Date: Mon, 9 Jul 2001 13:44:23 +0300 (IDT) From: Eli Zaretskii X-Sender: eliz AT is To: djgpp-workers AT delorie DOT com Subject: bfdsymify Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Reply-To: djgpp-workers AT delorie DOT com With the proliferation of use of non-COFF debugging by DJGPP users (I expect that to go up even more once GCC 3.0 and GDB 5.1 are released, with support for DWARF2), I think we should seriously consider distribution a version of SYMIFY linked with the BFD library (a.k.a. bfdsymify). It is IMHO unreasonable to expect users to recompile their programs just to get a human-readable backtrace. (An alternative to the backtrace is the core dump support, but since it's only being developed now, I think it's too early to consider it as a replacement for SYMIFY-type functionality.) I have bfdsymify working for me for a couple of years with no visible problems. So it can be considered stable enough to be used by every DJGPP user. Do people think it's a good idea to release bfdsymify to the masses? If so, I think we should discuss the practical ways of doing so. When this issue surfaced some time ago, DJ suggested to submit the program for inclusion in the Binutils package, so that it will be always available withe every port of Binutils. But this requires that bfdsymify be portable to platforms other than DJGPP, which is not easy, given the screen capture code. (Btw, do GNU/Linux systems have a device yet that allows direct video memory access?) I don't have time to do such ports, and don't have an easy access to GNU/Linux to test the ported code. So, if we agree on this approach, Someone (tm) other than myself will have to sit down and make it happen. Alternatively, we could make bfdsymify part of djdev, and augment the build process to build it only if libbfd.a is installed. Yet another possibility is to distribute it as a separate package. Thoughts?