home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: gnu.utils.bug
- Path: sparky!uunet!wupost!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!smart.bo.open.DE!tom
- From: tom@smart.bo.open.DE (Thomas Neumann)
- Subject: oleo CPU usage
- Message-ID: <1i9ka5INN4nh@smart.bo.open.de>
- Sender: gnulists@ai.mit.edu
- Organization: News at smart
- Distribution: gnu
- Date: Mon, 4 Jan 1993 15:15:16 GMT
- Approved: bug-gnu-utils@prep.ai.mit.edu
- Lines: 28
-
- I've just compiled oleo-1.2.1 on a NeXT running NeXTSTEP 3.0. Things that
- seem noteworthy:
-
- - oleo uses CPU time like crazy even when no user interaction
- takes place. Looks like it is waiting for keyboard input in raw
- mode a little bit too busy (just a guess).
-
- - the header file dependencies in Makefile.in look like some
- `make depend' generated stuff. It's not a good idea to have
- this in Makefile.in, which should be as generic as possible.
- I noticed this because Makefile.in contains a direct dependency to
- /usr/include/ctype.h which NeXTSTEP 3.0 does not have (it's
- in /NextDeveloper/Headers/ansi/ctype.h).
- This is of course trivial to fix, but nevertheless annoying.
-
- - when a floating point value is entered into a cell it gets
- reformatted to use the maximum number of possible decimal
- digits, e.g. one enters 42.1 and the cell gets filled with
- something like 42.10000000001.
- I know that this is due to a bug (or non-POSIX conformance)
- in NeXTs sprintf formatting code and therefore not strictly
- a bug in oleo, however, I suspect there may be more platforms
- that have this problem so maybe oleo should do it's own formatting,
- at least when it comes to screen output.
-
-
- bye -- Thomas
-
-