This is the mail archive of the mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: bash/cmd CTRL-C problem...


I addressed my message to you out of politeness since it was intended for you. I
addressed the message to cygwin since I thought it might also be of interest to the
readers of the mailing list. It's not an unusual convention to follow on mailing
lists, however, I won't do this with you in the future.

When thinking about the purpose of cygwin prior to drafting my message, it did
occur to me that interop between cygwin and win32 apps might simply be incidental
(as you state) but there is also evidence that cygwin has gone to some efforts to
explicitly provide support for win32 (e.g. the cygpath utility). This leads me, and
clearly others, to the apparently mistaken conclusion that cygwin strives to
support win32 apps. This is what prompted me, and a number of others to report what
we saw as a cygwin bash bug.

You state that your message was intended to stem the flow of bug reports but the
cygwin bug reports page specifically states that bug reports should be sent to the
mailing list. Furthermore, looking over the contents of this and related threads I
can see no "me too" postings. All the postings shed additional light on the problem
so they all contribute to the eventual solution of the problem. Even if you argue
that the "problem" discussed in this thread isn't really a problem because cygwin
doesn't currently provide support for win32 apps, it certainly doesn't hurt to
discuss it in the off chance that members of the cygwin community feel that cygwin
should begin to explicitly support win32 apps. After all, it is the community that
defines the direction of cygwin by making contributions to it, and the mailing list
is where the community discusses issues like this.

The goals of cygwin notwithstanding, I agree wholeheartedly with your comments
regarding community software development. As you say, this is a process in which
"core developers" and others are free to take part in if they wish. Just so you
know that I believe in this approach to software development in general, and to
cygwin in particular, I should mention that I have made a modest contribution to
the cygwin project myself which solved a problem that I (and others) had posted bug
reports for. In the case of the bug reported in the current thread, I recognize
that I have neither the background, nor the time to acquire the background
necessary to solve this problem, so I am willing to live with it.... (but if Robert
Collins is looking at it, my thanks go out to him!)

The tone of your message indicates that you are fed up with "bug reports" that are
nothing more than veiled requests to "core developers" to fix a problem. I can
certainly sympathize with you since I see these "bug reports" myself on the mailing
list. On the other hand, you should recognize that, to a cygwin newcomer, the
codebase is more than a little daunting, even to an experienced Unix programmer.
Furthermore, as cygwin gets easier to install and use, cygwin users are
increasingly Unix users, not Unix system programmers. While I completely agree with
setting people straight on the "community development" aspect of the cygwin
project, my guess is that the reason for these requests is more more likely due to
a misunderstanding of how the cygwin project works rather than due to sloth.


Christopher Faylor wrote:

> On Mon, Jan 07, 2002 at 12:30:33PM -0500, Gregory W. Bond wrote:
> >(I just returned from vacation so excuse me for responding to an old
> >message but I really feel compelled to respond to this one...)
> Why did you Cc me when I specifically set the Reply-To to
> and am obviously reading that mailing list?
> >Christopher,
> >
> >This reply puzzles me. What is the point of cygwin if not to support
> >interoperability between win32 and unix environments? If all I wanted
> >was an all-unix environment on a 386 box, I'd run linux not cygwin. To
> >me, the great advantage of cygwin over linux is that I can (mostly) mix
> >and match Unix and Win32 apps (win32 Java under cygwin bash being the
> >notable exception :-)
> The point of cygwin is to provide a UNIX environment on Windows.  A
> cygwin app has lots and lots of extra stuff in it to support that
> functionality.  One of the extra things in a cygwin app is a substantial
> amount of code to make signals work right.  Native windows apps don't
> have this.  Having native windows apps work with cygwin is something that
> we get pretty much "for free" and many native apps work fine.
> However boundary cases like "My win32 process doesn't CTRL-C right" are
> of little interest to me.  My focus, when I am working on cygwin (which
> is less and less these days) is to improve UNIX functionality.
> Since cygwin is a community free software project, you have one privilege
> and possibly one detriment over commercial software:
> Privilege: It's free.  Source is available.  You can fix problems yourself.
> Detriment: It's free.  It is controlled by volunteers.  The volunteers are
>            under zero obligation to look at problems which don't interest them.
> I responded to the original message because I thought it was better to
> do that than let people continue to send "bug reports" and "me toos"
> here.  I was trying to set expectations.
> I also thought that it was just vaguely possible that someone would
> actually take some time to contribute something back to the cygwin
> project beyond the bug reports and me toos.
> It looks like Robert Collins may be investigating the problem now so
> you're lucky.  It's too bad that, as always, it reverts to one of the
> "core developers" to investigate something like this.  Very occasionally
> we actually get people who are reporting bugs to attempt fixes.  It's
> a shame that no one was willing to do that in this case.
> cgf

Gregory W. Bond
AT&T Labs - Research
180 Park Avenue, Rm. D273, Bldg. 103
P.O. Box 971, Florham Park, NJ, 07932-0971, USA
tel: (973) 360 7216 fax: (973) 360 8187

Unsubscribe info:
Bug reporting:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]