Dear David, > On Thu, Nov 22, 2007 at 03:26:31PM +0700, Martin Hosken wrote: > >> Dear srezic, >> >>> t/fret.......... >>> # Failed test at t/fret.t line 8. >>> # Looks like you failed 1 test of 1. >>> > When I've been doing PostScriptish stuff and converting to other > formats, then minor differences in Ghostscript from one machine to the > next were enough to "break" my output files, despite them appearing to > be identical. Is this catching you out here? >
No, since I am generating the PDF entirely in Perl. I think what is happening is that I have been outputting floating point numbers into strings without using sprintf and so there are slight rounding differences on different architectures/compilers. I thought I had nailed all the occurrences, but find I hadn't. So I've tried again and there is a 0.10.1 in the wild that the test engines should pick up. Failing that, I am going to need to ask for copies of the failed temporary file t/testfont.pdf so that I can see where the differences are coming from. The test only deletes the temporary file if the test passes. Alternatively I could just pull the test ;) Thanks, Yours, Martin
