[bug #68484] pkg install <pkg> errors on missing `"` in case of packages using configure-make

--===============6370270020553669788==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: base64
Content-Disposition: inline

UGxlYXNlIHVzZSB0aGUgYnVnIHRyYWNrZXIgdG8gcG9zdCB1cGRhdGVzIHRvIGEgYnVnIHJlcG9y
dC4gIFRoZSBtYWlsaW5nIGxpc3QgaXMgaW50ZW5kZWQgYXMgYSByZWFkLW9ubHkgbm90aWZpY2F0
aW9uIHN0cmVhbS4gIEluZm8gcG9zdGVkIHRvIHRoaXMgbWFpbGluZyBsaXN0IGFkZHJlc3Mgd29u
J3QgYXBwZWFyIGluIHRoZSB0cmFja2VyIGRhdGFiYXNlIHdoZXJlIGl0IGlzIG1vc3QgdXNlZnVs
LgoK

--===============6370270020553669788==
Content-Type: MULTIPART/signed; micalg=pgp-sha256; protocol="application/pgp-signature"; BOUNDARY="8323329-1189641421-1782827003=:546562"
Content-Disposition: inline

--8323329-1189641421-1782827003=:546562
Content-Type: TEXT/plain; CHARSET=utf-8
Content-Transfer-Encoding: QUOTED-PRINTABLE
Content-Disposition: inline

Follow-up Comment #9, bug #68484 (group octave):=0A=0AI imported the cset f=
rom comment #5, built Octave-12 on Linux (0 FAILs),=0Abacked out the last c=
set of mxe-octave as suggested by Markus in comment #6,=0Aand let mxe-octav=
e do its thing.=0A=0AAfter crossbuilding Octave-12 and testing on Windows 1=
0:=0A=0AI tested a few packages and most of them gave 0 (zero) FAILs.  The =
signal pkg=0Acrashes Octave but that=0As a known issue (bug #66990).=0A=0AT=
he windows package is still mis-compiled in mxe-octave.  'pkg test windows'=
=0Acrashes Octave-12 at the very first code line of the first test in=0A__C=
OM__.cc-tst .=0A=0ATrying to install windows-1.7.0 gets beyond the configur=
e-make stage, so that=0Alooks better now, but afterwards things seem to be =
mixed up:=0A=0Aoctave:9> pkg install -verbose c:/users/philip/downloads/win=
dows-1.7.0.tar.gz=0A:=0A:=0Achecking   octave_kbhit or octave::kbhit...   o=
ctave_kbhit=0A:=0A:=0Agrab.cc:155:16: note: in expansion of macro 'OCTAVE__=
KBHIT'=0A  155 |           ch =3D OCTAVE__KBHIT( 0 );=0A      |            =
    ^~~~~~~~~~~~~=0Aconfig.h:19:23: error: 'octave_kbhit' was not declared =
in this scope; did you=0Amean 'octave_quit'?=0A   19 | #define OCTAVE__KBHI=
T octave_kbhit=0A      |                       ^~~~~~~~~~~~=0A:=0A=0Abut ma=
ybe that is a windows package issue?=0A=0AInstalling control package locall=
y: =0AAlso gets beyond configure-make stage, compiles w/o a hitch, and pass=
es 383=0Atests w. 10 XFAILs; same as the cross-compiled package =3D=3D>  OK=
.=0A=0AI guess the issue reported in this bug report has been fixed; but wh=
y the=0Awindows package gets mis-compiled in mxe-octave while before the co=
debase=0Areshuffle it got built largely correctly still feels fishy.=0AI'll=
 check the mxe-octave build log for the of-windows package.=0A=0A=0A    ___=
____________________________________________________=0A=0AReply to this ite=
m at:=0A=0A  <https://savannah.gnu.org/bugs/?68484>=0A=0A__________________=
_____________________________=0AMessage sent via Savannah=0Ahttps://savanna=
h.gnu.org/=0A
--8323329-1189641421-1782827003=:546562
Content-Type: APPLICATION/pgp-signature; name=signature.asc

-----BEGIN PGP SIGNATURE-----

iHQEABYIAB0WIQQk97aszIMMAvLLwm6qLAuaBUf3TgUCakPH+wAKCRCqLAuaBUf3
Tv+RAPd+c6Q7WE8EPkot9BU5qFg6JQ1q7RCztaRSNqrpk17NAQDF7trG0dVSy1DZ
S00chiWVYNnPW+fBwIVciW/2vNYjBA==
=/323
-----END PGP SIGNATURE-----

--8323329-1189641421-1782827003=:546562--



--===============6370270020553669788==--