[olsr-dev] HEAD build problems

[Patrick McCarty]

On 5/2/07, Bernd Petrovitsch <(spam-protected)> wrote:
> More background info: I'm thinking and playing with the "deploy one
> coding style" theme/issue. And "indent" cannot handle such
> pseudo-statements properly so this change/cleanup was a necessary
> prerequisite.

I figured something like this was the case. No worries.

> Probably because no one used tms_buf (and glibc and/or the Linux kernel
> is usually graceful on passed NULL pointers). However, reading the man
> page once more, I find:
> ----  snip  ----
>        On Linux, the buf argument can be specified as NULL,  with  the  result  that  times()  just
>        returns  a  function result.  However, POSIX does not specify this behaviour, and most other
>        Unix implementations require a non-NULL value for buf.
> ----  snip  ----
> I better readd it anyways with a comment .....

Correct, best I could tell the returned structure was completely unused.

> Hmm, running the bison commandline under ltrace or strace (doea OpenBSD
> has these things?) may show something (which fails but shouldn't) but I
> won't bet on it.

Heh.. I'm not aware if in fact they do.

I'm actually a Linux guy, but the project I'm working on calls for a
bit of a BSD flavor.

I ran bison on the file, and generated a .output file showing the
parse run. It does not indicate any conflicts or shifts that it can't
handle.

FWIW, these two tokens are unused, but not causing issues:
   TOK_SEMI
   TOK_PLNAME

I'm pretty well stumped. It's not throwing any errors (as any good
program should when it blows up).

I had to step away from this and work on something else...

> BTW we are working to get OpenBSD running under qemu. So let's see .....

I'm running OpenBSD under VMware here (For the devel)... mostly works...

-- patrick