> > I have tested it in two very different machines and It fails as soon > > as you format a cell. > > no, my laptop runs -current.
Oh great, it means that they have fixed it! > > > It is not "every possible combination of program and input", it is a > > failure that appears as soon as you format a cell. > > you're missing the big picture. it is "every possible combination of > program and input" - but we can't investigate stuff unless people > report it. and people won't report bugs in features they never use. > and we hardly ever backport fixes to old ports. > > programs mysteriously conflict with each other all the time. authors > change APIs or dependencies and then everything breaks. that's why it > took us so long to upgrade gettext... and i'm pretty sure that it's > not hard to find gnumeric users who don't format cells. i hardly ever > do: people send me excel spreadsheets, i look at them, say yes or no, > and that's the end of it. Well, I understand, but formating a cell under a spreadsheet is so important thing that I can not conveive a spreadcheet without it. You need formating a numeric cell for example to stablish the quantity of digits on the right of the decimal point. It is just an example, but it is heavily used. If gnumeric is broken, do not release it, I would have choosen kspread without complaining here. I think that I am saying very reasonable things. Some people start getting angry very quickly. I of course appreciate your work and the work of all the people who contribute to OpenBSD. I just want to say that I was disappoited on how many basic bugs I found on very few time on OpenBSD ports. I just wanted to do constructive comments. > > > That could be an interesting thing, the problem is that my C > > programming skills are very limited, just enough to write very simple > > programs. > > > > I am pleased to do any test you need. > > well this is a good opportunity to learn C, then. > > read bsd.port.mk(7) - pay attention to the parts about compiler flags. > build gnumeric and all its dependencies with -g, don't strip the > output, and see if you can get a crash dump. then use gdb or ddd to > look at the crash. if nothing else, that'll give you someplace to > start writing a bug report. you might even get lucky and not have to > build with debug symbols - look at the core and see if it tells you > where it crashed. Thank you very much for the interesting information. I will start reading, studying and understanding what you recomended. It will be hard work, I am 35 and thing I am too old to understand such complex software ;-). Programming is something one has to learn very early, in order to understand it easily. I appreciate very much your constructive email. Regards. Ramiro. > > when fixing ports, don't waste your time on a released version (3.8, > 3.9, etc). those are frozen and rarely ever change. if you find a in a > release, see if it exists in -current and try fix or report it. > > CK > > -- > GDB has a 'break' feature; why doesn't it have 'fix' too? > ;-)
