> -----Original Message-----
> From: ASSI
> Sent: Tuesday, May 18, 2021 12:37 AM
>
> Jason Pyeron writes:
> > What is the historic rationale behind the "OR true" after the make
> > check?
>
> Not historic for the most part, I'd say. Cygport can also do
> cross-builds of packages and in those cases the tests will seldomly work
> (at all or at least partly) unless upstream walked the extra mile.
> Also, due to ATWIL syndrome and other factors, Cygwin is often not
Which is why I want the tests to fail - so it causes patches to be created...
and pushed upstream.
> explicitly considered a target platform or (even then) treated wrongly
> in different ways, so even when building natively you will encounter
> your fair share of spurious test fails.
>
> > It seems silly to have to redefine src_test() as {
> > cd ${B}
> > make check
> > }, just to have a failure exit code if the test fail.
>
> It would be equally silly the other way around in a different set of
> circumstances. But yes, making this configurable might be useful, but
> this should then be done in the local configuration, not in the cygport
> file.
If I put it in ~/.config/cygport.conf it will impact all packages, not just the
"one". The only place it is limited to applicable package is in the
package.cygport file, or am I missing something?
e.g.
$ cat pdfgrep.cygport
NAME="pdfgrep"
VERSION=2.1.2
RELEASE=1
CATEGORY="Text"
SUMMARY="Command-line utility for searching text in PDFs"
DESCRIPTION="Pdfgrep is a tool to search text in PDF files. It works much
like grep, with one distinction: it operates on pages and not on lines."
HOMEPAGE="https://pdfgrep.org/";
SRC_URI="https://pdfgrep.org/download/pdfgrep-${VERSION}.tar.gz";
# git format-patch --stdout v2.1.2..cygwin-2.1.2 > v2.1.2..cygwin-2.1.2.patch
PATCH_URI="v2.1.2..cygwin-2.1.2.patch"
BUILD_REQUIRES="asciidoc gcc-g++ libpoppler-cpp-devel libgcrypt-devel
libpcre-devel dejagnu texlive-collection-latex"
src_test() {
cd ${B}
make check
}
