603159 - implement exslt-date:date-time()

Status: RESOLVED FIXED

(Core :: XSLT, enhancement)

Description

[Julian Reschke]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.10) Gecko/20100914 Firefox/3.6.10 (.NET CLR 3.5.30729)
Build Identifier: 3.6.10

The EXSLT date-time module contains many functions (http://www.exslt.org/date/index.html). However, implementing just date-time() out of these would be very useful, because it enables XSLTs to get the current system date (something that currently can't done at all).

Reproducible: Always

Steps to Reproduce:
Navigate to http://greenbytes.de/tech/webdav/xslt-test.xml
Actual Results:  
"Date-Time Support

exslt:date-time(): No"

Expected Results:  
"Date-Time Support

exslt:date-time(): Yes, and returns: 2010-10-10T09:40:29.095+01:00"

(with the current date)

Comment 1

[Julian Reschke]

Attachment: implementation of exslt:date-time()

1) the string handling probably should be double-checked

2) is there a canonical place for a test case?

Comment 2

[C. M. Sperberg-McQueen]

Having date-time() would certainly be useful.  I hope it can happen.

Comment 3

[Julian Reschke]

Any chance to get this into FF5?

Comment 4

[Julian Reschke]

Attachment: implementation of exslt:date-time(), plus a minimal test case

The test case is very minimal in that it only compares the prefix of the generated string (the year), but at least the extension function gets called.

Comment 5

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Use nsPrintfCString and CopyASCIItoUTF16 instead (yes, our string classes are not easy to use).

You can see how the generate-id function is implemented here:

http://mxr.mozilla.org/mozilla-central/source/content/xslt/src/xslt/txGenerateIdFunctionCall.cpp?force=1#83
http://mxr.mozilla.org/mozilla-central/source/content/xslt/src/xpath/txMozillaXPathTreeWalker.cpp#616

Also, is the string returned from this function parsable using the Date object in javascript? If so, it might be nicer to do so and check that the difference to Date.now() is less than, say, 30 minutes.

Comment 6

[Julian Reschke]

(In reply to comment #5)
> Use nsPrintfCString and CopyASCIItoUTF16 instead (yes, our string classes are
> not easy to use).
> 
> You can see how the generate-id function is implemented here:
> 
> http://mxr.mozilla.org/mozilla-central/source/content/xslt/src/xslt/txGenerateIdFunctionCall.cpp?force=1#83
> http://mxr.mozilla.org/mozilla-central/source/content/xslt/src/xpath/txMozillaXPathTreeWalker.cpp#616

Will do.

> Also, is the string returned from this function parsable using the Date object
> in javascript? If so, it might be nicer to do so and check that the difference
> to Date.now() is less than, say, 30 minutes.

I tried that last week and got an exception. I didn't investigate further as the the documentation on the 8601 aspect on Date.parse is a bit weak. It would be good to know what exactly it *is* expected to parse.

Comment 7

[Julian Reschke]

Maybe I can use

  https://developer.mozilla.org/en/JavaScript_code_modules/ISO8601DateUtils.jsm#Using_the_ISO.c2.a08601_date_utilities

in the test case?

Comment 8

[Julian Reschke]

Attachment: implementation of exslt:date-time(), plus a minimal test case

Comment 9

[Julian Reschke]

(In reply to comment #6)
> (In reply to comment #5)
> > Use nsPrintfCString and CopyASCIItoUTF16 instead (yes, our string classes are
> > not easy to use).
> > 
> > You can see how the generate-id function is implemented here:
> > 
> > http://mxr.mozilla.org/mozilla-central/source/content/xslt/src/xslt/txGenerateIdFunctionCall.cpp?force=1#83
> > http://mxr.mozilla.org/mozilla-central/source/content/xslt/src/xpath/txMozillaXPathTreeWalker.cpp#616
> 
> Will do.

Done.

> > Also, is the string returned from this function parsable using the Date object
> > in javascript? If so, it might be nicer to do so and check that the difference
> > to Date.now() is less than, say, 30 minutes.
> 
> I tried that last week and got an exception. I didn't investigate further as
> the the documentation on the 8601 aspect on Date.parse is a bit weak. It would
> be good to know what exactly it *is* expected to parse.

I retried and got it working (I think I assumed wrong that Date.parse returns a Date). Set the accepted diff to 30 minutes as suggested.

Comment 10

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Comment on attachment 526998 [details] [diff] [review]
implementation of exslt:date-time(), plus a minimal test case


>+            // http://exslt.org/date/functions/date-time/
>+            // format: YYYY-MM-DDTTHH:MM:SS.sss+00:00
>+            char formatstr[] = "%04hd-%02ld-%02ldT%02ld:%02ld:%02ld.%03ld%c%02ld:%02ld";
>+            size_t max = sizeof("YYYY-MM-DDTHH:MM:SS.sss+00:00");

Make this |const size_t max|.

>+  ok(now_ms - xslt_ms < accepted_diff, "generated timestamp should be not more than "
>+    + accepted_diff + " ms before 'now', but the difference was: " + (now_ms - xslt_ms));

This needs to be Math.abs(now_ms - xslt_ms), in both places.

r=me with those fixed.

If you attach a new patch (preferrably "hg export"ed) then I can land it.

Comment 11

[Julian Reschke]

(In reply to comment #10)
> Comment on attachment 526998 [details] [diff] [review]
> implementation of exslt:date-time(), plus a minimal test case
> 
> 
> >+            // http://exslt.org/date/functions/date-time/
> >+            // format: YYYY-MM-DDTTHH:MM:SS.sss+00:00
> >+            char formatstr[] = "%04hd-%02ld-%02ldT%02ld:%02ld:%02ld.%03ld%c%02ld:%02ld";
> >+            size_t max = sizeof("YYYY-MM-DDTHH:MM:SS.sss+00:00");
> 
> Make this |const size_t max|.

Will do.

> >+  ok(now_ms - xslt_ms < accepted_diff, "generated timestamp should be not more than "
> >+    + accepted_diff + " ms before 'now', but the difference was: " + (now_ms - xslt_ms));
> 
> This needs to be Math.abs(now_ms - xslt_ms), in both places.

Really? isn't now_ms guaranteed to be >= xslt_ms? Anyway; I'll make that change...

Comment 12

[Julian Reschke]

Attachment: implementation of exslt:date-time(), plus a minimal test case

Comment 13

[Julian Reschke]

Any chance to get this in before FF6 Aurora is branched?

Comment 14

[Shane Bundy]

http://hg.mozilla.org/mozilla-central/rev/0abb7330c64c

Comment 15

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Yup, checked in. Leaving the honors of marking it fixed to Julian.

(btw, the patch here contained windows lineendings which made it not apply. Easy fix though).

Comment 16

[Julian Reschke]

(In reply to comment #15)
> Yup, checked in. Leaving the honors of marking it fixed to Julian.

Thanks. I build the Nightly from m-c and tested against

  http://greenbytes.de/tech/webdav/xslt-test.xml

and all is well.

> (btw, the patch here contained windows lineendings which made it not apply.
> Easy fix though).

Sorry for that.