| Restore IDL code from .sav file? [message #37478] |
Wed, 24 December 2003 12:13  |
Jonathan Joseph
Messages: 69
Registered: September 1998
|
Member |
|
|
Twas the night before christmas and I did something incredibly stupid!
I edited a couple of .pro files in the wrong location (old versions of
files). In a moment of temporary insanity, I didn't realize that they
were old and different from the files in my working CVS directory, so I
copied them right over the top of my good files, which had MANY changes
I had made in the past 5 days and was just (really just) about to check
back in.
After testing the code, I realized what I had done. Aaaaagh!
Date of last incremental backup: 5 days ago, just after I started with
these changes. Nothing useful to recover from tape.
Emacs ~ backup files: None. (an emacs setting that I haven't changed yet
makes it so that when I edit files in a checked out CVS directory, it
doesn't make ~ backups).
Single strand of hope: I happen to have another recent IDLDE session
open. It has compiled versions of the pre-trashed files (but does not
have the source code in a buffer). I just made a .sav file of all the
routines, so at least the compiled versions are not lost.
Question: Is there any way to recover the actual source from the .sav
file or from my IDLDE session that has the compiled versions of the good
code?
Thanks and Merry Christmas.
-Jonathan
|
|
|
|
| Re: Restore IDL code from .sav file? [message #37555 is a reply to message #37478] |
Wed, 31 December 2003 16:34  |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
On Mon, 29 Dec 2003 13:24:58 -0700, Craig Markwardt wrote:
> "Haje Korth" <haje.korth@jhuapl.edu> writes:
>
>> Jonathan,
>> hey that story sounds familiar, happened to me too once. Unfortunately
>> there is probably nothing you can do about it at this point. Craig
>> Markwardt had routines that do exactly what you want (and what I
>> needed) included in his cmsvlib for a few days. But then a herd of
>> people including RSI started beating on him for publishing proprietary
>> information, so he withdrew the routine prodis, which caused all this
>> hassle. [ ... ]
>
> After Kodak threatened a potential lawsuit, and implied a threat to my
> job, I did indeed withdraw the core translator for IDL save files with
> code.
>
> I believed, and I still believe, that there are legitimate uses for a
> save file translator, and it is not illegal. We as users are not
> required to protect RSI's flawed business models. I developed the
> translator code because I myself accidentally deleted some source code,
> and needed to recover it from a save file. Obviously other people have
> the same problem. I have been recommending that people with this same
> need contact RSI and explain their dilemma. And further, I recommend
> against using save files for code.
The basic issue, I believe, is that Kodak/RSI would like to move IDL
away from being a tool used by scientist who, almost without
exception, share their code with each other, to a "platform" for
developers and software entrepreneurs, ala Java. The ironic thing is
how completely open most IDL code is, despite the very closed nature
of the IDL core itself. And frankly, it is this spirit of open
sharing that has kept IDL alive. Even people like David Fanning, who
make a living selling IDL-related services and information, have
adopted an essentially open-source philosophy (though they may curse
it at times). Imagine postings like:
"I found a new solution to this tricky sorting problem which uses
HISTOGRAM to achieve speed-up of 10x or more. You can find it on
shareware.com as a .SAV file. It only costs $14.95, and you'll get
free updates for life after that."
That just would never work. That's not the kind of language IDL is;
it doesn't address questions that people want immediate, canned,
not-too-expensive answers to, like an mp3 player shaped like a banana
or an easy to use calendar widget with color-coded phases of the moon.
For better or worse, IDL is a tinkerer's language... and I'd be
willing to bet it will remain that, or fade into obscurity.
JD
P.S. In the interest of full disclosure, I should say that even I am
not immune to the lure of "compiled" IDL programs running under the
free IDLVM. Mea culpa.
|
|
|
|
| Re: Restore IDL code from .sav file? [message #37560 is a reply to message #37478] |
Tue, 30 December 2003 09:24 |
Jonathan Joseph
Messages: 69
Registered: September 1998
|
Member |
|
|
Thanks for your posts.
