----- Original Message ----- From: "Denis Vlasenko" <[EMAIL PROTECTED]> To: "JÃrgen Hovland" <[EMAIL PROTECTED]>; "Henrik Nordstrom" <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Monday, March 15, 2004 6:39 AM Subject: Re: [squid-users] invalid url
> On Monday 15 March 2004 03:56, JÃrgen Hovland wrote: > > Ok I just commented out the few lines in src/url.c starting at line 309 > > (#if CHECK_HOSTNAMES....) > > > > It works! Henrik Nordstrom> No it does not, it only looks like it works. You are wrong, it works. You can try for yourself. > > Here comes the however: > > When using MSIE the hostname points you to the wrong website. > > If I disable proxy in MSIE I get to the correct website. There seems to be > > an encoding problem somewhere. I telneted manually to squid and requested > > www.jÃrgen.nu in ISO-8859-1 and UTF-8 mode and it worked. Seems like IE use > > urlencoding when using proxy, or IE use it all the time ? Is this a IE bug > Denis Vlasenko> Use tcpdump to find out Good idea. 15:07:44.500574 client.3617 > webserver.80: P 1:213(212) ack 1 win 65000 <nop,nop,timestamp 811567 958736353> (DF) 0x0000 4500 0108 0428 4000 8006 b669 d5b3 39ef E....(@....i..9. 0x0010 d4b5 5b06 0e21 0050 ece9 947e 73f6 4f0a ..[..!.P...~s.O. 0x0020 8018 fde8 bbb7 0000 0101 080a 000c 622f ..............b/ 0x0030 3925 27e1 4745 5420 2f20 4854 5450 2f31 9%'.GET./.HTTP/1 0x0040 2e31 0d0a 4163 6365 7074 3a20 2a2f 2a0d .1..Accept:.*/*. 0x0050 0a41 6363 6570 742d 4c61 6e67 7561 6765 .Accept-Language 0x0060 3a20 6e6f 0d0a 4163 6365 7074 2d45 6e63 :.no..Accept-Enc 0x0070 6f64 696e 673a 2067 7a69 702c 2064 6566 oding:.gzip,.def 0x0080 6c61 7465 0d0a 5573 6572 2d41 6765 6e74 late..User-Agent 0x0090 3a20 4d6f 7a69 6c6c 612f 342e 3020 2863 :.Mozilla/4.0.(c 0x00a0 6f6d 7061 7469 626c 653b 204d 5349 4520 ompatible;.MSIE. 0x00b0 362e 303b 2057 696e 646f 7773 204e 5420 6.0;.Windows.NT. 0x00c0 352e 313b 202e 4e45 5420 434c 5220 312e 5.1;..NET.CLR.1. 0x00d0 312e 3433 3232 290d 0a48 6f73 743a 2077 1.4322)..Host:.w 0x00e0 7777 2e6a f872 6765 6e2e 6e75 0d0a 436f ww.j.rgen.nu..Co 0x00f0 6e6e 6563 7469 6f6e 3a20 4b65 6570 2d41 nnection:.Keep-A 0x0100 6c69 7665 0d0a 0d0a live.... 15:10:48.714521 client.3620 > proxy.80: P 1:231(230) ack 1 win 64240 (DF) 0x0000 4500 010e 0557 4000 8006 d63a d5b3 39ef [EMAIL PROTECTED]:..9. 0x0010 d5b3 3902 0e24 0539 efa8 38c2 0ead 4ffe ..9..$.9..8...O. 0x0020 5018 faf0 1de0 0000 4745 5420 6874 7470 P.......GET.http 0x0030 3a2f 2f77 7777 2e6a 2543 3325 4238 7267 ://www.j%C3%B8rg 0x0040 656e 2e6e 752f 2048 5454 502f 312e 300d en.nu/.HTTP/1.0. 0x0050 0a41 6363 6570 743a 202a 2f2a 0d0a 4163 .Accept:.*/*..Ac 0x0060 6365 7074 2d4c 616e 6775 6167 653a 206e cept-Language:.n 0x0070 6f0d 0a50 726f 7879 2d43 6f6e 6e65 6374 o..Proxy-Connect 0x0080 696f 6e3a 204b 6565 702d 416c 6976 650d ion:.Keep-Alive. 0x0090 0a55 7365 722d 4167 656e 743a 204d 6f7a .User-Agent:.Moz 0x00a0 696c 6c61 2f34 2e30 2028 636f 6d70 6174 illa/4.0.(compat 0x00b0 6962 6c65 3b20 4d53 4945 2036 2e30 3b20 ible;.MSIE.6.0;. 0x00c0 5769 6e64 6f77 7320 4e54 2035 2e31 3b20 Windows.NT.5.1;. 0x00d0 2e4e 4554 2043 4c52 2031 2e31 2e34 3332 .NET.CLR.1.1.432 0x00e0 3229 0d0a 5072 6167 6d61 3a20 6e6f 2d63 2)..Pragma:.no-c 0x00f0 6163 6865 0d0a 486f 7374 3a20 7777 772e ache..Host:.www. 0x0100 6ac3 b872 6765 6e2e 6e75 0d0a 0d0a j..rgen.nu.... I see that IE sends an url encoded GET line, which is what it is supposed to do. I would put my 5 euro in that this is a non-implemented squid feature. Henrik Nordstrom> This is not supposed to work. HTTP is specified ASCII only for host names. Henrik Nordstrom> Henrik Nordstrom> If you want to browse an IDN name you MUST today use a browser which has Henrik Nordstrom> IDN support. For MSIE this involves installing one of the IDN plugins. Im simply interested in getting things to work. Permitting such domains will acomplish this task. I have no control over all the browsers that will be using our proxy. Henrik Nordstrom> Henrik Nordstrom> > Seems like IE use urlencoding when using proxy, or IE use it all the Henrik Nordstrom> > time ? Is this a IE bug or a non-implemented squid feature (and will it Henrik Nordstrom> > be implemented) ? Henrik Nordstrom> Henrik Nordstrom> Your browser is not IDN capable. As you are trying to use your browser in Henrik Nordstrom> situations it is not designed for odd things happens as it is trying to Henrik Nordstrom> comply to standards given junk input. As you probably are aware of, there are probably more browsers out there not IDN capable than capable of IDN. Rejecting such domains in a proxy software is not going to help anyone. The smartest thing would be to automaticly translate to IDN in squid directly (as an optional choice of course). Cheers, Joergen
