Ok, found the issue (probably specific to my environment). One of the
cookies seems to be "badly formatted", specifically "lib/pb_server.py"
(still trying to find where that one comes from).
In web2py main.py, the following code will only load the cookies up to the
bad one:
# ##################################################
# load cookies
# ##################################################
if env.http_cookie:
try:
request.cookies.load(env.http_cookie)
except Cookie.CookieError, e:
pass # invalid cookies
Not sure whether it would be worthwhile making it more resilient?
Regards,
Dominic.
On Tuesday, 4 December 2012 00:33:45 UTC+1, Dominic Cioccarelli wrote:
>
> Hi Massimo,
>
> sure, the parsing code is as follows:
>
> def getCookie(name):
> cookies = self.request.env.http_cookie.split(';')
> for cookie in cookies:
> cookie = cookie.strip()
> if cookie.startswith(name):
> temp = cookie.split('=', 1)
> return temp[1].strip('"')
> return None
>
> The cookie header (partially modified for obvious reasons) is included at
> the bottom of this post. My function can return the value of, for example,
> "monsterCookie" which is otherwise not visible via the standard cookie
> class.
>
> Regards,
> Dominic.
>
> zzeuck_answ=0; SI_baq09Z0lnXXAFA8_intercept=true; OAX=UodP6FB8EyoAAAIE;
> ab-test=A; UPS_AnonymousID=hJA3HAAAMhYXK; zzdddc=37;
> com_zz_jumpid=em_r11999_au/au/hho/pps/eppint-id999-12cw42_em_edm_pu_login_pd/employee-offer_str/20121019;
>
> uid=D+CoIlCJtIOVpgdbDp1vAg==; SI_ai2DPVK1I1MVhFG_intercept=true;
> s_pers=%20s_vnum%3D1437515522943%2526vn%253D2%799937515522943%3B%20s_invisit%3Dtrue%7C1399941895778%3B%20s_visit%3D1%999351941895781%3B%20aVc1%3D1%7C1351941895784%3B%20s_MC%3D1%7C9994532095788%3B%20s_LrI%3D1%7C1354532095790%3B;
>
> ATZZ_COOKIE=prefCountry%3DDE|prefViewGlobalContent%3D0|prefLang%3Dde;
> ATZZ_RETURN_URL=
> http://itsupport.zz.com/portal/site/sg/documentdetail?docid=KM0XXXXX;
> internalwikidb_mw__session=4hv22njzzzfbq0bdigxxxultp5;
> wikidb_mw__session=pxxxr0mvu84gja3srs7aobjsl0;
> Remember_me=3ac1c95e346d26150a963eb8225909eb36a536f6e10a3d7df56f12bxxx3ea424a02c853f69b69fcb;
>
> wikidb_mw_UserID=4055; wikidb_mw_UserName=bozo.hound%40zz.com;
> wikidb_mw_Token=eac777e6fcb2xxx38ac008874e656dcb;
> session_id_hotfixdb=199.195.185.83-570b5331-a11a-416a-9c81-393145fe1c4a;
> ovrd_last_page=https%3A//
> support.frogcompanydingo.zz.com/casemanager/case/SD00160439;
> lib/pb_server.py.session_id=4tuumlSfG16/4vJFNhwBKFHkHf4=; lang=en; cc=GB;
> OVRD_Cookie=editprofiletimestamp=1121122000044Z;
> CL_Cookie=?UTF-8?B?cHJlZmVycmVkbGFXXXVhZ2U9ZW58aHBjbG5hbWU9ZGNpb2NjYXJ8aHBjbGlkbnVtYmVyPTJhYTI1ZDUyMWQ0MGUyYjU5OTJhYWI3N2QyNGRlODY3fGhwcmVzaWRlbnRjb3VudHJ5Y29kZT1BVXxzbj1DaW9jY2FyZWxsaXxnaXZlbm5hbWU9RG9taW5pY3xlbWFpbD1kb21pbmljLmNpb2NjYXJlbGxpQGhwLmNvbXxjcmVhdGV0aW1lc3RhbXA9MjAwMi0wNC0xNyAwMjozNDo1NHxtb2RpZnl0aXXXc3RhbXA9MjAxMi0xMC0xNSAxMzo1MTo1NXxjbGFuZz1BVS1lbg==;
