Mailing-List: contact notifications-help@freemarker.incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@freemarker.incubator.apache.org
Date: Fri, 29 Sep 2017 07:30:00 +0000 (UTC)
From: "Daniel Dekany (JIRA)" <jira@apache.org>
To: notifications@freemarker.incubator.apache.org
Message-ID: <JIRA.13105900.1506639121000.234089.1506670200483@Atlassian.JIRA>
In-Reply-To: <JIRA.13105900.1506639121000@Atlassian.JIRA>
References: <JIRA.13105900.1506639121000@Atlassian.JIRA> <JIRA.13105900.1506639121800@jira-lw-us.apache.org>
Subject: [jira] [Commented] (FREEMARKER-76) New built-in ?if(condition,
 otherwise) to override left-hand expr. conditionally
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Fri, 29 Sep 2017 07:30:05 -0000


    [ https://issues.apache.org/jira/browse/FREEMARKER-76?page=3Dcom.atlass=
ian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D1=
6185453#comment-16185453 ]=20

Daniel Dekany commented on FREEMARKER-76:
-----------------------------------------

I assume {{expensiveCall()?as("cached")!false}} really meant to be {{expens=
iveCall()!false?as("cached")}}, so the {{?if(cache)}} will work. Anyway, th=
e use case seems to be a variation of this pattern:

{{code}}
<#if someExpensiveExpression??>
   Print something around ${someExpensiveExpression}.
</#if>
{{code}}

which could be addressed as:

{{code}}
<#if someExpensiveExpression?as(x)??>
   Print something around ${x}.
</#if>
{{code}}

It's longer than your version, but your is very strange because, apparently=
, it tries to print something inside a string literal only so that it can u=
se the {{?something}} syntax. That's problematic, among others because of h=
ow auto-escaping works.

There's an issue with my above application of {{?as}} as well though... bec=
ause of the fail-fast behavior of FM2 with {{null}}-s, it's quite unusual t=
hat {{?as}} can continue even if its left side is {{null}}, and much worse,=
 in that case all I could implement is returning with {{null}}, and not doi=
ng the assignment. Because storing {{null}} in a variable just would make i=
t transparent, that is, reading it would cause a fallback to a higher scope=
... (Yip, there are reasons why FM3 was started.) Though in the above issue=
 it doesn't mater, because if it was {{null}}, then {{??}} will be {{false}=
}, and so you don't enter the {{if}} block where {{x}} would be in scope.

> New built-in ?if(condition, otherwise) to override left-hand expr. condit=
ionally
> -------------------------------------------------------------------------=
-------
>
>                 Key: FREEMARKER-76
>                 URL: https://issues.apache.org/jira/browse/FREEMARKER-76
>             Project: Apache Freemarker
>          Issue Type: New Feature
>          Components: engine
>    Affects Versions: 2.3.26-incubating
>            Reporter: Ondra =C5=BDi=C5=BEka
>
> I propose this built-in:
> {code}
> expr?if(condition, otherwise)
> {code}
> This would check condition, and if true, it would pass along the {{expr}}
> If condition was false, it would pass along the {{otherwise}}.
> The diffrence from {{condition?then(expr, otherwise)}} is that {{expr}} w=
ould be evaluated even if  {{condition}} was false. The otherwise could be =
optional and default to "".
> For instance,
> {code}
> ${ "Hi, ${user}!"?if(user??) }
> {code}
> The true benefit of {{?if}} would be when used with {{as?}}, proposed in =
FREEMARKER-75. =20
> {code}
> ${ "Hi, ${expensiveCall()?as("cached")!false}, again!"?if(cached) }
> {code}
> Without these goodies, this would look like this:
> {code}
> <#assign cached =3D expensiveCall() >
> <#if cached?has_content >
>     Hi, ${ cached }, again!
> </#if>
> {code}
> I hope I didn't overlook some other obvious solution :)


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