From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail1.windriver.com (mail1.windriver.com [147.11.146.13]) by mx.groups.io with SMTP id smtpd.web11.4327.1585860213767087191 for ; Thu, 02 Apr 2020 13:43:34 -0700 Authentication-Results: mx.groups.io; dkim=missing; spf=pass (domain: windriver.com, ip: 147.11.146.13, mailfrom: matthew.zeng@windriver.com) Received: from ALA-HCA.corp.ad.wrs.com (ala-hca.corp.ad.wrs.com [147.11.189.40]) by mail1.windriver.com (8.15.2/8.15.2) with ESMTPS id 032KhVMZ006915 (version=TLSv1 cipher=AES256-SHA bits=256 verify=FAIL); Thu, 2 Apr 2020 13:43:31 -0700 (PDT) Received: from mt-manjaro (128.224.21.163) by ALA-HCA.corp.ad.wrs.com (147.11.189.50) with Microsoft SMTP Server (TLS) id 14.3.487.0; Thu, 2 Apr 2020 13:43:30 -0700 References: <87o8sbf8e4.fsf@windriver.com> <871rp6t1jg.fsf@windriver.com> <21293c7f-669d-8153-2104-f9de823a7362@windriver.com> User-agent: mu4e 1.2.0; emacs 28.0.50 From: "Matthew" To: Randy MacLeod CC: Andreas =?utf-8?Q?M=C3=BCller?= , Subject: Re: Regarding poppler auto PACKAGECONFIG when qt5-layer exists In-Reply-To: Date: Thu, 2 Apr 2020 16:43:28 -0400 Message-ID: <87mu7tehzj.fsf@windriver.com> MIME-Version: 1.0 X-Originating-IP: [128.224.21.163] Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Andreas M=C3=BCller writes: > On Thu, Apr 2, 2020 at 7:27 PM Randy MacLeod > wrote: >> >> On 2020-04-02 10:17 a.m., Mingde (Matthew) Zeng wrote: >> > >> > Andreas M=C3=BCller writes: >> > >> >> On Wed, Apr 1, 2020 at 7:02 PM Mingde (Matthew) Zeng >> >> wrote: >> >>> >> >>> Hello, >> >>> >> >>> In meta-oe/recipes-support/poppler/poppler_0.85.0.bb line 17, there = is the line >> >>> >> >>> PACKAGECONFIG ??=3D "jpeg openjpeg png tiff nss ${@bb.utils.con= tains('BBFILE_COLLECTIONS', 'qt5-layer', 'qt5', '', d)}" >> >>> >> >>> Let's say I add meta-qt5 to bblayers.conf, but I don't plan to use i= t immediately in a build. Therefore I haven't make any additional change to= local.conf. >> >>> >> >>> However I will be receiving the following error when starting a buil= d: >> >>> >> >>> WARNING: qtbase is not whitelisted, figuring out PNWHITELIST... >> >>> ERROR: Nothing PROVIDES 'qtbase' (but .../meta-openembedded/met= a-oe/recipes-support/poppler/poppler_0.85.0.bb DEPENDS on or otherwise requ= ires it) >> >> Guess I know how this is going to end so just out of curiosity: What >> >> configuration is causing 'qtbase is not whitelisted' ? >> > >> > This is the problem. I don't have any configuration in local.conf rela= ted to qt5, because I don't plan to use it in my build. It is only in bblay= ers.conf so I can use qt5 layer when I feel necessary. >> >> Mathew is building xfce using the whitelist bbclass that some people >> use. When he added meta-qt5 and whitelisted a few recipes, the >> build broke. I don't object to adding changing how poppler works >> when qt is present but by default as Matthew said, adding a layer >> should not change anything. On the other hand, I can understand how >> it would be very convenient to have the dependencies and configuration >> change automatically. Could the be a tunable that is off by default? >> The obvious downside to such an approach is testing but the >> PACKAGECONFIGs already suffer from lack of testing so this only makes >> it incrementally less tested! ;-) >> >> >> There are only two such dynamic package configurarations in meta-oe: >> $ grep -r BBFILE_COLLECTIONS | grep "PACKAGECONFIG " >> meta-oe/recipes-support/poppler/poppler_0.85.0.bb:PACKAGECONFIG ??=3D >> "jpeg openjpeg png tiff nss ${@bb.utils.contains('BBFILE_COLLECTIONS', >> 'qt5-layer', 'qt5', '', d)}" >> meta-networking/recipes-support/wireshark/wireshark_3.2.2.bb:PACKAGECONF= IG >> ?=3D "libpcap gnutls libnl libcap sbc >> ${@bb.utils.contains('BBFILE_COLLECTIONS', 'qt5-layer', 'qt5 plugins', >> '', d)}" >> >> >> and the only other such usage that I have in the meta-foo layers in my >> local collection is: >> meta-security/meta-tpm/recipes-tpm/swtpm/swtpm_0.2.0.bb:PACKAGECONFIG += =3D >> "${@bb.utils.contains('BBFILE_COLLECTIONS', 'filesystems-layer', 'cuse', >> '', d)}" >> >> so fixing this isn't a huge impact. >> >> Matthew, >> If Andreas agrees, please send a patch removing this behaviour >> unless someone replies this week, providing a mechanism and a >> local.conf setting to make it configurable. Sounds good, let's see if there is an alternative mechanism. >> >> ../Randy > Ahh - thanks for explanation. Send out whatever is 'correct' I won't > complain and find a way to not break the qt-based pdf readers / > poppler consumers. > > Andreas -- Mingde (Matthew) Zeng