[ http://jira.dspace.org/jira/browse/DS-640?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Stuart Lewis updated DS-640: ---------------------------- Assignee: Stuart Lewis Status: Open (was: Received) [09:12am] tdonohue: Interal System Error when browsing with wrong argument : http://jira.dspace.org/jira/browse/DS-640 [09:12am] PeterDietz: looks like I've duplicated this with: http://jira.dspace.org/jira/browse/DS-684 [09:13am] ksclarke joined the chat room. [09:13am] tdonohue: Ok, we need to mark one as a dupe then [09:14am] tdonohue: any volunteers to implement? This would be good to fix for 1.7 (in my opinion) [09:14am] richardrodgers: Probably should 404? [09:14am] sandsfish joined the chat room. [09:14am] mhwood: Reporter says there is a patch, but didn't attach one. [09:14am] stuartlewis: 404 sounds good. [09:14am] tdonohue: +1 -- yea, 404 would probably be appropriate [09:14am] stuartlewis: Assign to me if you want. [09:15am] tdonohue: Ok DS-640 - Assign to stuartlewis for 1.7. Should return 404 (mark DS-684 as a dupe and close) [09:15am] grahamtriggs: -1 on 404... I think.... remember, it's the same base url (with parameters) as the valid pages. Will 404 have consequences? [09:15am] mhwood: I've been meaning to ask: how do we get contact information for issue reporters, so we can ask for more details etc? [09:15am] tdonohue: mhwood -- if you add a comment to an issue, it will email them [09:16am] richardrodgers: alternate suggestion grahamtriggs ? [09:16am] mhwood: Hah, of course. Thanks. [09:16am] grahamtriggs: Friendly message - 'No such browse' ? [09:17am] tdonohue: yea -- that could be a good point, grahamtriggs. Not sure if all search engines will treat those as separate pages (as they are just parameterized) [09:18am] richardrodgers: I'm a little worried about the search engine case - could expand the URL space searched.... [09:19am] richardrodgers: but I don't know enough to say one way or another [09:19am] PeterDietz: then maybe 301 to default browse [09:19am] tdonohue: Or, it could just send you to a page listing all the valid browse options? [09:21am] tdonohue: suggest that we have stuartlewis investigate options and suggest a route -- 404, friendly error, etc [09:21am] stuartlewis: xmlui for /browse/ returns 404 and doesn't seem to cause any harm [09:22am] stuartlewis: Actually, returns 200, but message of page not found [09:22am] mhwood: Ugh, that 200 needs fixing. [09:22am] grahamtriggs: but then http://demo.dspace.org/xmlui/browse/?type=title doesn't return a browse page either [09:23am] tdonohue: yea -- I think doing the same as http://demo.dspace.org/xmlui/browse/?type=SomethingNotIndexed should be fine [09:24am] stuartlewis: But what http status code? [09:24am] tdonohue: oh, but without the slash is an error still: http://demo.dspace.org/xmlui/browse?type=SOMETHINGNOTINDEXED [09:24am] stuartlewis: 404, or 200 [09:24am] mdiggory: http://www.google.com/search?sourceid=chrome&ie=UTF-8&q=dspace+browse#sclient=psy&hl=en&safe=off&q=+browse+site:http%3A%2F%2Fdspace.knust.edu.gh%2F&aq=f&aqi=&aql=&oq=&gs_rfai=&pbx=1&fp=5a996d56de453056 [09:24am] ThomasAJohnson left the chat room. (Ping timeout: 265 seconds) [09:24am] mdiggory: different parameters = different hits [09:24am] stuartlewis: mdiggory: Isn't that < 1.5 browse? [09:24am] grahamtriggs: tdonohue: yes, because any url that includes the slash gets the not found - even when you specify a valid type (ie. title) [09:24am] mhwood: Should not be 200 if it's an incorrect request. [09:25am] tdonohue: I thought we had another JIRA issue already open around the 200 vs. 404 issues in XMLUI? [09:25am] mhwood: Yes, from way back, dependent on Cocoon problems that they never fix. [09:25am] mdiggory: don't know I was just huntingfor an example of browse indexed in google still [09:27am] PeterDietz: yeah they have different results for: dspace/browse-date and dspace/browse-date?order=oldestfirst in google [09:27am] mdiggory: http://www.google.com/search?q=browse++Nuclear+site%3Ahttp%3A%2F%2Fdspace.mit.edu [09:27am] mdiggory: Mit still does it [09:28am] tdonohue: ok, maybe we don't have a 200 v. 404 JIRA issue opened -- I thought we did though, as I recall this discussion coming up recently [09:28am] grahamtriggs: I think we should err on the side of caution and say that if we need a path to return valid documents for some parameters, then it should never return 404 for any invalid parameters [09:29am] mhwood: Hmmm, well, what's HTTP for "invalid parameters"? [09:29am] PeterDietz: 400 [09:29am] PeterDietz: Bad Request [09:29am] mdiggory: more specifically, for just /browse? http://www.google.com/search?q=browse++Nuclear+site%3Ahttp%3A%2F%2Fdspace.mit.edu%2Fbrowse [09:29am] stuartlewis: Sounds sensible - so in this case, what should the response be? "200, sorry, invalid browse" [09:30am] PeterDietz: its a client error that they gave you bad input [09:30am] tdonohue: yea -- 400 Bad Request is probably the right one for bad params. [09:30am] tdonohue: "400, Sorry, invalid browse" seems better [09:30am] tdonohue: http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html [09:30am] mdiggory: true [09:30am] mhwood: We should not return OK for an uninterpretable request. A well-formed request which finds nothing should return OK. [09:30am] stuartlewis: It all depends on google et al differentiating between pages with query string parameters. Surely they do? So many sites work that way. [09:31am] stuartlewis: 404 OK for some parameters, 200 for others. [09:31am] sandsfish: what we've done is prohibit invalid browse URL processing from adding anything the DRI body, which in turn activates the Manakin PageNotFound transformer. [09:31am] tdonohue: Google does differentiate between pages w/query string params. Top 2 results here are same, except for querystring: http://www.google.com/search?q=browse++Nuclear+site%3Ahttp%3A%2F%2Fdspace.mit.edu [09:32am] grahamtriggs: what about proxies? [09:32am] mbklein joined the chat room. [09:33am] mhwood: We're 35 minutes in. Take this topic to the dev list? [09:33am] tdonohue: ok -- we're at 30 mins any final decisions, or should we take ithis to email? [09:33am] tdonohue: yea, thanks mhwood [09:33am] stuartlewis: I'll summarise options on JIRA and transfer to dspace-devel > Interal System Error when browsing with wrong argument > ------------------------------------------------------ > > Key: DS-640 > URL: http://jira.dspace.org/jira/browse/DS-640 > Project: DSpace 1.x > Issue Type: Bug > Components: Documentation, JSPUI > Affects Versions: 1.5.2, 1.6.0, 1.6.1, 1.6.2 > Reporter: Hardik Mishra > Assignee: Stuart Lewis > > On Browsing Items: > If someone tries to use browse type for which browse index does not exist, > like browse=publisher > e.g. http://dspace.webinito.com/browse?type > OR > e.g. http://dspace.webinito.com/browse?type=xyz > You will get Internal System Error -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.dspace.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira ------------------------------------------------------------------------------ Beautiful is writing same markup. Internet Explorer 9 supports standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. Spend less time writing and rewriting code and more time creating great experiences on the web. Be a part of the beta today. http://p.sf.net/sfu/beautyoftheweb _______________________________________________ Dspace-devel mailing list Dspace-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dspace-devel