Re: [brlcad-devel] red not working with EDITOR=vi, BRL-CAD 7.18.2, Ubuntu

[Christopher S. M. <br...@ma...>]

On Mar 2, 2011, at 10:37 AM, Tom Browder wrote:

> The bug is not on the trunk version.  We need a new release soon.

We are in the middle of release testing now (which was actually halted to investigate this and a couple other issues).  The release show-stoppers are listed in TODO, but should hopefully all get investigated and fixed today/tomorrow at the latest.

The concern I have is that there have been no changes I can find that have been made to the red and ted sources since 7.18.2 was released.  So if .2 is broken and .3 is not, that's highly suspect.  It's A) either not broken in .2 or B) it's still broken for some case in .3 or C) I've missed some relevant commit.

The .2 release was around revision r43153 according to the Changelog.

Following the code, 'red' is a callback in src/mged/setup.c that calls f_red() in src/mged/utility1.c; last edit to that file was r42581.

Following f_red(), it calls ged_red() from src/libged/red.c; last edit to that file was r43216 and is the only commit after 43152, but that change was an equivalent conversion from NEAR_ZERO() to ZERO(), so it's logically unchanged.

So I'm at a loss as to what was broken or what was fixed, unless it was some underlying bu or rt routine that red's implementation is using.  Nothing comes to mind, though.

I'm willing to turn a blind eye if everyone else is confident that it's "really" working on trunk, now, but I'm not so convinced there is enough understanding to declare that.

Especially since this particular issue has resurfaced several times now, due diligence would suggest that someone check out r43153, verify it's not working, the progress forward through commits until it's working again to isolate the change that "fixed" it.  Takers?

Cheers!
Sean