| Re: Catching I/O errors on RESTORE [message #29671] |
Mon, 11 March 2002 21:03  |
alt
Messages: 28
Registered: August 2001
|
Junior Member |
|
|
>
> I would very much like to hear reports on how well the CMSVLIB SAVE
> library does with corrupted files. It should report errors back to
> user programs without crashing.
>
It really does. Thanks. At least as far as I had opportunity to check.
I had problems with ordinary RESTORE and none of them is present now.
I am going to get huge statistics on CMSVLIB stability during my
software package running (save-restore every 10 sec through network)
this summer (forest fire season). Certainly I will inform you about
results.
Craig, it would be impudence to wish more but if your library had an
ability to replace (expand, squeeze, delete) variables in sav-files it
would be absolutely indispensable for organizing of IDL file container
and would be used very widely as I think. Just a wish, sorry :-))
Best regards,
Altyntsev Dmitriy
Remote Sensing Center, ISTP
Irkutsk, Russia
http://ckm.iszf.irk.ru
|
|
|
|
|
|
| Re: Catching I/O errors on RESTORE [message #29688 is a reply to message #29684] |
Sun, 10 March 2002 23:40  |
R.Bauer
Messages: 1424
Registered: November 1998
|
Senior Member |
|
|
"M. Katz" wrote:
>
> Here's in I/O error-trapping question for the IDL gurus out there...
>
> Working in IDL 5.4 on SunOS UNIX, I'd like to catch errors
> that occur when reading a bad or incomplete--but existing--
> IDL SAVE file. I want to use ON_IOERROR, but it doesn't
> seem to catch errors that occur with the RESTORE function.
> Here's more information...
>
> I create a smallish (100-200 bytes) disk file with
>
> SAVE, x, y, z, filename="myfile.dat"
>
> Since I have lots of programs running simultaneously, different
> parts of my program or different IDL sessions may be
> SAVEing and RESTOREing these files in quick succession.
>
> Before I RESTORE, I always check for the existence of the file with
> FRELUN, 10
> OPENR, 10, filename, ERROR=err
>
> Sometimes, the file exists but is empty (!), then IDL tries to RESTORE
> and comes back with an END OF FILE error.
>
> SO, I put in an error catcher:
>
> ON_IOERROR, error_label
>
> Yet the program still halts with an END OF FILE, and the error
> trap is never activated. Any ideas?
> I would like the RESTORE file to trip the error trap and go
> nicely to my error-handling code.
>
> Thanks, MKatz
Dear MKatz,
I have had a similiar problem.
RESTORE,filename=file
PRINT,!error_state.name
IF !error_state.name NE 'IDL_M_SUCCESS' THEN MESSAGE,'FILE BAD',/INFO
regards
Reimar
--
Reimar Bauer
Institut fuer Stratosphaerische Chemie (ICG-I)
Forschungszentrum Juelich
email: R.Bauer@fz-juelich.de
------------------------------------------------------------ -------
a IDL library at ForschungsZentrum Juelich
http://www.fz-juelich.de/icg/icg1/idl_icglib/idl_lib_intro.h tml
============================================================ =======
|
|
|
|
| Re: Catching I/O errors on RESTORE [message #29698 is a reply to message #29688] |
Sat, 09 March 2002 23:40 |
alt
Messages: 28
Registered: August 2001
|
Junior Member |
|
|
I have the similar problem while organizing simple IDL client-server
applications. Suppose client is trying to read sav-file before server
end up with writing...
Some possible useful ideas:
* CATCH (not ON_IOERROR) catches all errors including RESTORE errors
(unfortunately not all)
* FILE_TEST() checks file existence without opening it.
* file size can be obtained via (FSTAT(f)).SIZE
* FILE_INFO() (only in IDL 5.5) can get file size without opening it.
* sometimes (very rare) RESTORE can not restore all saved variables
and restore only part of them messaging "Truncating..." and CATCH does
not catch this situation (AFAIK). So I recommend to use CMSVLIB
library of Craig B. Markwardt
(http://cow.physics.wisc.edu/~craigm/idl/cmsave.html) which he kindly
grants to IDL community.
Best regards,
Altyntsev Dmitriy
Remote Sensing Center, ISTP
Irkutsk, Russia
http://ckm.iszf.irk.ru
MKatz843@onebox.com (M. Katz) wrote in message news:<4a097d6a.0203081046.1deaf7e@posting.google.com>...
> Here's in I/O error-trapping question for the IDL gurus out there...
>
> Working in IDL 5.4 on SunOS UNIX, I'd like to catch errors
> that occur when reading a bad or incomplete--but existing--
> IDL SAVE file. I want to use ON_IOERROR, but it doesn't
> seem to catch errors that occur with the RESTORE function.
> Here's more information...
>
> I create a smallish (100-200 bytes) disk file with
>
> SAVE, x, y, z, filename="myfile.dat"
>
> Since I have lots of programs running simultaneously, different
> parts of my program or different IDL sessions may be
> SAVEing and RESTOREing these files in quick succession.
>
> Before I RESTORE, I always check for the existence of the file with
> FRELUN, 10
> OPENR, 10, filename, ERROR=err
>
> Sometimes, the file exists but is empty (!), then IDL tries to RESTORE
> and comes back with an END OF FILE error.
>
> SO, I put in an error catcher:
>
> ON_IOERROR, error_label
>
> Yet the program still halts with an END OF FILE, and the error
> trap is never activated. Any ideas?
> I would like the RESTORE file to trip the error trap and go
> nicely to my error-handling code.
>
> Thanks, MKatz
|
|
|
|
| Re: Catching I/O errors on RESTORE [message #29802 is a reply to message #29671] |
Tue, 12 March 2002 16:06 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
alt@iszf.irk.ru (Altyntsev Dmitriy) writes:
>>
>> I would very much like to hear reports on how well the CMSVLIB SAVE
>> library does with corrupted files. It should report errors back to
>> user programs without crashing.
>>
> It really does. Thanks. At least as far as I had opportunity to check.
> I had problems with ordinary RESTORE and none of them is present now.
> I am going to get huge statistics on CMSVLIB stability during my
> software package running (save-restore every 10 sec through network)
> this summer (forest fire season). Certainly I will inform you about
> results.
Thanks for the report!
> Craig, it would be impudence to wish more but if your library had an
> ability to replace (expand, squeeze, delete) variables in sav-files it
> would be absolutely indispensable for organizing of IDL file container
> and would be used very widely as I think. Just a wish, sorry :-))
You're right. It would be impudence.
Craig
[ Joking of course. ] Replacing, expanding, squeezing, deleting, are
all much harder operations than what CMSAVE can do right now. The
/APPEND option was "easy" in the sense that all you do is tack on some
more data at the end.
You might be amazed, at least I was, that IDL does not have a way to
trim an existing file. I'm talking about the same thing that
ftruncate() does in C. So "squeezing" would be very hard.
Probably then the easiest way to "replace" a variable is to restore
the file, replace the variable you want, then save it again. At least
at the moment.
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@cow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
