We are trying to build a Rails web-application, and there has been this weird issue with random string being inserted to the URL right after the host domain (haven't seen it happening anywhere else in the URL). One example is like this:
https://www.<domain>.com/ZQpfZ/<some route>
This bug/error doesn't seem to be happen every time, but definitely a bit unpredictably as we generate this URL using path helpers from Rails. Another occasion that this would happen is I would directly type into the URL bar: www.<domain>.com, and it occasionally would insert some random string right after I press enter (with no auto-fill) as such:
https://www.<domain>.com/d4rsQf
What would be the reason for this and how should we go about solving this? It has been causing a lot of problem as we have been sending email with account confirmation link embedded, and sometimes the link won't work just because of the random string being added.
Any help would be greatly appreciated!
Related
I have a problem passing special parameters from a bash script to a browser URL bar.
I would like to open a PDF file on a certain page, and with a certain zoom factor, and at a certain position (relative to a page corner, I think that's called "focus") in Brave, from the command line.
I thought that should work like this:
$ brave-browser file:/home/user/foo.pdf#page=5&zoom=150,50,50
(the ",50,50" would be the "focus") but it doesn't. It just provokes an error message.
Copy/pasting the entire line with all the special characters manually to the URL bar works fine, but I would like to do that without the mouse or keyboard, from the CL. In other words, copying the string
file:/home/user/foo.pdf#page=5&zoom=150,50,50
and pasting that straight into the URL bar will open the file the way I want in Brave.
This shortened command works from the CL:
$ brave-browser file:/home/user/foo.pdf#page=5
but adding the &zoom=150,50,50 provokes an error message from Brave.
I just can't get my head around how to format the rest of the line so that the file opens in the browser as intended (zoomed and focussed), when called from the CL.
I begin to wonder if that is possible at all.
I have tried everything that was suggested here:
Shell script to open a URL
and a couple of other ideas I had (mainly, creative use of quotation marks)
I also read this, but I am not sure if it refers to the same question - and the post is 11 years old. Was hoping that things changed:
Can a website pass focus to the browsers url field?
Thanks a lot for any help!
Just found out that all it takes is a backslash in front of the ampersand...
Strange: no matter how I do write that here, with the *** before and after the code, or without the three *** , I can never see the backslash in the post's preview!
browser file:/home/user/foo.pdf#page=5 \ &zoom=150,50,50
I have to type blanks before and after the hash to make it visible in this post as above. Looks wrong but this is the only way I can show what I mean. Also strange: the hash # works without the backslash \
Thanks, should someone have looked into this in the meantime!
I'm trying to use Google Federated Login REST API. I can succesfully reach out to the google server and validate a user but I cannot pull parameters from the return url
for example:
http://mysite.com/login/return?openid.ns=http%3A%2F%2Fspecs.openid.net%2Fauth%2F2.0&openid.mode=id_res&openid.op_endpoint=https%3A%2F%2Fwww.google.com%2Faccounts%2Fo8%2Fud...
All the variables in that return string are not accessible in the params array. I have no idea how to get them out. requst.url, request.query_parameter, and all similar calls do not return the query string.
I think i found the issue. I was using the open-uri library to make the call to google's endpoint url so it may have been stepping outside of the normal rails response/request cycle. I've since used Net::HTTP requests and parse the information from the response.
So I have a very similar issue, where I'm actually building a Rails-based openid provider but being consumed by another Rails app. I basically adapted the code from
The whole URL was:
http://localhost:3000/openid?openid.assoc_handle=%7BHMAC-SHA1%7D%7B5193d33f%7D%7BdBrUwQ%3D%3D%7D&openid.claimed_id=http%3A%2F%2Flocalhost%3A3000%2Fopenid%2Fwarren&openid.identity=http%3A%2F%2Flocalhost%3A3000%2Fopenid%2Fwarren&openid.mode=checkid_setup&openid.ns=http%3A%2F%2Fspecs.openid.net%2Fauth%2F2.0&openid.ns.sreg=http%3A%2F%2Fopenid.net%2Fextensions%2Fsreg%2F1.1&openid.realm=http%3A%2F%2Flocalhost&openid.return_to=http%3A%2F%2Flocalhost%2Fsession%3F_method%3Dpost%26return_to%3D&openid.sreg.required=nickname%2Cemail
I had a similar problem where the only parameters being reported were:
{"action"=>"index", "controller"=>"openid"}
So, suspecting that some parameter (maybe a period?) was causing it to hiccup, I went through and deleted them one by one until I found that deleting the following parameter enables the entire thing to go through correctly:
openid.mode=checkid_setup
That left all the remaining parameters correctly being parsed:
{"openid.assoc_handle"=>"{HMAC-SHA1}{5193d33f}{dBrUwQ==}",
"openid.claimed_id"=>"http://localhost:3000/openid/warren",
"openid.identity"=>"http://localhost:3000/openid/warren",
"openid.ns"=>"http://specs.openid.net/auth/2.0",
"openid.ns.sreg"=>"http://openid.net/extensions/sreg/1.1",
"openid.realm"=>"http://localhost",
"openid.return_to"=>"http://localhost/session?_method=post&return_to=",
"openid.sreg.required"=>"nickname,email",
"action"=>"index",
"controller"=>"openid"}
I'm now trying to find why openid.mode causes this issue. It fails even if I change it to openid.mode=5, so it's the key, not the value, causing the problem.
