How to route in USP code?

[shivshankardayal]

Consider I have an usp file say one.usp which gives me one.so so that my URL served is http://localhost/one. Now if I want to serve http://localhost/one/two then how can I achieve that?

[victorstewart]

just create a directory named "one" in your document root directory, and put two.so inside it

[stefanocasazza]

Hi,

I think that better than words is a prototype:
one.usp
one.test
one.req

[shivshankardayal]

Thanks for the prototype. This mean routing is non-trivial. I have another question. Consider url to be http://localhost/one/100 where 100 is some id in database on which selection is to be done. How can I extract 100 from the URL? Which class contains this information? I believe UHTTP will have that but in doxygen information I could not find a method which is obvious.

[victorstewart]

if you construct the url like this... http://localhost/one?dbID=100

then either...

1. you can access the raw text in the UClientImage_Base::body variable
   OR
2. you can declare "dbID" in the args section.

Here's a quick example:

benchmarking.usp

[shivshankardayal]

No I do not intend to do that. That conflicts with SEO based URL design.

[stefanocasazza]

Hi, in the prototype I exposed a more general routing, you can also simple make a checking: if (UClientImage_Base::request_uri->equal(U_CONSTANT_TO_PARAM("/one/two"))) ...; else .... so you can see that the information about the URL is in this variable... Bests Stefano 2017-09-14 17:26 GMT+02:00 Shiv Shankar Dayal <notifications@github.com>:

…

Thanks for the prototype. This mean routing is non-trivial. I have another question. Consider url to be http://localhost/one/100 where 100 is some id in database on which selection is to be done. How can I extract 100 from the URL? Which class contains this information? I believe UHTTP will have that but in doxygen information I could not find a method which is obvious. — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#30 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAGkXorO92lzXYwey1TwbdilvqUBvSJVks5siUWwgaJpZM4PWInF> .

[victorstewart]

I'm not sure you can achieve what you deserve to though. You're asking for conditional url parsing, with no clear differential between cases of pure resource paths vs resource paths with an identifier tagged on.

This is the whole reason ?x=y etc exists.

If you point the request at "/one/100", that will be interpreted as a resource path.. and ULib will look for a resource at your-document-root/one/100.so

What you're asking for would require parsing that violates general URL standards / specifications.

[shivshankardayal]

No it does not. For example, look at URL of a question on stackoverflow. The post id of question is embedded in URL. For example, this is achievable easily in frameworks like Django.

[stefanocasazza]

Hi,

If you point the request at "/one/100", that will be interpreted as a resource path.. and ULib will look for a resource at your-document-root/one/100.so

in case of using the more general routing exposed in the prototype it does not. ULib will look for a resource at your-document-root/one.so cause of the config var: USP_AUTOMATIC_ALIASING /one Gived the url: https://stackoverflow .com/questions/1657484/can-you-give-an-example-of-stack-overflow-in-c can you describe which methods you wish to extract information, I can add them to ULib... Stefano 2017-09-14 17:56 GMT+02:00 Shiv Shankar Dayal <notifications@github.com>:

…

No it does not. For example, look at URL of a question on stackoverflow. The post id of question is embedded in URL. For example, this is achievable easily in frameworks like Django. — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#30 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAGkXlihl-V5PLetH7QByVdYOQZtaVpVks5siUy1gaJpZM4PWInF> .

[shivshankardayal]

Given the URL https://stackoverflow .com/questions/1657484/can-you-give-an-example-of-stack-overflow-in-c I want it to be served by questions.so and 1657484 is what needs to be extracted from URL.

[stefanocasazza]

Hi, ok. I imagine in this way. config var: USP_AUTOMATIC_ALIASING /questions questions.usp: <!--#cpath id; --> it is ok for you? Also can you give an example of what you consider a trivial routing? Bests Stefano 2017-09-15 10:44 GMT+02:00, Shiv Shankar Dayal <notifications@github.com>:

…

Given the URL https://stackoverflow .com/questions/1657484/can-you-give-an-example-of-stack-overflow-in-c I want it to be served by `questions.so` and 1657484 is what needs to be extracted from URL. -- You are receiving this because you modified the open/close state. Reply to this email directly or view it on GitHub: #30 (comment)

[stefanocasazza]

I made as general rule:
/name_of_usp/cpath1/cpath2/... be served by name_of_usp.so.
It is no more needed the config var USP_AUTOMATIC_ALIASING
I have update the prototype:
one.usp
two.usp
one.test
one.req

[shivshankardayal]

Thanks a lot for helping me out Stefano. Now one more problem remains which I raised in another issue which is persistent connection to DB.

[stefanocasazza]

it is show inside one.usp

[shivshankardayal]

So you mean to say that since it is static connection it will live as long as userver will keep running. I think that this issue should be closed now. Thanks a lot Stefano for all the support.