This is the mail archive of the cygwin@cygwin.com 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]

Re: Setup use on win2k - a me too ...



----- Original Message -----
From: "Pavel Tsekov" <ptsekov@syntrex.com>
To: "Gareth Pearce" <tilps@hotmail.com>
Cc: <cygwin@cygwin.com>
Sent: Tuesday, November 06, 2001 11:34 PM
Subject: Re: Setup use on win2k - a me too ...


> Gareth Pearce wrote:
> >
> > Due to the fact that recompiling to latest cvs appears to have fixed the
> > problem misteriously ... already ... I was not supprised that adding
those
> > lines produced
> >  ...
> > site: ftp://planetmirror.com/pub/sourceware/cygwin
> > get_url_to_string ftp://planetmirror.com/pub/sourceware/cygwin/setup.ini
> > ftp > 220 PlanetMirror FTP server ready.
> > FTP command is USER anonymous
> > ftp > 331 Anonymous login ok, send your complete email address as your
> > password.
> > ftp > 230-
> > ftp >   Welcome to the FTP archives at planetmirror.com.   You
> > ftp >   can also access these via a web interface reachable from:
> > ftp >
> > ...
> >
> > notice that the log records the 220 before the USER...
>
> This is so because there is a read just after the connect which is
> logged

umm ... so the client doesnt wait for any information from the ftp server
before entering the username?
I thought that waiting for 220 was something that would be done... maybe I
am dreaming though :)
Maybe I should read some of the code sometime ;)
>
> > the log from the time where it consumed memory evilly, it recorded x  as
the
> > first thing after the get_url_to_string...
> >
> > so I would guess that the bug stems previous to that...
>
> IMO if the problem is triggereddepends on who compiled the
> setup.exe. The pointers net_ftp_user and net_ftp_passwd are not
> zeroed in the code so it's up to the compiler if these are
> set to zero or just some random garbage. For example the MS compiler
> sets all unitialized variables to zero if you're building debug
> version of an app and leaves these as is if you're building release
> unless you specify otherwise explicitly with an compiler option.

quite possible ... although in the 3 occasions I tested, they were all done
on the same computer with gcc, 1 failed the others didnt, the one that
failed was version 128, the ones that didnt were 129... this could be
spurious ... since I have had programs where adding a variable declaration
in file A, resulted in program stopping working due to variable B in file C
not being initialised to 0 anymore... (since I hadnt initialised it at all
in the first place ... oops)

>
> --
> Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
> Bug reporting:         http://cygwin.com/bugs.html
> Documentation:         http://cygwin.com/docs.html
> FAQ:                   http://cygwin.com/faq/
>
>

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


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