Hi Paul! On So, 08 Sep 2019, Paul Gevers wrote:
> pass fail > ghostscript from testing 9.28~~rc2~dfsg-1 > fig2dev from testing 1:3.2.7a-7 > versioned deps [0] from testing from unstable > all others from testing from testing > > I copied some of the output at the bottom of this report. (The output is > rather dense. I would love a more verbose output, so that others like me > have something to look at.) > 28: compare arrow tips with template FAILED (output.at:41) The problem here is the following command in the fig2dev test suite: gs -sDEVICE=pgmraw -dEPSCrop -r1200 -dNOPAUSE -dBATCH -dQUIET -sOutputFile=a.pgm ../.././data/arrows.eps (called in fig2dev/tests/testsuite.dir/28) With gs from testing this generates a.pgm, while with the new gs 9.28~~rc2 it fails with the following message: (sid)root@dinghy:/tmp/autopkgtest.QBY3nQ/build.pm5/src/fig2dev/tests/testsuite.dir/28# gs -sDEVICE=pgmraw -dEPSCrop -r1200 -dNOPAUSE -dBATCH -dQUIET -sOutputFile=a.pgm ../.././data/arrows.eps Error: /undefinedfilename in (../.././data/arrows.eps) Operand stack: Execution stack: %interp_exit .runexec2 --nostringval-- --nostringval-- --nostringval-- 2 %stopped_push --nostringval-- --nostringval-- --nostringval-- false 1 %stopped_push Dictionary stack: --dict:731/1123(ro)(G)-- --dict:0/20(G)-- --dict:75/200(L)-- Current allocation mode is local Last OS error: Permission denied GPL Ghostscript RELEASE CANDIDATE 2 9.28: Unrecoverable error, exit code 1 Removing the "./" from the input filename the error disappears. Sadly the testsuite generates a filename with /./ in it... From my point of view ../.././data/arrows.eps is a valid file name on Unix systems, so gs should accept it. Is there a good reason to reject this file name in gs? Then we have to change the testsuite, otherwise I'd suggest fixing the behavior of gs. Greetings Roland
signature.asc
Description: PGP signature