Suspecting the ".mode" part of the string for the trouble (maybe ".mode" is a filetype or something being parsed by the routing?) I am looking towards this post on allowing periods, but it only applies to the value, not the key: rails routing and params with a '.' in them
Will report back if I find more.
Update: I tried, in another Rails app, adding ?openid.mode=0 to the end of a URL -- ".mode" does not result in a parameter being read, but ".modes=" does and so does ".mod=". This confirms that ".mode" is causing a params parsing error.
Update 2: yikes... actually "?a.mode=0" does work. So far, only the complete string "openid.mode" does not work, and this is across various Rails apps. "?openid.mode" with nothing else results in: Parameters: {"openid.mode"=>nil}, but "?openid.mode=" with nothing after the "=" fails to pass any parameters besides action & controller. Very odd.
Update 3: OK, figured it out, I believe -- the params were getting sanitized i.e. deleted by the rack-openid gem, in that gem's path: /lib/openid.rb:168, "sanitize_query_string". This seems to be incompatible with the example I was working with: https://github.com/openid/ruby-openid/tree/master/examples/rails_openid. Going to override that method.
Final update: I replaced this line:
oidreq = server.decode_request(params)
with this line, since we could no longer use the now-empty params hash:
oidreq = server.decode_request(Rack::Utils.parse_query(request.env['ORIGINAL_FULLPATH']))
If an extra character (like a period, comma or a bracket or even alphabets) gets accidentally added to URL on the stackoverflow.com domain, a 404 error page is not thrown. Instead, URLs self correct themselves & the user is led to the relevant webpage.
For instance, the extra 4 letters I added to the end of a valid SO URL to demonstrate this would be automatically removed when you access the below URL -
https://stackoverflow.com/questions/194812/list-of-freely-available-programming-booksasdf
I guess this has something to do with ASP.NET MVC Routing. How is this feature implemented?
Well, this is quite simple to explain I guess, even without knowing the code behind it:
The text is just candy for search engines and people reading the URL:
This URL will work as well, with the complete text removed!
The only part really important is the question ID that's also embedded in the "path".
This is because EVERYTHING after http://stackoverflow.com/questions/194812 is ignored. It is just there to make the link, if posted somewhere, if more speaking.
Internally the URL is mapped to a handler, e.g., by a rewrite, that transforms into something like: http://stackoverflow.com/questions.php?id=194812 (just an example, don't know the correct internal URL)
This also makes the URL search engine friendly, besides being more readable to humans.
I found this. It seems kind of ugly to just throw this into environment.rb. The question is kind of old. I just wanted to ask if there was a better way to do this now.
The problem (if you don't want to click through), is including your custom log message formatter. (The problem I'm solving is that I want to assign a guid to every request, prepend all log messages from that request with the guid, and then return the guid in meta data to the client a.k.a. request id)
Rails logger format string configuration
You can put it in an initializer.
Initializers can be any name and go in config/initializers. Every file in this directory is loaded at startup; it's a great place to put miscellaneous startup code that doesn't seem to fit anywhere else.
I tried this:
Test
but it stops after the first :S. Anyone know why?
It will show a link for:
S:US:1123">Test
I'm confused. :)
Working fine for me, but as a general rule, always ensure your query string parameters are URL Encoded, e.g:
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=250778518281&ssPageName=ADME%3AB%3ASS%3AUS%3A1123