On Mon, 9 Jul 2007, Mark Morgan Lloyd wrote:
> Vincent Snijders wrote: > > >> I can put a bit more time into this, but I don't know SPARC very well > >> at the machine level and my experience with gdb is limited. > > > > Can't you copuy the TTextStyle declaration (and dependants) into a test > > program and assign the alignment field as in the snippet? > > I don't think it's a simple alignment problem. I've just spent a few hours > eliminating the attractive heresy that somebody had slipped in a reserved word > "Alignment" that was screwing things, and am now looking at an apparent > situation where CPU state is in a mess on entry to TCustomStringGrid.Create > with a bus error very shortly after- however I don't know gdb well enough to > be able to look at things in detail. > > I'm reminded of the problem in gtkwinapiwindow.pp that I fixed a week ago by > inserting a DebugLn() in front of the suspect code. That was with 2.0.4 and > 0.9.22, I checked earlier and I didn't have to apply the same hack this time > to get it working. > > Does anybody know what the current situation is with regards to automated > tests for FPC and Lazarus? I tried the FPC tests earlier and while on x86 they > completed and appeared to give me sensible results (although not being an FPC > guru I don't know whether they were the /right/ sensible results :-) but on > SPARC they bombed early on- I interpreted that as bad maintenance of the test > suite. It's not bad maintenance. On the contrary. The tests are run on a daily basis, also for Sparc. The daily FPC test suite results can be viewed on-line: http://www.freepascal.org/cgi-bin/testsuite.cgi You can compare with your results. The tests are processor independent, i.e. the idea is that all CPU's get more or less the same result. Michael. _________________________________________________________________ To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe" as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
