El dom, 23-10-2005 a las 09:52 -0400, Stuart Brorson escribió: > Lots of good ideas in here! Sorry if my responses are short -- I > gotta run soon. > > > - Noweb: I don't know anyone who has benefit of using it. > > I am not a fan of noweb, but I have learned to live with it. Anyway, > I understand that Patrick, who has contributed a large number of very > valuable patches, is the primary motivator behind noweb. I wouldn't > want to alienate him since his work has been valuable in the past, and > will likely be valuable in the future. Therefore, I wouldn't dump > noweb unless Patrick himself wants to do it.
Such a change like switching to C files has to be agreed. That's the reason I asked for opinions. [snip] > BTW: Did you look at how the buttons work in gattrib? I have > forgotten . . . . I don't know, but it can take the code from gschem when it is done. [snip] > > - Autosaving: > > Great idea! Also, I prefer real_name.sch~. And don't delete teh > autosave file since a backup with valid stuff in it is important. > > > - Autosaving when crashing: > > Here's a suggestion, whcih is a variant of your proposal: > > 1. gschem generates a backup file of the old schematic when you read > it in. Call it real_name.sch~ > > 2. Gschem generates an autsave file every 1 minute. Call it > #real_name.sch#. > > I believe this is how emacs works, and since emacs makes all computing > simple we should follow its example. I like it. Emacs also deletes #real_name.sch# after exiting. This way, you will have only the real_name.sch~ backup copy after working with real_name.sch . > Also, rather than making gschem handle the question of which file to > import, let the user just mv the files around on the command line (or > let the project manager deal with it). It's better that the next time gschem loads real_name.sch, it looks for #real_name.sch#. If it exists, then a previous execution crashed (it didn't delete the autosave copy), and then gschem asks the user what to do. I think the user shouldn't know how gschem works, but he should know that his work is safe. > BTW: Have you looked at the project manager "geda" and how it > handles these issues? I am thinking that we should try to bring it up > to the same level as the rest of gEDA/gaf. I agree, but I'd like to implement some things in gschem first.... Cheers, Carlos
