Starting with 5403b7384b203c6b462f8ed3190b7962c929769c a couple of the nightly regression tests fail with a /VMerror. Here is an example command line: bin/gs -o test.pxl -sDEVICE=pxlcolor -r300 ./Bug692688.pdf The other file that shows the same error is Bug692933.pdf
Both of those give me VMerrors with commit prior to the one mentioned - as a result of trying to allocate more than 2Gb in the transparency code. Assigning back to Marcos for further information.......
I've double checked and there is definitely a /Vmerror in 5403b7384b203c6b462f8ed3190b7962c929769c and not in 693c516ad22bb79d7c8c85329ff6a53dc82505fc, however I agree that allocating 2 gigs of memory is not reasonable so presumably the change just exposed the problem. Assigning to Michael since he is the transparency go to person.
This should be resolved by my pending change to use a clist when the pdf14 device would consume too much memory. I suggest that this be made a duplicate of bug 689805. If you agree, Marcos or Michael, please go ahead and resolve this as a duplicate.
*** This bug has been marked as a duplicate of bug 689805 ***