>
> c_lang=?UTF-8?B?TEFXXX9DT0RF; clang=?UTF-8?B?QVUtZW4=;
> ovrdtarget=?UTF-8?B?aHR0cXXXLy9zdXBwb3J0Lm9wZW52aWV3LmhwLmXXXS9lbnRpdGxlbWVudC9jc3MvY29udHJhY3RzXXXzcw==;
>
> internalwikidb_mw_UserID=73; internalwikidb_mw_UserName=bozo.hound%
> 40zz.com; internalwikidb_mw_Token=951621afe99965XXXfb3daf6ac7ba405;
> s_sv_sid=999905932487; zzeuck_prefs=1111;
> com_zz_hope=g**H**v**2**r**2**s**1999922368246**x**1920x1080**m**h%7C0%7C0%7C0%3Xo%7C19%7C0%7C0%3Ba%7C1%7C0%7C0**t**0%3Aa%7C3%7C0%7C0%3A12454%3Aa%7C1%7C0%7C0**l**0;
>
> has_js=1; prop12=r10978; EMID=; s_cc=true;
> s_sq=zzhqintraatzzintranetportaltest%3D%2526pid%253AAAZZ%252520%25257C%252520Home%252520%2526pidt%253D1%2526oid%253Dhttps%25253A%252F%
> 252Fwww.rooms.zz.com%252Fsignin.aspx%2526ot%253DA;
> s_vi=[CS]v1|28316ZZZ850125E4-40000109990D67EF[CE]; cert_auth_cookie=1;
> auth_level_cookie=15; zzGlobalID=ZZKRML; zzEmail="bozo.ho...@zz.com";
> cn="Bozo Hound"; givenName=Bozo; sn=Hound; employeenumber=00527999;
> bozo=bozo;
> monsterCookie="MRXW22LONFRS4Y3JN5RWGYLSMVWGY2XXXBYC4Y3PNURUI33NNFXGSYZAINUW6Y3DMFZGK3DMNERUI33NNFXGSYZDINUW6Y3DMFZGK3DMNERTAMBVGI3TKMJWENNFUS2SJXXX====";
>
> session_id_admin=199.195.185.82-25582396-caaa-4308-bad3-337c7bebc5a4;
> SMSESSION=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;
>
> session_id_m3t=199.195.185.82-50aaa5a5-aa99-4873-a3aa-57ba1d1932d8
>
>
> On Monday, 3 December 2012 22:52:04 UTC+1, Massimo Di Pierro wrote:
>>
>> Can you post your cookie parsing code? It would help us understand the
>> problem.
>>
>> On Monday, 3 December 2012 15:28:12 UTC-6, Dominic Cioccarelli wrote:
>>>
>>> Hi all,
>>>
>>> I have discovered an issue with the way web2py treats the "cookies"
>>> header. Specifically, when the header grows bigger than a certain size,
>>> web2py doesn't retrieve all the required cookies for the domain in question.
>>>
>>> In my case, the application is running on an intranet for a rather big
>>> company. This means that many cookies are potentially available to an
>>> application (those that are stored at the domain level) meaning that the
>>> length of the overall cookies header grows.
>>>
>>> If I iterate over all the headers in Java (JSP) I get a complete set. If
>>> I then redirect to a web2py application, I only get the first x headers
>>> (where x is dependant of the previous cookie's lengths). The *
>>> request.env.http_cookie* variable is complete however.
>>>
>>> Whilst I've managed to code around this issue (by writing my own
>>> cookie retrieval function that parses request.env.http_cookie) there is
>>> a problem with the inbuilt session management code as it uses the
>>> underlying (broken) cookie code to store the session ID.
>>>
>>> I haven't had a chance to look at the source code for cookie management
>>> in web2py but wanted to check whether anyone else had experienced this
>>> issue?
>>>
>>> Regards,
>>> Dominic.
>>>
>>>
>>>
>>>
--
