Libinsane 1.0.2

Announcements
1

Mail posted on the old mailing-list the 2019/10/14:

Hello,

I’ve released Libinsane 1.0.2.
It seems I also forgot to send the email for Libinsane 1.0.1, sorry :/.

1.0.1 and 1.0.2 are bugfix-only releases.

 Main changes are:

2019/08/18 - 1.0.1:

  • Sane: make sure sane_cancel() is called at the end of all scans
    (required by some drivers)
  • Sane: Fix memory leak when opening an invalid device. (thanks to
    Elliott Sales de Andrade)
  • WIA/Twain: Fix reading bits-per-pixel from BMP header on big-endian
    CPU (thanks to Elliott Sales de Andrade)

19/10/14 - 1.0.2:

  • Libinsane/gobject:
    Added mising flag Libinsane.CAPABILITY_INACTIVE.

  • Sane:
    when the application expect one page, do not request a second one (fix
    unittests timeout).

  • WIA2:
    Use the option ‘dev_name’ to figure out the manufacturer name and the
    device name. It is the only reliable option.

  • WIA2:
    HP Photosmart C4200, Epson Perfection V30/V300:
    When getting the scanned image, allow calls to seek(SEEK_SET, current_position).
    Scan report 280 - EPSON Perfection V30/V300 - OpenPaperwork
    Scan report 350 - HP Photosmart C4200 - OpenPaperwork

  • WIA2:
    Ignore sources that are not explicitely a flatbed or an ADF (for
    instance 0000\Root\Auto: It is simply not supported by Libinsane).

  • TWAIN: Fujitsu fi-6130 / fi-6130dj:
    Keep reading Windows message queue while waiting for the DSM callback
    to be called with the value ‘ready’. If we don’t, the DSM callback never
    gets called.

  • normalizers/opt_aliases:
    Fix: When using option aliases ‘tl-x’, ‘tl-y’, ‘br-x’ and ‘br-y’, with
    WIA2, if the lowest values for ‘xpos’, ‘ypos’, ‘xextent’ and ‘yextent’
    don’t have (0, 0) as minimum values, inaccurate
    values were set in WIA2 frame options.
    Rapport de numérisation 335 - Brother MFC-J5620DW - OpenPaperwork

  • workarounds/cache:
    Fix: do not drop the item type (root, adf, flatbed, etc).

  • normalizers/resolution:
    If the driver does not specify any constraint for the option
    ‘resolution’, put one default constraint.

  • normalizers/safe_defaults: Sane + Fujitsu fi-6130dj:
    Cannot set the option ‘page-height’ to the maximum: Scanner rejects
    it. Option must be set to slightly less than the maximum.
    Rapport de numérisation 392 - FUJITSU fi-6130dj - OpenPaperwork

  • normalizers/safe_default: Sane + Epson Perfection 1250:
    If the default value for an option is out of the range specified by
    the option constraint, do not change it.
    Rapport de numérisation 328 - Epson Perfection 1250/Photo - OpenPaperwork

  • workaround/opt_names: Sane + Canon Pixma MX520:
    Take into account that somes backend provide both an option
    ‘resolution’ and ‘scan-resolution’. In that case, do not generate an
    option ‘resolution’ that would be an alias for ‘scan-resolution’, just
    keep ‘resolution’ as is.

  • normalizers/source_names: Sane + Canon CanoScan Lide 30:
    Normalize the weird source names specified by the plustek driver.
    Rapport de numérisation 298 - Canon CanoScan N1240U/LiDE30 - OpenPaperwork

  • Sane + Canon Lide 220:

    • Rapport de numérisation 279 - Canon LiDE 220 - OpenPaperwork
    • workarounds/check_capabilty: Some options are marked as non-readable
      (INACTIVE) but may still be writable (for instance the option ‘source’).
    • normalizers/source_nodes: If the option ‘source’ is marked inactive,
      still try to set it, but do not return an error if setting it fails.

As usual, information regarding Libinsane are available at

Detailed ChangeLog is available here:

Best regards,