Hi Steven, Thanks a lot for your response & explanations! I think, I can see where you are coming from, being a user myself for (roughly) 99.999% of all software I use. ;-) However, I want to see and fix (or at least understand) a potential problem rather then refreshing the screen to hide it. Then again, please feel heartily invited to participate on the Mondo Rescue mailing list and to get involved with development!
That said, your suggestion that this may be a symptom of the issue you raised in #379966, sounds quite plausible to me. So, would you be able to re-run with mondo-2.09-2 to double-check whether the problem persists with the change for #379966 in place? Thanks a lot & best regards! Andree On Sat, 2006-09-02 at 17:24 -0400, Steven M. Robbins wrote: > Hello Andree, > > Apologies for the silence: I have been away from my Debian computer > for most of the last month. > > Quoting Andree Leidenfrost <[EMAIL PROTECTED]>: > > >> I can think of two strategies to mitigate this. Perhaps you want to > >> do both. Strategy #1 is to make sure all subprocesses have their stdout > >> and stderr redirected appropriately, possibly to the log file. Strategy > >> #2 is to clear and re-draw the screen completely before each status > >> update, to guard against failure to apply strategy #1. > > > > #1 is what should be pretty much in place. Your problem shows that this > > is certainly not 100% the case. > > Yep; and since there are always bugs, there's a reasonable chance that > at any given time one of them is going to mess up the display. That's > why I would suggest that strategy #2 is a good idea. > > > > #2 strikes me as not very efficient and more of a workaround to gloss > > things over. > > I don't accept any efficiency argument. Mondo is reading and writing > gigabytes of data during the backup. Adding, say, an extra 80x40 > characters every half second will be lost in the noise. > > On the other hand, I have some sympathy for the "gloss things over" > sentiment, coming from a computer science / programmer background. > However: when I put on my "user" hat that argument cuts no ice when > considered against the probability that the display is so messed up to > be unusable. > > > > I wouldn't want to refresh the screen frequently because it > > may contain something important. > > The refresh should certainly keep all the important text up-to-date. > > > > That said, I'd like to get to the root of the evil here and either do #1 > > or fix the underlying problem. To me this looks like a system() or > > popen() call in mondoarchive might have a syntax error in its argument. > > > > Would you be able to narrow down what exactly it is doing? Is there any > > way to match this against things that appear in the log (you could use > > '-K 9999' to increase logging verbosity)? Could you come up with a > > small test example that I could use to reproduce the issue like you did > > for the other bug? > > I'm virtually certain that this *is* caused by the other bug > (#379966); i.e. that some filename has a backquote in it. > > Cheers, > -Steve > -- Andree Leidenfrost @ Debian Developer Sydney - Australia
signature.asc
Description: This is a digitally signed message part
