ethernet module has a control flow bug

On Tuesday 19 Oct 2010 01:56:38 Faustus wrote:
> First,
>
> /usr/share/laptop-mode-tools/modules/ethernet: if ! [ -z
> $path ]; then
> /usr/share/laptop-mode-tools/modules/ethernet: if ! [ -z
> $dev_path ] && [ -f $dev_path/enable ]; then
>
> are confusing without the quotes around the variable given to -z (but it
> still works in Bash).
>

So does it not work with dash ?
laptop-mode-tools uses dash and I haven't seen this bug report. In case you
see problems when run with dash, please report.

> Second, in the ETHERNET devices loop, if $path is empty, previous value
> for $dev_path will be used (redundantly operating on the device).

Thank you. I have fixed this in the upstream branch. Will be part of the next
(1.56) release.

Ritesh

--
Ritesh Raj Sarraf
RESEARCHUT - http://www.researchut.com
"Necessity is the mother of invention."

> So does it not work with dash? laptop-mode-tools uses dash and I haven't seen this bug report. In case you
see problems when run with dash, please report.

What is dash? On Gentoo, /usr/sbin/laptop_mode uses /bin/sh, which is a link to /bin/bash. It works, and I don't see a reason to use what looks like an obscure bug/feature for no reason.

This bug was fixed in the package laptop-mode-tools - 1.57-1ubuntu1

---------------
laptop-mode-tools (1.57-1ubuntu1) natty; urgency=low

  * Merge with Debian unstable. Remaining Ubuntu changes:
    - debian/rules: Do not ship acpi/apm scripts (we handle that ourselves)
  * Drop previous Ubuntu changes for /etc/default/laptop-mode. We don't
    install this package by default any more for a long time, so if people
    install it manually they presumably want to install it. 1.53 added its own
    global enable/disable switch, so this is obsolete, too.
  * This version removes the file conflict with pm-utils. (LP: #638307)

laptop-mode-tools (1.57-1) unstable; urgency=low

  * New Upstream Release
    + Pass calling application's arguments to the main function. Thanks to
      Radek for the report and the patch
    + Use governor policy from the config file. Thanks to Michael Orlov for
      the report.
    + Enable ENABLE_AUTO_MODULES by default. When a user installs the package,
      the first immediate expectation is to see some changes
    + USB suspend interface has changed with kernels 2.6.35 and above.
      (Closes: #589000)
    + Add support for Linux Runtime Power Management (Closes: #592661)
    + On many other distributions, wireless-tools packages is still old and
      does not return proper exit status on failures. Thanks to Luca Landolfi
      for the patch.
    + Add support to blacklist USB Devices by types. Thanks to Simon Que for
      the patch
    + Re-initialize dev_path before the next iteration. Thanks to Faustus for
      the bug report and the patch (LP: #662924)
    + ENABLE_LAPTOP_MODE_TOOLS was ignore because of wrong check. Thanks to
      Matus Harvan for reporting this. (Closes: #602278)
    + Add support to invoke laptop-mode-tools with hotplug events. This
      currently only has support for the usb-autosuspend module.
      Thanks to Simon Que for the patch
    + Add locking support using flock. (Closes: #566613)
    + Add 99-laptop-mode.rules to trap multiple kernel events
  * Drop pm-utils from recommends since it conflicts with us (Closes: #614024)
  * Bump Standards Version to 3.9.1 (No changes required)

laptop-mode-tools (1.55-1) unstable; urgency=low

  * Change address to my official Debian address
  * New Upstream Release
    + On Intel HD Audio, revert power saving changes when switching back to AC.
      Thanks to Christoph Langguth for the report and the patch
    + Install the pm-utils hook as 01laptop-mode to ensure proper execution in
      reverse ordering
    + Add wireless-power module, for generic non-Intel wireless interfaces that
      support the iwconfig "power" option.
    + Fixed upgrade path for configurations which didn't have the LM/NOLM
      options in the Intel SATA Power Management module config.
    + Add flush-* to ignored programs for lm-profiler, because it is the
      successor to pdflush.
    + Fix WoL regression introduced by DISABLE_ETHERNET_ON_BATTERY. Thanks
      to Matthias Dienstbier for the report and the patch
    + Update comments in wireless-iwl-power.conf. (Debian BTS: #580730)
    + Check for old hook in pm-utils and remove it
    + Update manpage for al...