2012/11/12 Damien Cassou <damien.cassou@gmail.com>:
it's a backup! You probably never need any backup... until you do :-).
Exactly! ;-)
In a perfect world Pharo would be perfect, never crash and never loose any data.
To be fair, I have not had a VM crash in quite some time. What I do manage to archieve is to make the image unresponsive (either by halting where I am not supposed to, or introduce delays/infinite-recursions in the UI thread, etc.). Then I have to kill the VM. It is sad having to kill the VM, not being able to rescue your unsaved work. But Mariano is right, and I *am* aware of the "recover lost changes" functionality (it was not always the case though), and I would like to address his comments/concerns...
On Mon, Nov 12, 2012 at 8:42 PM, Mariano Martinez Peck <marianopeck@gmail.com> wrote: Now, I would really like to understand if this is a real need of a new tool or a demonstration that the current tools need some improvements. As you may know, we always have the .changes and therefore we can tools to recover lost data from there. We have the change sorter, the "recover lost code", etc. We are even able to load the .changes in another image. So...my question is, what is the actual problem/limitations with those which are "solved" with this new tool?? or in other words, why do we need this new tool? should we improve the already existing tools instead?
Mariano, you raise very valid points and questions. I know about the "recover lost code" functionality, and have had to use it myself a couple of times already. And was very glad and relieved after the fact! :-) When I post this, I don't do it to question or challenge core Pharo functionalities. I do it as a "user" sharing a solution that I programmed for myself, for my own uses, and hoping it might be useful for somebody else. Maybe it *is* confusing if I post something like this on the "pharo-project" mailing list. I apologize for this. Now, having said this, let me tell you that *I* (me) am not always happy with the "recover lost changes" way of recovering my changes. For one, I find it tedious to have to sort through do-its that just *might* be important, or maybe not... It takes too long, it is something that demands too much concentration from me, and a process where I can do many things wrong (what happens if I forget a couple of methods? can I be sure I didn't forget anything). And then there are the renames, class deletions, etc. On top of that, I think that not all might land on the ".changes"... I had the case where a class creation was not there... Maybe it was a class rename, or something like that. Me, if I have to decide between sorting through a long list of changes, or just doing "file-in" from a couple of files and be done in seconds, I take the latter approach ;-) Again, this is a solution I hacked for me, and am just sharing with the world :-). It's not meant as a "reconver lost changes" replacement... But I understand your concerns, and am glad you raised them. Cheers, Sebastian