svn commit: r493407 - head/graphics/xpdf4

[Kubilay Kocak]

On 20/02/2019 5:52 pm, Tobias Kortkamp wrote:
> On Wed, Feb 20, 2019 at 05:37:58PM +1100, Kubilay Kocak wrote:
>> On 20/02/2019 5:34 pm, Tobias Kortkamp wrote:
>>> On Wed, Feb 20, 2019 at 06:14:57AM +0000, Cy Schubert wrote:
>>>> Author: cy
>>>> Date: Wed Feb 20 06:14:57 2019
>>>> New Revision: 493407
>>>> URL: https://svnweb.freebsd.org/changeset/ports/493407
>>>>
>>>> Log:
>>>>     Remove whitespace in USE_QT, otherwise qmake_build is ignored.
>>>>     
>>>>     Reported by:	Paul Beard <paulbeard at gmail.com>
>>>>
>>>> Modified:
>>>>     head/graphics/xpdf4/Makefile
>>>>
>>>> Modified: head/graphics/xpdf4/Makefile
>>>> ==============================================================================
>>>> --- head/graphics/xpdf4/Makefile	Wed Feb 20 04:50:53 2019	(r493406)
>>>> +++ head/graphics/xpdf4/Makefile	Wed Feb 20 06:14:57 2019	(r493407)
>>>> @@ -27,8 +27,7 @@ SUB_FILES=	xpdf-man.conf
>>>>    
>>>>    USES=		compiler:c++11-lang cmake localbase:ldflags \
>>>>    		desktop-file-utils qt:5
>>>> -USE_QT=		buildtools_build,concurrent,core,gui,network,printsupport,\
>>>> -		qmake_build,svg,widgets
>>>> +USE_QT=		buildtools_build,concurrent,core,gui,network,printsupport,qmake_build,svg,widgets
>>>
>>> Next step: Replace the , with spaces to actually add Qt dependencies
>>> to the port.  Check with build-depends-list, run-depends-list,
>>> stage-qa.  Currently the port has no Qt 5 dependencies and it fails
>>> to build in Poudriere:
>>>
>>> ===> Checking for items in pkg-plist which are not in STAGEDIR
>>> Error: Missing: libexec/xpdf/xpdf
>>> Error: Missing: %%DATADIR%%/man/man1/xpdf.1.gz
>>> ===> Error: Plist issues found.
>>>
>>
>> If USE_* values are never comma-separated (always space separated),
>> sounds like a good portlint check to have.
> 
> Wouldn't it be much nicer if the framework would consistently check
> them on its own and refuse to proceed with the build if there are
> unknown components in a USE_*?

* > null, i just suggested the first thing that came to mind :)

>> If we have some comma-separated support framework bits (I didn't check),
>> maybe we should standardize on not doing so
> 
> USES=qt already complains about components that it does not recognize.
> I'm surprised that it accepts a comma-separated list silently.
>