Re: Add a new exec_exec_file_name RPC

[Justus Winter]

Hi,

Quoting Samuel Thibault (2014-04-08 02:18:13)
> Well, I'm not the one to be convinced: it's only Roland which can ack
> the glibc part, and thus the whole idea of the RPC addition.  One "just"
> needs to explain him why we really need it.

I thought that this patch series was necessary for fakeroot-hurd.  In
order to demonstrate the problem I dropped the patch series from my
hurd-ci packages, but the problem did not reappear:

teythoon@hurdbox ~ % cat test.bash
#!/bin/bash
echo $0
teythoon@hurdbox ~ % fakeroot-tcp ./test.bash
./test.bash
teythoon@hurdbox ~ % fakeroot-hurd ./test.bash
./test.bash

It used to say /dev/fd/3 when run under fakeroot-hurd when I started
looking into the issue.  I now think that the original patch series
introduced this problem for programs run under fakeroot-hurd.  The one
presented here has been amended with the fix I posted some time ago
that went into the debian package as exec_filename_fix.patch.

Emilios patch series addresses [0].  I can confirm that this problem
as described in the bug report still exists.

0: https://savannah.gnu.org/bugs/?28934

Stock Debian/Hurd packages with the patch series:

root@debian:~# ./foo ./bar
$0: ./bar
root@debian:~# ./foo bar
$0: bar

Hurd packages from hurd-ci without the patch series:

teythoon@hurdbox ~ % ./foo ./bar
$0: ./bar
teythoon@hurdbox ~ % ./foo bar
$0: /dev/fd/3

I have rebased and fixed the patch series so that it doesn't disturb
fakeroot-hurd.  The problem is clearly described in the bug report.  I
do not know what else to do.  I have dropped the patch series from the
hurd-ci builds so that it will not disturb my workflow any longer.

Justus