iraf.net

[b:aa520eeeb1]From: [/b:aa520eeeb1]"Dmitry Bizyaev" <dmbiz@noao.edu ([email]dmbiz@noao.edu[/email])>
[b:aa520eeeb1]Date: [/b:aa520eeeb1]July 28, 2005 4:19:00 PM MST
[b:aa520eeeb1]To: [/b:aa520eeeb1]iraf@noao.edu ([email]iraf@noao.edu[/email])
[b:aa520eeeb1]Subject: [/b:aa520eeeb1][b:aa520eeeb1]weird thing in IRAF's WCS[/b:aa520eeeb1]
Dear Sirs,
there's an at least inconvenience found in headers produced by IRAF V2.12.2.
Everything is OK when I use one reference spectra for the wavelength calibration of echelle frames. If I use 2, there is always no space between the descriptions of spectral solution for the same order from different calibration files. An example:
WAT0_001= 'system=multispec'
WAT1_001= 'wtype=multispec label=Pixel units=angstrems'
WAT2_001= 'wtype=multispec spec1 = "1 153 2 5062.8166030989 -0.046661351312468 '
WAT2_002= '1200 0. 15.86 21.06 0.5 0. 1 7 1. 1200. 5035.12802691391 -27.9630193'
WAT2_003= '538527 -0.285805572201651 -0.0113631740304677 -0.00178860718736775 5'
WAT2_004= '.13511469672007E-4 -5.74987772021264E-40.5 0. 1 7 1. 1200. 5035.1327'
WAT2_005= '3663838 -27.9623027639898 -0.28473703778715 -0.0108575670041065 -0.0'
WAT2_006= '0140573815535319 6.91237577741873E-5 -7.80622795856407E-4" spec2 = "'
WAT2_007= '2 152 2 5096.1364373927 -0.046962093668028 1200 0. 33.98 39.18 0.5 0 .....
DCLOG1  = 'REFSPEC1 = cfs0002.ec 0.5'
DCLOG2  = 'REFSPEC2 = cfs0010.ec 0.5'
--- end of example ---
Look at WAT2_004 string: instead of "-5.74987772021264E-4 0.5" it shows -5.74987772021264E-40.5.  And it happens in all orders!
I believe IRAF handles it somehow, since it works with the files properly. But it is very inconvenient to handle within any external code. Is it available to fix? Is there any way to post-fix files already produced with this feature?
Best regards,
Dmitry

[b:aa520eeeb1]From: [/b:aa520eeeb1]Frank Valdes <valdes@noao.edu ([email]valdes@noao.edu[/email])>
[b:aa520eeeb1]Date: [/b:aa520eeeb1]August 15, 2005 9:46:42 AM MST
[b:aa520eeeb1]To: [/b:aa520eeeb1]Dmitry Bizyaev <dmbiz@noao.edu ([email]dmbiz@noao.edu[/email])>
[b:aa520eeeb1]Subject: [/b:aa520eeeb1][b:aa520eeeb1]Re: weird thing in IRAF's WCS[/b:aa520eeeb1]
Hello Dmitry,
I am not aware of this bug but it definitely is a bug.  Is this the latest version of IRAF?  Could you send me the first two lines of the login banner and tell me the type of host machine you are using.  I don't know if when we can track this down and fix it for a new distribution.  But at a minimum we can document the bug for others.  The fact that I am not aware of it means no one else has noticed a problem but this would be a little hard to see.  How did you find it? In other words, does it manifest itself as clearly bad coordinates?
To postfix this my only suggestion is to use HFIX individually on each file.  If you can find a pattern and some kind of "sed" type of command then HFIX could be automated but this would be very tricky.  When you use HFIX you are directly manipulating the headers and so you must take care not to get the FITS format wrong.  The WAT system of storing information is particularly tricky since keywords
are written to the end of the line.  Rather than trying to reformat all the lines I suggest you add the space and then delete the least significant digit of a number on the same line.  As you see the formating puts out a lot of digits which are not
really significant.  So in your example
"-5.74987772021264E-40.5" ->" -5.7498777202126E-4 0.5"
I'm sorry for the problem.
Yours,
Frank Valdes
[b:aa520eeeb1]From: [/b:aa520eeeb1]"Dmitry Bizyaev" <dmbiz@noao.edu ([email]dmbiz@noao.edu[/email])>
[b:aa520eeeb1]Date: [/b:aa520eeeb1]August 24, 2005 5:03:53 PM MST
[b:aa520eeeb1]To: [/b:aa520eeeb1]Frank Valdes <valdes@noao.edu ([email]valdes@noao.edu[/email])>
[b:aa520eeeb1]Subject: [/b:aa520eeeb1][b:aa520eeeb1]Re: weird thing in IRAF's WCS[/b:aa520eeeb1]
Hi Frank,
thank you for your answer. These are first two lines:
NOAO PC-IRAF Revision 2.12.2a-EXPORT Wed Jul 14 20:45:34 MST 2004
This is the EXPORT version of PC-IRAF V2.12 supporting most PC systems.
I run it under Linux Slackware
Linux sim 2.6.11.6 #1 Sun May 15 18:32:34 MST 2005 i686 unknown unknown GNU/Linux
and I used some older versions of kernel before (like 2.4.20 and older).
There are no error messages or wrong pictures within IRAF when I calculate or plot this such a spectrum. Probably the IRAF itself handles it very well. But if I try to deal with this header from any other external program - a bug emerges (as it should actually be).
I could transfer you corresponding files to finalize the wavelength calibration at your machine an let the problem reveal itself (or not).
Best wishes,
Dmitry[b:aa520eeeb1]From: [/b:aa520eeeb1]Frank Valdes <valdes@noao.edu ([email]valdes@noao.edu[/email])>
[b:aa520eeeb1]Date: [/b:aa520eeeb1]August 26, 2005 12:17:58 PM MST
[b:aa520eeeb1]To: [/b:aa520eeeb1]Dmitry Bizyaev <dmbiz@noao.edu ([email]dmbiz@noao.edu[/email])>
[b:aa520eeeb1]Subject: [/b:aa520eeeb1][b:aa520eeeb1]Re: weird thing in IRAF's WCS[/b:aa520eeeb1]
Thank you Dmitry,
We can try and see if we can reproduce this.  If you would like to send me a file and the command that introduces the WAT keywords (I assume this is DISPCOR?) we can see if we can find the bug.  You can put a file to our ftp site at iraf.noao.edu in the pub directory or you can tell me how to retrieve it from your site.  If the file is not too large you could also attach it to may mail address.
Yours,
Frank