After posting that message, and kicking around a few chairs, I spent
about 4 hours reconstructing the changes I had made to the easier of the
two files. Then, as I was looking again through my other recent IDLDE
session, lo and behold, I found that I actually *did* have the second
file open in the IDLDE editor and the code was sitting there with all of
the changes in tact. I carefully did a save-as, ignoring the warnings
that the file had changed on disk and didn't I really want that version
instead.
So with a little luck and a good deal of patience, I was able to
recover/reconstruct the lost changes. Even without the source code for
the file I had truly lost, I was at least able to run to compiled
version to corroborate that with reasonable certainty I was able to
re-derive most of the changes I had made - I even found a spot where I
had accidentally forgotten to include something the first time around.
A very frustrating Christmas eve, but I think I've got the code back to
where I want it.
-Jonathan
Craig Markwardt wrote:
> "Haje Korth" <haje.korth@jhuapl.edu> writes:
>
>
>> Jonathan,
>> hey that story sounds familiar, happened to me too once. Unfortunately there
>> is probably nothing you can do about it at this point. Craig Markwardt had
>> routines that do exactly what you want (and what I needed) included in his
>> cmsvlib for a few days. But then a herd of people including RSI started
>> beating on him for publishing proprietary information, so he withdrew the
>> routine prodis, which caused all this hassle. [ ... ]
>
>
> After Kodak threatened a potential lawsuit, and implied a threat to my
> job, I did indeed withdraw the core translator for IDL save files with
> code.
>
> I believed, and I still believe, that there are legitimate uses for a
> save file translator, and it is not illegal. We as users are not
> required to protect RSI's flawed business models. I developed the
> translator code because I myself accidentally deleted some source
> code, and needed to recover it from a save file. Obviously other
> people have the same problem. I have been recommending that people
> with this same need contact RSI and explain their dilemma. And
> further, I recommend against using save files for code.
>
> Craig
>
>
>
>
>> ... His documentation gives you an overview and you will find, why RSI
>> was upset about that particulat version of cmsvlib, even though he does not
>> elaborate on the particular info.
>
>
> P.S. By the way, the on-line documentation is for data only, no code.
>
>
|
|
|
|
| Re: Restore IDL code from .sav file? [message #37565 is a reply to message #37478] |
Mon, 29 December 2003 12:46 |
Haje Korth
Messages: 651
Registered: May 1997
|
Senior Member |
|
|
Rick,
you are asking the right question: How much is it worth? And the answer is
probably not enough to justify expenses. That is why I have given up. I will
just rewrite the code using my weekends while ticking off my family for
hiding in my office. :-)
By the way, proving ownership is actually not that hard, since detailed
information is actually contained in the sav files. Craig's documentation
describes the sav file format quite well and it is easy to retrieve these
info. However, his documentation skims over the way functions and routines
are encoded (probably for a reason). Consequently, it is more time consuming
to figure out how to disassemble the sav files in general then it is to
rewrite my code. Therefore, I have not bothered since.
Haje
PS: I bet all the fuss is about the embedded IDL licensing that could be
uncovered. And just looking at the sav files and the inofficial
documentation, you can see that you do not need prodis to fake the
licensing. But again, who cares???
--
"Rick Towler" <rtowler@u.washington.edu> wrote in message
news:bspt4f$l22$1@nntp6.u.washington.edu...
> I would guess that someone at RSI could do this. Either they have their
own
> internal tools, a copy of prodis, or both. The only question is whether
> they would be willing to perform this service and how much they would
> charge. Oh, and they would probably want some assurance that you were the
> original author of the code which could prove to be tricky unless your
> installation # is embedded in the save file.
>
> So how much is the code worth?
>
> -Rick
>
>
>
> "Haje Korth" wrote in message ...
>> Jonathan,
>> hey that story sounds familiar, happened to me too once. Unfortunately
> there
>> is probably nothing you can do about it at this point. Craig Markwardt
had
>> routines that do exactly what you want (and what I needed) included in
his
>> cmsvlib for a few days. But then a herd of people including RSI started
>> beating on him for publishing proprietary information, so he withdrew
the
>> routine prodis, which caused all this hassle. I tried to search the
> internet
>> forever to retrieve the library to recover my code that I spent quite
some
>> time on, without success. So the frustrating answer is that there is a
way
>> to recover source code from sav file, but it is illegal. If you are
>> ambitious, you can download info on the sav file format from Craig's
>> website. His documentation gives you an overview and you will find, why
> RSI
>> was upset about that particulat version of cmsvlib, even though he does
> not
>> elaborate on the particular info.
