From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Sat, 11 Jan 2020 17:08:45 +0100 Subject: [Buildroot] [PATCH 1/1] package/poppler: really disable test programs In-Reply-To: References: <20200111115054.864819-1-fontaine.fabrice@gmail.com> <20200111153537.66974300@windsurf> Message-ID: <20200111170845.1239c29a@windsurf> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net On Sat, 11 Jan 2020 15:44:44 +0100 Fabrice Fontaine wrote: > > What is the problem with C++ executables, since poppler depends on C++ > > anyway ? > > > > Of course, it is good to disable building tests if possible, but I'm > > just curious to understand the motivation for the change: did you > > encounter a build failure? Or it is just that you realized the tests > > were being built while they should not? > Following your comment on the leveldb patch that was fixing the > -latomic build failure, I'm going through the different -latomic > workarounds used in the cmake packages. > I'm trying to replace these workarounds by upstreamable patches. > However, in some cases such as libraries (like poppler) these > workarounds/patches are not strictly needed if we don't build any > executable. It's not true: if libpoppler.so uses some symbols that are defined in libatomic.so, then it should link against libatomic.so (i.e libpoppler.so should have a DT_NEEDED entry for libatomic.so). Otherwise any application linking against libpoppler.so will have to know that it should also link against libatomic.so, which is not good. Thomas -- Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com