Well, actually, it's because LIST/STAT is acting in a case-sensitive
manner on that server.

I've actually written by own FTP server (which works with most other
browsers and clients, but it hasn't been tested too much.)  However,
it doesn't work with Google Chrome, and this is because it has a bug:
LIST filename doesn't work, only LIST directory.

This is Chrome trying to download a file:

Notice: Accepted connection (127.0.0.1:3439)
SC: 220 Hello
CS: USER ftp
SC: 331 What's the password?
CS: PASS ftp
SC: 230 You're in.
CS: CWD /Testing/GestureMatch.zip
SC: 550 No can do.
CS: TYPE A
SC: 200 Ok.
CS: PASV
SC: 227 I'm waiting. (127,0,0,1,193,115)
CS: LIST /Testing/GestureMatch.zip
SC: 150 Opening data connection.
SC: 226 File doesn't exist; nothing to list.
Notice: Closed connection

In contrast, MSIE looks more like this:

Notice: Accepted connection (127.0.0.1:4131)
SC: 220 Hello
CS: USER ftp
SC: 331 What's the password?
CS: PASS ftp
SC: 230 You're in.
CS: TYPE I
SC: 200 Ok.
CS: PASV
SC: 227 I'm waiting. (127,0,0,1,204,103)
CS: SIZE /Testing/GestureMatch.zip
SC: 213 25678
CS: RETR /Testing/GestureMatch.zip
SC: 150 Opening BINARY mode data connection (25678 bytes)
Notice: Waiting for data connection...
SC: 226 Done.

This is of course a bug in my implementation; LIST should return info
about even a single file.  However, it points to what's happening here
with Microsoft's server (I typed this into telnet but you get the
idea):

220-Microsoft FTP Service
220 MARS-FTP2
USER anonymous
331 Anonymous access allowed, send identity (e-mail name) as password.
PASS [EMAIL PROTECTED]
230 Anonymous user logged in.
CWD /pub/tr/tr-2001-50.pdf
550 /pub/tr/tr-2001-50.pdf: The directory name is invalid.
TYPE A
200 Type set to A.
STAT /pub/tr/tr-2001-50.pdf
213-status of /pub/tr/tr-2001-50.pdf:
213 End of status.
STAT /pub/tr/TR-2001-50.pdf
213-status of /pub/tr/TR-2001-50.pdf:
05-01-01  11:58AM                78322 TR-2001-50.pdf
213 End of status.
QUIT
221

Doing it the way IE (and Firefox iirc) do it, using SIZE, seems much
more logical and bug-proof to me.

-[Unknown]


On Sep 9, 10:20 pm, Darin Fisher <[EMAIL PROTECTED]> wrote:
> I have no idea.  If you have time, you should run a TCP tracing tool like
> WireShark to see what the data looks like over the wire.  You can compare
> the different browsers to see what Chrome is doing that is funny here.-Darin
>
> On Tue, Sep 9, 2008 at 10:14 PM, Wei Hu <[EMAIL PROTECTED]> wrote:
>
> > Well, I suppose ftp.research.microsoft.com is running Windows, so the
> > server shouldn't be case sensitive, right? Then why didn't Chrome
> > download the file?
>
> > On Sep 10, 1:03 am, Darin Fisher <[EMAIL PROTECTED]> wrote:
> > > Interesting.  We are currently using WinInet for FTP, which is the same
> > code
> > > that Internet Explorer uses.  I wonder why it doesn't work.
> > > -Darin
>
> > > On Tue, Sep 9, 2008 at 9:47 PM, Wei Hu <[EMAIL PROTECTED]> wrote:
>
> > > > OK, Chrome is very strict about upper cases and lower cases in the
> > > > file names. Chrome is happy to download
> > > >ftp://ftp.research.microsoft.com/pub/tr/TR-2001-50.pdf,
> > > > but notftp://ftp.research.microsoft.com/pub/tr/tr-2001-50.pdf.
> > > > Strangely enough, Chrome doesn't care about the directory names. For
> > > > example,ftp://ftp.research.microsoft.com/PuB/tR/TR-2001-50.pdfworks
> > > > just fine. I'm wondering what the FTP protocol has to say about this?
>
> > > > Even if Chrome is conforming to the standard in this case, IE,
> > > > Firefox, wget, and ncftp all seem to be able to download both
> > > > tr-2001-50.pdf and TR-2001-50.pdf.
>
> > > > On Sep 4, 9:06 pm, Wei Hu <[EMAIL PROTECTED]> wrote:
> > > > > E.g.,ftp://ftp.research.microsoft.com/pub/tr/tr-2001-50.pdf
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Chromium-dev" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/chromium-dev?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to