>>
>> Haje
>>
>> PS: If you find prodis somewhere, let me know, I would love to recover
the
>> source of my code too. I tried the wayback machine, but they do not
store
>> binaries...
>>
>> --
>>
>>
>> "Jonathan Joseph" wrote in message...
>>>
>>> Twas the night before christmas and I did something incredibly stupid!
>>>
>>> I edited a couple of .pro files in the wrong location (old versions of
>>> files). In a moment of temporary insanity, I didn't realize that they
>>> were old and different from the files in my working CVS directory, so
I
>>> copied them right over the top of my good files, which had MANY
changes
>>> I had made in the past 5 days and was just (really just) about to
check
>>> back in.
>>>
>>> After testing the code, I realized what I had done. Aaaaagh!
>>>
>>> Date of last incremental backup: 5 days ago, just after I started with
>>> these changes. Nothing useful to recover from tape.
>>>
>>> Emacs ~ backup files: None. (an emacs setting that I haven't changed
yet
>>> makes it so that when I edit files in a checked out CVS directory, it
>>> doesn't make ~ backups).
>>>
>>> Single strand of hope: I happen to have another recent IDLDE session
>>> open. It has compiled versions of the pre-trashed files (but does not
>>> have the source code in a buffer). I just made a .sav file of all the
>>> routines, so at least the compiled versions are not lost.
>>>
>>>
>>> Question: Is there any way to recover the actual source from the .sav
>>> file or from my IDLDE session that has the compiled versions of the
good
>>> code?
>>>
>>> Thanks and Merry Christmas.
>>>
>>> -Jonathan
>>>
>>
>>
>
>
|
|
|
|
| Re: Restore IDL code from .sav file? [message #37566 is a reply to message #37478] |
Mon, 29 December 2003 12:24 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
"Haje Korth" <haje.korth@jhuapl.edu> writes:
> Jonathan,
> hey that story sounds familiar, happened to me too once. Unfortunately there
> is probably nothing you can do about it at this point. Craig Markwardt had
> routines that do exactly what you want (and what I needed) included in his
> cmsvlib for a few days. But then a herd of people including RSI started
> beating on him for publishing proprietary information, so he withdrew the
> routine prodis, which caused all this hassle. [ ... ]
After Kodak threatened a potential lawsuit, and implied a threat to my
job, I did indeed withdraw the core translator for IDL save files with
code.
I believed, and I still believe, that there are legitimate uses for a
save file translator, and it is not illegal. We as users are not
required to protect RSI's flawed business models. I developed the
translator code because I myself accidentally deleted some source
code, and needed to recover it from a save file. Obviously other
people have the same problem. I have been recommending that people
with this same need contact RSI and explain their dilemma. And
further, I recommend against using save files for code.
Craig
> ... His documentation gives you an overview and you will find, why RSI
> was upset about that particulat version of cmsvlib, even though he does not
> elaborate on the particular info.
P.S. By the way, the on-line documentation is for data only, no code.
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@REMOVEcow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
| Re: Restore IDL code from .sav file? [message #37567 is a reply to message #37478] |
Mon, 29 December 2003 10:49 |
Rick Towler
Messages: 821
Registered: August 1998
|
Senior Member |
|
|
I would guess that someone at RSI could do this. Either they have their own
internal tools, a copy of prodis, or both. The only question is whether
they would be willing to perform this service and how much they would
charge. Oh, and they would probably want some assurance that you were the
original author of the code which could prove to be tricky unless your
installation # is embedded in the save file.
So how much is the code worth?
-Rick
"Haje Korth" wrote in message ...
> Jonathan,
> hey that story sounds familiar, happened to me too once. Unfortunately
there
> is probably nothing you can do about it at this point. Craig Markwardt had
> routines that do exactly what you want (and what I needed) included in his
> cmsvlib for a few days. But then a herd of people including RSI started
> beating on him for publishing proprietary information, so he withdrew the
> routine prodis, which caused all this hassle. I tried to search the
internet
> forever to retrieve the library to recover my code that I spent quite some
> time on, without success. So the frustrating answer is that there is a way
> to recover source code from sav file, but it is illegal. If you are
> ambitious, you can download info on the sav file format from Craig's
> website. His documentation gives you an overview and you will find, why
RSI
> was upset about that particulat version of cmsvlib, even though he does
not
> elaborate on the particular info.
