On Thu, 16 Aug 2007 19:12:19 +0000, Greg Lee wrote: > On Wed, 15 Aug 2007 17:13:57 +0000, SciFi wrote: > >> So, let me go ahead to open a report >> just so we'll have it on record anyway. > > I read your comment #7 about clicking on the Subject header and getting > a new sort done on a previously selected field. Very interesting. My > guess is that it's a data alignment problem -- there's a number > somewhere keeping track of which field should be sorted, but only the > lower order 2-3 bits are correct. Something bad is happening to the > other bits -- maybe they're overwritten by other data. In that case, > changing your preferences about what fields you want to be displayed > might change the behavior. And it raises the possibility that it might > be a compiler bug, so you might consider updating your compilers. (I'm > not optimistic about actually finding the bug, but maybe you can kill > it.)
Hi Greg, Apple's compilers won't be updated until/unless they update their XCode system. We're stuck with gcc 4.0.1 & 3.3 ATM (I think to compile for intel-native machines requires 4.0.1, this is Apple's rule). IIUC Apple does selectively apply patches, so I wouldn't say it is 'exactly' 4.0.1 we're using. Yes I could go compile gcc-4.2.1 right now but I've never ever been able to make it work in the XCode's place -- there's just too much interwoven (the STL stuff for instance) under Apple's compiler tree. So much so that I'd bet the closed-source parts of OSX wouldn't work anymore. At any rate, I am expecting a big revamp of XCode when Leopard comes out (as I said in my bugreport, I can't afford a $pay-for$ ADC account to get to use the betas coming out, and I do not trust the p2p versions floating around; I have been publicly wishing some rich dude would "mentor" me so I could do this kind of testing with Leopard while it is still beta, thus affect its liklihood of working better with open source projects everywhere, yes even mentioned this idea on Apple's maillists etc. ... ... but I guess I'm the only one worried about such things, eh... and I'm digressing I know...). Another thought struck me: that it might be an endian issue, where some of gtk/glib/pan2/etc. detects 'darwin' but doesn't detect the 'endianness' of the CPU -- a common mistake in configure scripts when Apple made that "switch". But as I mentioned in my bugreport, I do not have access to my ppc machines ATM to see what might need patching. I suppose all I can do is wait for another Mac user to try testing this, or maybe another non-Mac user will see it... p.s. this discussion & any info may get lost if it's not entered into the bugreport. ;) _______________________________________________ Pan-users mailing list Pan-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/pan-users
