Backup autosave file... Bad idea?

So how is this routine supposed to get called? If you're going to go to the
trouble to call a routine to back up the AutoSave file, it seems to me it
would be just as simple to use the QSAVE command.
___

 

Hi all,

Here's what I've got, but I'm worried about unforeseen consequences. Does
anyone see any problems with this? Of course, recommendations for
improvement are always welcome as well.

;;; AutoSave makes a copy of AutoCAD's current autosave file.
(defun AutoSave (arglst / date sfp file1 file2)
(setq date (getvar "date")
sfp (getvar "savefilepath")
file1 (getvar "savefile")
file2 (strcat sfp "\\AutoSave " (getvar "dwgname"))
)
(if (and (or (not ASV:TimeToSave) (>= date ASV:TimeToSave))
file1
)
(progn
(prompt "\nBacking up autosave file ... \n")
(vl-file-delete file2)
(vl-file-copy file1 file2)
(setq ASV:TimeToSave (+ date (/ (getvar "savetime") 1440.0)))
)
)
(princ)
)
;;; EditorRxr is the reactor callback function called by the
"vlr-Editor-reactor"
(defun EditorRxr (calling-reactor commandInfo /)
(vl-load-com)
(AutoSave '(nil))
(princ)
)
;;; The Editor reactor
(vlr-Editor-reactor
nil
(list
'vlr-commandWillStart . EditorRxr)
)
)

 

Judging from the responses so far, maybe I should be more specific (and I
probably could have worded the subject line better...). This is intended to
run automatically (hence the reactor) in the background and allow the
"autosave" file to be retrieved if something happens to the primary working
file. It basically alleviates the (IMHO) problem of AutoCAD deleting its
own autosave file on any 'normal' exit. I'm sure this is what I want to do,
I'm just concerned about possibile ramifications of my code and/or approach
in terms of performance, stability, etc.

As far as copying the autosave file rather than 'just' saving again, I
figured the save has already been done by AutoCAD and a simple file copy
would be more efficient. If I'm wrong on that, someone please let me know.
Besides, I'm not sure how to save the current file to a different location
(leaving the current file active in the same location) without interrupting
the user. I've actually wanted to do that before for different purposes,
but that's another topic.

I appreciate your thoughts Paul and Dean.

 

OK, so set up your reactor to simply do a SAVEAS to a specific drive, rather
than depending on AutoCAD's flaky AutoSave process.

Honestly, in over 20 years with AutoCAD, we've had to depend on AutoSave
fewer than half a dozen times, and none of those in at least the past five
years. AutoSave is a poor substitute for good work habits.
___

 

Am I incorrect that a file copy is quicker / more efficient than a save? I
am basing this only on working observation. It just seems like I have to
wait longer for a save operation than a file copy operation on a given file.

I am content to let my solution *follow* AutoCAD's. As far as I can tell,
the built-in autosave works fine except for the file disappearing. I
believe circumventing that disappearing act is all I need to do.

I couldn't agree more. Do I want to depend on autosave? Absolutely not.
But..., the next time a user loses their work by whatever means (happened
again recently, hence my current focus), I believe it is in the company's,
the user's, and my best interests if I have found a simple way to quickly
get us back in business. That solution is what I think I have here. I only
wish I had done this sooner...

I appreciate everyone's input on this already, but am still hoping for some
technical feedback regarding performance/stability. There are 3 reasons I
went with the file copy as opposed to a save/as.

1. I do not want to change the active file's location.
2. I do not want to alter the currently saved primary file (i.e. saveas
over it to restore the original location), in case the user has
intentionally not saved yet.
3. As mentioned at first, I thought it was more efficient anyway.

Thank you all.

James

 

You don't want to do a SAVEAS, because that leaves you working in that new
file that it makes, rather than leaving you in your original file. You want
SAVE instead. And don't think you're doing this with AutoCAD's so-called
"Save" menu & toolbar items -- the ones that say "Save" or have a disk icon
actually do a QSAVE. The actual SAVE command can be invoked in a lisp
routine or macro or whatever, or typed in, but I don't think you can do it
from their standard menu items. Check out the difference in Help.

Kent Cooper, AIA


...

 

Wow - I have a totally different experience with autosave. I use it all the
time as we have flaky drawings that crash. My SAVETIME is five minutes to
minimize loss of work. I never have an issue with the .SV$ files
disappearing. The only bad luck is if a drawing currently being edited has
become permanently corrupt resulting in all saved files including autosaves
& bak's being unopenable. (pretty rare hasn't happened in a year).

To make it easier to find the autosave folder, I use this in a menu:
ID_ExploreTemp [Explore Temp Folder with AutoSave Files]^C^C^P(startapp
(strcat "explorer.exe /e,"(getvar "tempprefix")))

Dan

 

AutoSAVE is protection only against catastrophic failures and is not a good substitute for good work habits. If AutoCAD exits the program properly the SV$ is deleted. If any of AutoCAD's SAVE features are implemented the SV$ is deleted. That is the proper function of AutoSAVE.