Mailing-List: contact notifications-help@freemarker.incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@freemarker.incubator.apache.org
Date: Tue, 10 Oct 2017 16:18:00 +0000 (UTC)
From: "Daniel Dekany (JIRA)" <jira@apache.org>
To: notifications@freemarker.incubator.apache.org
Message-ID: <JIRA.13106196.1506742750000.19131.1507652280616@Atlassian.JIRA>
In-Reply-To: <JIRA.13106196.1506742750000@Atlassian.JIRA>
References: <JIRA.13106196.1506742750000@Atlassian.JIRA> <JIRA.13106196.1506742750715@jira-lw-us.apache.org>
Subject: [jira] [Comment Edited] (FREEMARKER-77) expr!! as a shorthand for
 (expr)!
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Tue, 10 Oct 2017 16:18:06 -0000


    [ https://issues.apache.org/jira/browse/FREEMARKER-77?page=3Dcom.atlass=
ian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D1=
6198539#comment-16198539 ]=20

Daniel Dekany edited comment on FREEMARKER-77 at 10/10/17 4:17 PM:
-------------------------------------------------------------------

Or rather, first I want a concrete working solution in FM3, and then see ho=
w to port ideas back with causing the least confusion. Though for many of t=
he changes back porting just wont be feasible...

The problem with {{(...)!default}} in FM2 is that parentheses already have =
a commonly understood meaning. In particular, if the evaluation order with =
and without the parenthesis is the same, like in the case of {{(a * b) + c}=
}, t can be removed. {{(...)!default}} works differently, which is confusin=
g. Can you show me an example where you suspect the FM3 approach won't work=
 well?

(Though you don't really know how FM3 meant to do it. The relevant discussi=
on is [here in the mail archive|https://lists.apache.org/thread.html/2715ec=
c7a47fbacf18996169dfd4ed4b82bc38c567f114cec86e0db1@<dev.freemarker.apache.o=
rg>], but it's admittedly very long and not easy to follow.)


was (Author: ddekany):
Or rather, first I want a concrete working solution in FM3, and then see ho=
w to port ideas back with causing the least confusion. Though for may of th=
e changes back porting just wont be feasible...

The problem with {{(...)!default}} in FM2 is that parentheses already have =
a commonly understood meaning. In particular, if the evaluation order with =
and without the parenthesis is the same, like in the case of {{(a * b) + c}=
}, t can be removed. {{(...)!default}} works differently, which is confusin=
g. Can you show me an example where you suspect the FM3 approach won't work=
 well?

(Though you don't really know how FM3 meant to do it. The relevant discussi=
on is [here in the mail archive|https://lists.apache.org/thread.html/2715ec=
c7a47fbacf18996169dfd4ed4b82bc38c567f114cec86e0db1@<dev.freemarker.apache.o=
rg>], but it's admittedly very long and not easy to follow.)

> expr!! as a shorthand for (expr)!
> ---------------------------------
>
>                 Key: FREEMARKER-77
>                 URL: https://issues.apache.org/jira/browse/FREEMARKER-77
>             Project: Apache Freemarker
>          Issue Type: New Feature
>            Reporter: Ondra =C5=BDi=C5=BEka
>            Priority: Minor
>
> a.very.long.expr!! could be a shorthand for {{(a.very.long.expr)!}}.
> The effect of !! would go back to the "root" of the expression, i.e. at t=
he start of the part ending with !! . Not sure if that would have clear eno=
ugh rules, though.
> The use case is, that often there is a long expression, (again, getting s=
omething from a deep graph), and anywhere on the path something can be miss=
ing. The only relevant thing is at the end. And the parentheses make the ex=
pression less readable.
> Example made-up situatuon:
> {code}
> (foo.bar.baz[(foo.transform((a.very.long.expression())!"defaultResult"))!=
fallbackKey(context)])!"-"
> {code}
> 1) The long expression can be missing at any point, so the transform woul=
d get some default value.
> 2) The transform may return null if the argument is somehow improper. For=
 that case, the {{fallbackKey()}} would return a key as per the {{context}}=
.
> 3) If the map doesn't contain the entry or any of {{foo.bar.baz}} is miss=
ing, the whole expression would turn into {{"-"}}.
> {code}
> foo.bar.baz[foo.transform(a.very.long.expression()!!"defaultResult")!!fal=
lbackKey(context)]!!"-"
> {code}
> Ok, this one is perhaps a candidate for splitting into few steps. But it'=
s just for illustration.
> Does it make sense? Is it doable?
> Marking it as minor as it's not really too much pain, just would be nice =
to have.


--
This message was sent by Atlassian JIRA
(v6.4.14#64029)