[Bug]: archaic hostbuild-step of gnushogi 1.4.2 throws errors

[robertkirkman]

Problem description

• gnushogi 1.4.2, which released on February 16, 2014, produces these errors during its build in termux-packages:

./pat2inc
make[1]: ./pat2inc: Text file busy
make[1]: *** [Makefile:224: pattern.inc] Error 127
make[1]: *** Waiting for unfinished jobs....

make[1]: Leaving directory '/home/builder/.termux-build/gnushogi/host-build/gnushogi'
make: [Makefile:46: gnushogi_compile] Error 2 (ignored)
gcc -g -O2 -DHASHFILE=\"/usr/local/lib/gnushogi/gnushogi.hsh\" -fsigned-char -funroll-loops -Wall -Wno-implicit-int -Wstrict-prototypes -ansi -pedantic -I. -I.. -I/home/builder/.termux-build/gnushogi/src -o gnushogi globals.o init-common.o pattern-common.o attacks.o book.o commondsp.o cursesdsp.o dspwrappers.o eval.o genmove.o init.o pattern.o rawdsp.o search.o tcontrl.o util.o main.o  -lcurses -lm
/usr/bin/ld: cursesdsp.o:/home/builder/.termux-build/gnushogi/src/gnushogi/cursesdsp.c:75: multiple definition of `DRAW'; globals.o:/home/builder/.termux-build/gnushogi/src/gnushogi/globals.c:200: first defined here
/usr/bin/ld: rawdsp.o:/home/builder/.termux-build/gnushogi/src/gnushogi/rawdsp.c:50: multiple definition of `mycnt2'; cursesdsp.o:/home/builder/.termux-build/gnushogi/src/gnushogi/cursesdsp.c:65: first defined here
/usr/bin/ld: rawdsp.o:/home/builder/.termux-build/gnushogi/src/gnushogi/rawdsp.c:50: multiple definition of `mycnt1'; cursesdsp.o:/home/builder/.termux-build/gnushogi/src/gnushogi/cursesdsp.c:65: first defined here
/usr/bin/ld: rawdsp.o:/home/builder/.termux-build/gnushogi/src/gnushogi/rawdsp.c:51: multiple definition of `DRAW'; globals.o:/home/builder/.termux-build/gnushogi/src/gnushogi/globals.c:200: first defined here
/usr/bin/ld: rawdsp.o:/home/builder/.termux-build/gnushogi/src/gnushogi/rawdsp.c:48: multiple definition of `MSCORE'; cursesdsp.o:/home/builder/.termux-build/gnushogi/src/gnushogi/cursesdsp.c:74: first defined here
/usr/bin/ld: rawdsp.o:/home/builder/.termux-build/gnushogi/src/gnushogi/rawdsp.c:47: multiple definition of `MV'; cursesdsp.o:/home/builder/.termux-build/gnushogi/src/gnushogi/cursesdsp.c:73: first defined here
collect2: error: ld returned 1 exit status
make[1]: Leaving directory '/home/builder/.termux-build/gnushogi/host-build/gnushogi'
make[1]: *** [Makefile:140: gnushogi] Error 1
make: [Makefile:55: bbk] Error 2 (ignored)

/bin/sh: 1: ./gnushogi: Exec format error
make[1]: *** [Makefile:234: gnushogi.bbk] Error 126
make[1]: Leaving directory '/home/builder/.termux-build/gnushogi/build/gnushogi'
make: [Makefile:55: bbk] Error 2 (ignored)
make[1]: warning:  Clock skew detected.  Your build may be incomplete.
make[1]: Leaving directory '/home/builder/.termux-build/gnushogi/build/gnushogi'
cd ./gnushogi && make install
make[1]: Warning: File 'pat2inc' has modification time 3599 s in the future
make[1]: Entering directory '/home/builder/.termux-build/gnushogi/build/gnushogi'
./pat2inc

/usr/bin/install: cannot stat 'gnushogi.bbk': No such file or directory
make[1]: *** [Makefile:252: install] Error 1
make[1]: Leaving directory '/home/builder/.termux-build/gnushogi/build/gnushogi'
make: [Makefile:72: gnushogi_install] Error 2 (ignored)

Sometimes, I have also seen this error:

cp: cannot stat '/home/builder/.termux-build/gnushogi/host-build/gnushogi/pat2inc': No such file or directory

• However, I couldn't help but notice that, after February 2014, upstream development continued on the master branch for almost a year before the last commit in the upstream gnushogi repository on November 19, 2014 - including cross-compilation-focused changes.
• I was unable to successfully cherry-pick the cross-compilation-focused commits without pulling all other commits in the whole master branch as dependencies, so I tested a snapshot tarball of gnushogi "1.5pre" dated to November 2014. For me, that causes all the errors listed above to disappear.
• The upstream developer is active on GitHub in 2024. @ydirson, it has been more than 10 years since the last release of gnushogi. Will you make another release that includes improvements to cross-compilation support?

What steps will reproduce the bug?

1. scripts/run-docker.sh ./build-package.sh gnushogi
2. look for errors (return code may be 0, but under certain circumstances, the errors will interrupt the build. The errors are visible in the log under all circumstances.)

What is the expected behavior?

No visible errors when running scripts/run-docker.sh ./build-package.sh gnushogi

System information

N/A

[robertkirkman]

A couple of notes:

• i suspect that ydirson might be capable of releasing a tarball that could be cross-compiled without any hostbuild-step, simultaneously cleaning up both cross-compilation mode and non-cross-compilation mode. If ydirson is available for comment, they might be able to confirm and/or further clarify that.
• ydirson places gnushogi version "1.5~git20140725-2.1" in official Debian. The code of that version comes from a repository where development has continued seemingly without "GNU official" releases for 10 years. That gives me hope that there could be a high chance of ydirson giving permission to package an unreleased branch.