> to the
> >>> point of considering that path.
> >>
> >> I've made some progress but haven't figured out the fix yet. First,
> for
> >> easier debugging, here's a simpler way to reproduce the problem:
> >>
> >> $ cat proc_bug.c
> >> #include <unistd.h>
> >> #include <stdio.h>
> >> #include <stdlib.h>
> >> #include <fcntl.h>
> >>
> >> int
> >> main ()
> >> {
> >> int fd[2];
> >> char fname[100];
> >>
> >> if (pipe (fd) < 0)
> >> {
> >> perror ("pipe");
> >> exit (1);
> >> }
> >> sprintf (fname, "/dev/fd/%d/foo", fd[0]);
> >> if (open (fname, O_RDONLY) < 0)
> >> {
> >> perror ("open");
> >> exit (1);
> >> }
> >> }
> >>
> >> $ gcc -o proc_bug proc_bug.c
> >>
> >> $ ./proc_bug.exe
> >> assertion "p >= path" failed: file
> >> "../../../../newlib-cygwin/winsup/cygwin/path.cc", line 3065, function:
> int
> >> symlink_info::check(char*, const suffix_info*, fs_info&,
> path_conv_handle&)
> >> Aborted (core dumped)
> >>
> >> Here's what happens. The program is trying to open /dev/fd/3/foo,
> where file
> >> descriptor 3 is the read end of a pipe. path_conv check resolves this
> to
> >> /proc/<pid>/fd/3/foo, creates an fhandler_process for this path, and
> calls (at
> >> path.cc:782) fhandler_process::exists. The latter fills the filebuf
> with
> >> "pipe:[xxxxxxxxxx]/foo" and returns virt_fsdir. We're now at
> path.cc:808, and
> >> everything is set up for the assertion failure.
> >
> > This is now fixed. David, you can test it as soon as Corinna has a
> chance to
> > make a new Cygwin snapshot.
>
> That's now done:
>
> https://cygwin.com/snapshots/
>
> Ken
>
Thank you!
My first time through
https://cygwin.com/snapshots/
and
https://cygwin.com/faq.html#faq.setup.snapshots
Installed the dll alone (
https://cygwin.com/snapshots/x86_64/cygwin1-20200908.dll.xz )
and the tests I performed were successful
$ ack 2 <(echo 2)
2
prev parent reply other threads:[~2020-09-09 3:48 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-06 18:43 David Dyck
2020-09-06 19:07 ` Eliot Moss
2020-09-06 22:15 ` David Dyck
2020-09-06 23:19 ` Eliot Moss
2020-09-06 19:47 ` Ken Brown
2020-09-06 20:28 ` Ken Brown
2020-09-07 20:35 ` Ken Brown
2020-09-08 19:26 ` Ken Brown
2020-09-08 19:32 ` Ken Brown
2020-09-09 3:48 ` David Dyck [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAJtSMgvAyB6PgJFsjzUX9fmYX7Tqm17OJgUOy+kFb5TSSt05rQ@mail.gmail.com \
--to=david.dyck@gmail.com \
--cc=cygwin@cygwin.com \
--cc=kbrown@cornell.edu \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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 read-only IMAP folder(s) and NNTP newsgroup(s).