>
> Haje
>
> PS: If you find prodis somewhere, let me know, I would love to recover the
> source of my code too. I tried the wayback machine, but they do not store
> binaries...
>
> --
>
>
> "Jonathan Joseph" wrote in message...
>>
>> Twas the night before christmas and I did something incredibly stupid!
>>
>> I edited a couple of .pro files in the wrong location (old versions of
>> files). In a moment of temporary insanity, I didn't realize that they
>> were old and different from the files in my working CVS directory, so I
>> copied them right over the top of my good files, which had MANY changes
>> I had made in the past 5 days and was just (really just) about to check
>> back in.
>>
>> After testing the code, I realized what I had done. Aaaaagh!
>>
>> Date of last incremental backup: 5 days ago, just after I started with
>> these changes. Nothing useful to recover from tape.
>>
>> Emacs ~ backup files: None. (an emacs setting that I haven't changed yet
>> makes it so that when I edit files in a checked out CVS directory, it
>> doesn't make ~ backups).
>>
>> Single strand of hope: I happen to have another recent IDLDE session
>> open. It has compiled versions of the pre-trashed files (but does not
>> have the source code in a buffer). I just made a .sav file of all the
>> routines, so at least the compiled versions are not lost.
>>
>>
>> Question: Is there any way to recover the actual source from the .sav
>> file or from my IDLDE session that has the compiled versions of the good
>> code?
>>
>> Thanks and Merry Christmas.
>>
>> -Jonathan
>>
>
>
|
|
|
|
| Re: Restore IDL code from .sav file? [message #37568 is a reply to message #37478] |
Mon, 29 December 2003 06:09 |
Haje Korth
Messages: 651
Registered: May 1997
|
Senior Member |
|
|
Jonathan,
hey that story sounds familiar, happened to me too once. Unfortunately there
is probably nothing you can do about it at this point. Craig Markwardt had
routines that do exactly what you want (and what I needed) included in his
cmsvlib for a few days. But then a herd of people including RSI started
beating on him for publishing proprietary information, so he withdrew the
routine prodis, which caused all this hassle. I tried to search the internet
forever to retrieve the library to recover my code that I spent quite some
time on, without success. So the frustrating answer is that there is a way
to recover source code from sav file, but it is illegal. If you are
ambitious, you can download info on the sav file format from Craig's
website. His documentation gives you an overview and you will find, why RSI
was upset about that particulat version of cmsvlib, even though he does not
elaborate on the particular info.
Haje
PS: If you find prodis somewhere, let me know, I would love to recover the
source of my code too. I tried the wayback machine, but they do not store
binaries...
--
"Jonathan Joseph" <jj21@cornell.edu> wrote in message
news:3FE9F361.9060109@cornell.edu...
>
> Twas the night before christmas and I did something incredibly stupid!
>
> I edited a couple of .pro files in the wrong location (old versions of
> files). In a moment of temporary insanity, I didn't realize that they
> were old and different from the files in my working CVS directory, so I
> copied them right over the top of my good files, which had MANY changes
> I had made in the past 5 days and was just (really just) about to check
> back in.
>
> After testing the code, I realized what I had done. Aaaaagh!
>
> Date of last incremental backup: 5 days ago, just after I started with
> these changes. Nothing useful to recover from tape.
>
> Emacs ~ backup files: None. (an emacs setting that I haven't changed yet
> makes it so that when I edit files in a checked out CVS directory, it
> doesn't make ~ backups).
>
> Single strand of hope: I happen to have another recent IDLDE session
> open. It has compiled versions of the pre-trashed files (but does not
> have the source code in a buffer). I just made a .sav file of all the
> routines, so at least the compiled versions are not lost.
>
>
> Question: Is there any way to recover the actual source from the .sav
> file or from my IDLDE session that has the compiled versions of the good
> code?
>
> Thanks and Merry Christmas.
>
> -Jonathan
>
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
