PR#1: Use RPMTAG_NOT_FOUND macro instead of -1 (for RPM 4.19) - rpms/rpminspect

rpms / rpminspect

#1 Use RPMTAG_NOT_FOUND macro instead of -1 (for RPM 4.19)

Merged a year ago by dcantrell. Opened a year ago by mdomonko.

rpms/ mdomonko/rpminspect rawhide into rawhide

Use RPMTAG_NOT_FOUND macro instead of -1 (for RPM 4.19)

Michal Domonkos • a year ago

ffa1692

rpminspect-1.11-use-RPMTAG_NOT_FOUND-macro.patch

file added

+31

		`@@ -0,0 +1,31 @@`
		`+ commit 0ce349a220025a6f2bbdeb111dab2bcd88e1c374`
		`+ Author: Michal Domonkos <mdomonko@redhat.com>`
		`+ Date: Thu May 18 15:58:53 2023 +0200`
		`+`
		`+ [fix] Use RPMTAG_NOT_FOUND macro instead of -1`
		`+`
		`+ In RPM 4.19, the internal representation of the rpmTagVal type has been`
		`+ changed from signed to unsigned. That means, a plain check for the`
		`+ value of -1 (tag not found) returned from rpmTagGetValue() now produces`
		`+ a compiler warning.`
		`+`
		`+ In order to fix that, we have to either explicitly cast the literal -1`
		`+ to (uint32_t)-1 or just use the RPMTAG_NOT_FOUND macro readily provided`
		`+ by RPM.`
		`+`
		`+ Choose the latter since that's the supported way, really, it's also now`
		`+ documented as such in the API docs for the rpmTagGetValue() function.`
		`+`
		`+ diff --git a/lib/inspect_modularity.c b/lib/inspect_modularity.c`
		`+ index 8b49421..3c96b1a 100644`
		`+ --- a/lib/inspect_modularity.c`
		`+ +++ b/lib/inspect_modularity.c`
		`+ @@ -181,7 +181,7 @@ static bool modularity_driver(struct rpminspect ri, rpmfile_entry_t file)`
		`+`
		`+ /* Find how to find the header */`
		`+ tv = rpmTagGetValue("modularitylabel");`
		`+ - if (tv == -1) {`
		`+ + if (tv == RPMTAG_NOT_FOUND) {`
		`+ add_result(ri, &params);`
		`+ free(params.msg);`
		`+ return false;`

rpminspect.spec

file modified

+6 -2

		`@@ -1,6 +1,6 @@`
		`Name: rpminspect`
		`Version: 1.11`
		`- Release: 1%{?dist}`
		`+ Release: 2%{?dist}`
		`Summary: Build deviation compliance tool`
		`Group: Development/Tools`
		`# librpminspect is licensed under the LGPLv3+, but 5 source files in`
		`@@ -30,6 +30,7 @@`
		`Source0: https://github.com/rpminspect/rpminspect/releases/download/v%{version}/%{name}-%{version}.tar.xz`
		`Source1: https://github.com/rpminspect/rpminspect/releases/download/v%{version}/%{name}-%{version}.tar.xz.asc`
		`Source2: gpgkey-62977BB9C841B965.gpg`
		`+ Patch0: rpminspect-1.11-use-RPMTAG_NOT_FOUND-macro.patch`
		`Requires: librpminspect%{?_isa} = %{version}-%{release}`

		`BuildRequires: meson`
		`@@ -160,7 +161,7 @@`

		`%prep`
		`%{gpgverify} --keyring='%{SOURCE2}' --signature='%{SOURCE1}' --data='%{SOURCE0}'`
		`- %autosetup`
		`+ %autosetup -p1`


		`%build`
		`@@ -197,6 +198,9 @@`


		`%changelog`
		`+ * Mon May 22 2023 Michal Domonkos <mdomonko@redhat.com> - 1.11-2`
		`+ - Use RPMTAG_NOT_FOUND macro instead of -1 (for RPM 4.19)`
		`+`
		`* Fri Mar 03 2023 David Cantrell <dcantrell@redhat.com> - 1.11-1`
		`- Upgrade to rpminspect-1.11`

no initial comment

mdomonko commented a year ago

I recognize that the release model here is not really patch-based so please let me know if you'd prefer to create a new tarball instead.

dcantrell commented a year ago

This is fine for now. I have Copr builds which are regenerated on each merge upstream and those builds are used in production by the gating services. Periodically I do a new release and update the packages in Fedora, but I don't have a strict schedule around when I do a new versioned release. So I will merge this PR for now and then on the next x.y release I do, this patch can go away in the SRPM.