HTTP Reason Phrase

[Erkki Lindpere]

It seems Lift does not support custom HTTP Reason Phrases in
responses. I would like to send error messages in the Reason Phrase
(along with a vaguely applicable HTTP status code) in a RESTful API
I'm providing. My understanding of the HTTP spec is that the reason
phrase is meant to be human readable and does not have to contain the
recommended messages (i.e. "Not Found").

But maybe it would not be wise to do this? I'm not actually aware of
any API-s that send error messages in the Reason Phrase.

[Ross Mellgren]

I think it would be fine to have different text there, probably better than having the standard text if you have refined detail. I don't think it'd be a good idea to conditionalize on the response text in client code - that's always fragile. If you want to give additional machine-readable detail, I'd put it in a response header or in the body as a JSON or XML field or what have you.

You can specify custom text there, but you may have to sidestep the usual response classes, depending on which one. The one you gave, not found, can have the message customized though, just do new NotFoundResponse("the message").

-Ross

> --
> You received this message because you are subscribed to the Google Groups "Lift" group.
> To post to this group, send email to lif...@googlegroups.com.
> To unsubscribe from this group, send email to liftweb+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/liftweb?hl=en.
>

[Erkki Lindpere]

The NotFoundResponse (and others) puts the custom message into the
request body. That works as well, but to be really lean (mainly for
bragging rights :)) I'd like to remove any unnecessary elements from
the rest api. Most of the error messages are going to be simple one-
line messages (and short sentences). For some errors I might provide a
detailed response and it could go into the body, as XML/JSON/...
That's inconsistent if the other errors have a plain text message in
the body.
So I could either go with structured details for all messages or in
simple cases use the HTTP headers or status line. A custom header
would work, but the status line is standard and also more easily
accessed with some client libraries.

And last but perhaps not least, for debugging purposes, the HTTP
Reason Code should show up better in web developer tools (for example
FireBug, Chrome's tools). My web UI also goes through the REST API so
it would be nice to read error messages right in the listing in
firebug's net panel.

So I'm suggesting that perhaps Lift would like to provide only the
possibility of changing that value in user code. But I completely
understand if it doesn't. Currently it doesn't seem to be supported in
Lift's http.provider package and even in javax.servlet the
setStatus(int, String) method is deprecated (I'm not sure if
sendError(int, String) uses the reason phrase).

Erkki L

[Timothy Perrett]

If you want to alter the Reason-Phrase, you can already do that - objects like NotFoundResponse are just helpers on InMemoryResponse... nothing stopping you adding your own helpers that set headers with customised reason codes; this should give you what you what. I haven't managed to find an RFC that lists reason-phrase as anything but a particular header in the HTTP response.

Moreover, its not the wrong thing to return a plain text response if the request mime was text/plain... indeed, it would be even less helpful if it returned JSON or such. IMHO, the response type should match what was asked for by the caller - i.e. its an implementation issue not a framework level issue.

Tim

[Erkki Lindpere]

Actually, the reason phrase is not a normal HTTP header, but part of
the status line:
http://www.w3.org/Protocols/rfc2616/rfc2616-sec6.html#sec6.1

Status-Line = HTTP-Version SP Status-Code SP Reason-Phrase CRLF

Examples:
HTTP/1.1 200 OK
HTTP/1.1 404 The user with id 8 does not exist

The only way of setting this in Java Servlets as far as I know is
through HttpServletResponse.setStatus(int, String), which
unfortunately is deprecated. A non-deprecated possibility is
sendError(int, String), but that goes to the container's default error
page (or the one defined in web.xml, I think) so it's not exactly what
I would like.

Also, I checked that FireBug actually does display the custom reason
phrase, but Chrome displays the standard one instead.

Erkki L

[Ross Mellgren]

Yeah you're very correct. It's unfortunate, but I think since it's deprecated in the container we should probably not add support for it. I can't believe they deprecated it for the reason they did, but there it is.

-Ross

[Timothy Perrett]

I agree Ross... I would be very reluctant to have Lift rely on some deprecated method in the servlet spec - even if it is in servlet 3.0.

Cheers, Tim

[Erkki Lindpere]

Ok. The feature is not really that important for me, just something
that would be nice to have. I think some hack could be made with
sendError(int, String) and web.xml config, but that would be worse
than using the deprecated method.

Erkki L

[David Pollak]

I think this issue rests with Marius.  He's done most of the interface between Lift and the servlet containers.  Let's see what he has to say.

Lift, the simply functional web framework http://liftweb.net
Beginning Scala http://www.apress.com/book/view/1430219890
Follow me: http://twitter.com/dpp
Surf the harmonics

[Marius]

I think it worth to add this support even if servlet API is a little
weird sometimes but Lift HTTP API could provide a nicer support.
Furthermore such Lift API support would be handy for non JEE
containers therefore I'd vote for it.

Erkky would you please open a defect (https://www.assembla.com/spaces/
liftweb/tickets) and assign it to me ?

Br's,
Marius

> > liftweb+u...@googlegroups.com<liftweb%2Bunsu...@googlegroups.com >

> > .
> > > >>>>>> For more options, visit this group athttp://
> > groups.google.com/group/liftweb?hl=en.
>
> > > >>>> --
> > > >>>> You received this message because you are subscribed to the Google
> > Groups "Lift" group.
> > > >>>> To post to this group, send email to lif...@googlegroups.com.
> > > >>>> To unsubscribe from this group, send email to

> > liftweb+u...@googlegroups.com<liftweb%2Bunsu...@googlegroups.com >

> > .
> > > >>>> For more options, visit this group athttp://
> > groups.google.com/group/liftweb?hl=en.
>
> > > >> --
> > > >> You received this message because you are subscribed to the Google
> > Groups "Lift" group.
> > > >> To post to this group, send email to lif...@googlegroups.com.
> > > >> To unsubscribe from this group, send email to

> > liftweb+u...@googlegroups.com<liftweb%2Bunsu...@googlegroups.com >

> > .
> > > >> For more options, visit this group athttp://
> > groups.google.com/group/liftweb?hl=en.
>
> > > > --
> > > > You received this message because you are subscribed to the Google
> > Groups "Lift" group.
> > > > To post to this group, send email to lif...@googlegroups.com.
> > > > To unsubscribe from this group, send email to

> > liftweb+u...@googlegroups.com<liftweb%2Bunsu...@googlegroups.com >

> > .
> > > > For more options, visit this group athttp://
> > groups.google.com/group/liftweb?hl=en.
>
> > --
> > You received this message because you are subscribed to the Google Groups
> > "Lift" group.
> > To post to this group, send email to lif...@googlegroups.com.
> > To unsubscribe from this group, send email to

> > liftweb+u...@googlegroups.com<liftweb%2Bunsu...@googlegroups.com >

> > .
> > For more options, visit this group at
> >http://groups.google.com/group/liftweb?hl=en.
>
> --
> Lift, the simply functional web frameworkhttp://liftweb.net

> Beginning Scalahttp://www.apress.com/book/view/1430219890

[Timothy Perrett]

Ah yes - keep forgetting one day we'll have netty et al providers!

Interested to see what you come up with Marius

Cheers, Tim

> To unsubscribe from this group, send email to liftweb+u...@googlegroups.com.

[Erkki Lindpere]

Opened ticket:
https://www.assembla.com/spaces/liftweb/tickets/334-Support-for-HTTP-Reason-Phrase

In case you are busy with other things, I might have time this weekend
to create a patch.

Erkki L