On Sat, Nov 25, 2000 at 11:02:54AM -0500, Daniel Burrows wrote:
>
> How about this: "While it's possible to make a program deal with any user
> unput without crashing, this just won't ever happen." Based on empirical
> observation, this is just as true..hmm, maybe I should close a raft of bugs
> against my packages }=)
The difference between that and my original reasoning that
originally, I did not believe at that time that there was a solution
(mostly because I thought that rename() would also fail on a full
disk, an error that I am gratified to note has been made by at least
one other person in this thread). I have never claimed to be a master
programmer, and I claim the right to make mistakes. If he had
bothered to correct my mistake the first time I closed the bug, this
mess would never have gotten this far. If he hadn't felt the need to
taunt me on a public list, this might have been dealt with by now.
A better analogy to what happened from my perspective might be,
"While it's possible to make a program with an interactive display on
XWindows stay running and retain its state even when the user presses
CTL-ALT-BSP, to do so is unfeasable and to expect it to happen for
every program is unreasonable." Nick has a past history of being
unreasonable, and I've been under a bit of stress lately, so perhaps I
was less willing to extend him the benefit of the doubt than I should
have been. I apologize for this. I place as a footnote, however,
that having him blindly reopen the bug without bothering to comment
upon my reasoning frankly didn't make me any more inclined to take him
seriously.
> > Yes, it's good coding practice to check every function for successful
> > completion, but how often does that really happen in
> > non-critical code?
>
> Well, I'd hope quite often.
In this case, in my original reasoning, it wouldn't have helped,
since the damage would have been done by the time the error message
was read.
> Luckily, it looks like the maintainer finally understands that writing a
> temporary file and using rename() is a good idea, so maybe he'll forward it..
I understood that it was a good idea as soon as I realized that it
was actually a functioning solution. That happened as soon as Nick
bothered to respond to my closure explanation, which I will note
didn't happen until _after_ he posted to this list.
As is stated in the bug logs, it will get forwarded sometime next
week, once I've had time to calm down and get on a regular sleep cycle
again. I'm not normally so thin skinned, but as I said, I'm under a
bit of stress and the nerves are a little raw, and I don't want to be
composing a request to an upstream author who has sometimes been a
little touchy when every time I think about it my blood pressure
rises. Having to respond to this thread hasn't helped any, either.
--
Zed Pobre <zed@debian.org> a.k.a. Zed Pobre <zed@resonant.org>
PGP key and fingerprint available on finger; encrypted mail welcomed.
Attachment:
pgpMaS6ZbsIWH.pgp
Description: PGP signature