all devices regression report - 2017-11-13-02:46:39 - 5fe97ce30de048027da19a85bc574a8adfcf2116
| Newsgroups | gmane.comp.printing.ghostscript.regression |
|---|---|
| Message-ID | <20171113104640.8858820401E6@i7> |
Differences in all devices errors using gs/examples files Previous Revision: db2415d2b117946e5ec1d46936a134940f99c21d Current Revision: 5fe97ce30de048027da19a85bc574a8adfcf2116 commit 5fe97ce30de048027da19a85bc574a8adfcf2116 Author: Ken Sharp <[email protected]> AuthorDate: Mon Nov 13 09:07:03 2017 +0000 CommitDate: Mon Nov 13 09:07:03 2017 +0000 pdfwrite - fix an error handling problem Bug 698735 "Ghostscript: SEGV in pdfwrite with -Z@" This was, unfortunately, caused by the shift to cleaning up memory instead of relying on the garbage collector. When an error occurred writing the metadata, the code simply exited back to the caller, it did not close the temporary stream it was writing the metadata to. Due to the way that 'asides' are handled this led to the wrong stream being closed, and a pointer retained to a closed stream. The result was the seg fault. There may well be other conditions like this still present. devices/vector/gdevpdfe.c new errors: errors that went away: errors that are not whitelisted: === ghostpdl/examples/transparency_example.ps === cp50 === === ghostpdl/examples/transparency_example.ps === psdrgb ===