[Release] Libinsane 1.0.2

[Jerome Flesch]

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)`.
https://openpaper.work/scannerdb/report/280/
https://openpaper.work/scannerdb/report/350/
- 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.
https://openpaper.work/fr/scanner_db/report/335/
- 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.
https://openpaper.work/fr/scanner_db/report/392/
- 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.
https://openpaper.work/fr/scanner_db/report/328/
- 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.
https://openpaper.work/fr/scanner_db/report/298/
- Sane + Canon Lide 220:
* https://openpaper.work/fr/scanner_db/report/279/
* 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
https://gitlab.gnome.org/World/OpenpaperWork/libinsane
Detailed ChangeLog is available here:
https://gitlab.gnome.org/World/OpenpaperWork/libinsane/blob/master/ChangeLog



Best regards,