Astrometry eats bscale-card in fits-headers

[Matthias Heinzelmann]

Hello,

I gave my local astrometry.net installation (0.46) some fits-files and noticed, that the card containing the bscale-keyword got eaten:

SIMPLE  =                    T                                                 
BITPIX  =                   16                                                 
NAXIS   =                    2                                                 
NAXIS1  =                 4096                                                 
NAXIS2  =                 4096                                                 
CCD-TEMP=               -22.08                                                 
EXPTIME =                120.0                                                 
EXPOSURE=                120.0                                                 
DATE    = '2014-11-08T07:12:06'                                                
BZERO   =              32768.0    after here the original bscale is inserted                                             
COMMENT  reduced file                                                                   
COMMENT --Start of Astrometry.net WCS solution--                               
COMMENT --Put in by the new-wcs program--                                      
COMMENT                                                                        
WCSAXES =                    2 / no comment                                    
CTYPE1  = 'RA---TAN-SIP' / TAN (gnomic) projection + SIP distortions           
CTYPE2  = 'DEC--TAN-SIP' / TAN (gnomic) projection + SIP distortions           
EQUINOX =               2000.0 / Equatorial coordinates definition (yr)        
LONPOLE =                180.0 / no comment                                    
LATPOLE =                  0.0 / no comment                                    
CRVAL1  =        49.6344158735 / RA  of reference point                        
CRVAL2  =         41.361928016 / DEC of reference point                        
CRPIX1  =        2555.42550659 / X reference pixel                             
CRPIX2  =        2677.61062622 / Y reference pixel                             
CUNIT1  = 'deg     ' / X pixel scale units                                     
CUNIT2  = 'deg     ' / Y pixel scale units                                   
[snip]

Original Header:
SIMPLE  =                    T                                                 
BITPIX  =                   16                                                 
NAXIS   =                    2                                                 
NAXIS1  =                 4096                                                 
NAXIS2  =                 4096                                                 
BSCALE  =               1.0000                                                 
BZERO   =           32768.0000                                                 
CCD-TEMP=         -22.08000000                                                 
EXPTIME =         120.00000000                                                 
EXPOSURE=         120.00000000                                                 
DATE    = '2014-11-08T07:12:06'                                                
COMMENT  reduced 
END    


Is this normal behaviour because the default-value of bscale is 1.0 and in my files it is too, or a bug?

Best Regards,
Matt