TRA6SG@cms1.ucs.leeds.ac.uk (04/25/91)
I have compiled 3.0b on a SGI 4D/210VGX under IRIX 3.3.1 with cfront 2.0 . Everything seemed to going very well until I started using Doc for a short report. I am having problems with editing section headings, titles etc. The first sign of trouble was not always being able to insert characters at the begining or end of the heading, similarly if the whole heading is deleted by backspace (say to overwrite <section> if you run the macro on a blank area) it becomes impossible to enter anything into the header at all. I know this is not ajust an interviews problem as I have seen this work correctly on a SUN 3/50. When I was shown the SUN version by a friend, he also showed me that you could have multi line section headings by just entering a return into the heading. I tried that on my system only to find that doing this immediately causes my X server to core dump with a bus error ! Any ideas out there ? One other thing on Doc, the new menu system appears to be about 2-3 times slower than the 2.6 variety, I thought the new system was glyph based and therefore more efficient. Does anybody else out there feel the new menus are a bit slow ? Stephen Gallimore. TRA6SG%cms1.leeds.ac.uk@nsfnet-relay.ac.uk
linton@marktwain.rad.sgi.com (Mark Linton) (04/30/91)
In article <25.Apr.91.14:15:05.BST.#9564@UK.AC.LEEDS.CMS1>, TRA6SG@cms1.ucs.leeds.ac.uk writes: |> I tried that on my system only to find that doing this |> |> immediately causes my X server to core dump with a bus error ! |> |> Any ideas out there ? Unfortunately, all I can suggest is that you try to get an early version of IRIX 4.0. That is what I am using. |> One other thing on Doc, the new menu system appears to be about 2-3 times |> slower than the 2.6 variety, I thought the new system was glyph based and |> therefore more efficient. Does anybody else out there feel the new menus |> are a bit slow ? 3.0-alpha menus were slower than necessary, but I would be surprised if beta menus were slow. Perhaps the problem is doing the bevels with polygons, as well as copying from a pixmap for "double-buffering". The simplest way to examine performance is to run xscope and trace the calls that are being made.