trochee: (Default)
trochee ([personal profile] trochee) wrote2005-05-06 11:54 am
Entry tags:

bug squashed

irritating parser problem turns out to be that it fails on the latest version of g++. But the group releasing it still is back at 3.3.5. Recompiling everything with 3.3 (the closest match I have) seems to fix the problem. [edit: actually, it compiles just fine but has radically different behavior under the new stuff: it fails on sentences longer than one word. downgrading the compiler avoids this.]

*grumbles, shakes fist in direction of makefile, which does not invoke -Wall*

still no clean target. But the guy who helped me debug this asked me for a list of suggested changes to the architecture, and clean (etcetera) targets are at the top of the list. And I just sent him another one, suggesting -Wall.

[identity profile] xaosenkosmos.livejournal.com 2005-05-06 08:46 pm (UTC)(link)
which does not invoke -Wall*

Remember that list i mentioned before? That's on there, too =)

Good to hear that the source of the bug has been ferreted out, it's at least a step towards progress.

[identity profile] trochee.livejournal.com 2005-05-06 09:25 pm (UTC)(link)
Remember that list i mentioned before? That's on there, too =)
yah. and the use of g++ instead of $(CXX).

[identity profile] http://users.livejournal.com/_dkg_/ 2005-05-13 04:45 pm (UTC)(link)
In addition to -Wall, i like to use --pedantic. Working under hard constraints definitely helps me think through what i'm doing more.