dmidecode changed default output format and added new command line interface for scripts - please update laptop-detect

[Mantas Kriaučiūnas]

Hi,

laptop-detect script currently tries to parse dmidecode's version 2.6 output instead of using new command line interface, which appeared in dmidecode ver. 2.7:

dmidecode (2.7-1) unstable; urgency=low

  * New upstream release.
    - Include new command line interface making it easier for scripts to extract the values they want.
    - Default output format is slightly changed. This might break existing scripts, which should use the new command line options instead.

New command line interface are new parameters:

 --string KEYWORD Only display the value of the given DMI string
 --type TYPE Only display the entries of given type

Valid string keywords are:
  bios-vendor
  bios-version
  bios-release-date
  system-manufacturer
  system-product-name
  system-version
  system-serial-number
  baseboard-manufacturer
  baseboard-product-name
  baseboard-version
  baseboard-serial-number
  baseboard-asset-tag
  chassis-manufacturer
  chassis-version
  chassis-serial-number
  chassis-asset-tag
  processor-manufacturer
  processor-version

Valid type keywords are:
  bios
  system
  baseboard
  chassis
  processor
  memory
  cache
  connector
  slot

So, please change parsing dmidecode|grep Chassis -A 10|grep -m1 Type|sed -e 's/.*Type: \(.*\)/\1/' to new code, which uses new command-line parameters or at least check if changed default dmidecode output format doesn't couse any problems with detecting a laptop.

[Matt Zimmerman (mdz)]

It's too late in Dapper to migrate to the new interface, but the existing code should be updated to cope with changes in the human-readable interface

[Daniel Silverstone (dsilvers)]

The new format doesn't provide a tag which outputs the chassis type anyway.

The human-readable format has not changed sufficiently to cause any issues with laptop-detect.sh

[Matt Zimmerman (mdz)]

Mantas, can you confirm that there is actually no problem parsing the current output, and that this is just a proposal to use the alternative interface?

[Daniel Silverstone (dsilvers)]

I'm confirming, it's all okay (verified manually using grep etc on my laptop)

Since dmidecode doesn't offer the desired data in a tagged form, this bug is therefore invalid. rejecting.