This is the mail archive of the
mailing list for the Cygwin project.
Re: Sparse file criteria malfunction - binutils produces sparse .exe & .dll files
Christopher Faylor wrote:
> On Thu, Jun 05, 2003 at 07:29:51PM +0100, Max Bowsher wrote:
>> I assumed you would trust someone telling you whether the read-only
>> attribute of a file was set, without needing to see further evidence?
>> To me, this is an equivalent situation.
> Nope. I wouldn't. I'd ask for "attrib" output. That's tech support
> 101. And, I'd do it even though there is a system utility to do that.
> Strangely enough, I have this argument with our tech support guys all
> of the time. I recently had a situation where I lost a day of debugging
> because I "trusted" someone to have done the right thing and it turned
> out that the problem, which I couldn't duplicate, was due to someone
> incorrectly reporting a version number. They didn't provide the output
> of an rpm command or anything. They just said "I'm pretty sure it's
> version X".
Fine, I'll try not to cut corners in my information reporting:
I tested freshly-built new-cygwin1.dll cygwin0.dll and cygserver.exe with
both my tool and Pierre Humblet's tool - both agreed the files were sparse.
>>>> Personally I think "Don't risk anything if there is no potential gain"
>>>> is reasonably persuasive.
>>> Lets use the popular reasoning here. If there was no potential gain
>>> then Microsoft would not have provided the option, would they? Since
>>> they did there has to be *some* potential gain.
>> But their option isn't on by default. So, presumably, there must be
>> some circumstances when it is undesirable.
> That wasn't the point here. You were saying there is *no potential
> gain* not that there were situations in which it was undesirable. I'm
> just using your reasoning to show that there has to be a potential gain.
> I'm not necessarily agreeing with your reasoning but, IMO, you can't
> claim both that "If it was good Microsoft, would have turned it on by
> default" and "There is no potential gain", since in the former you trust
> Microsoft and in the latter you don't.
My fault, I've snipped too much context, and/or not explained myself well
enough in the first place. What I intended to say was:
I would like the ability to turn off sparse file creation, because *in the
ways I use Cygwin*, it cannot benefit me, but it could inconvenience me.
Now, AFAIK, sparse files will only benefit people using BitTorrent or
similar, or people in very large domains (lastlog).
I.e., this does not apply to a significant number of Cygwin users.
Are there any other cases when sparse files can help?
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html