🚀 go-pugleaf

David Masterson  <David.Masterson@synopsys.com> wrote:
 > Did this get out?

Yes, both copies.

 > Previously, I wrote:
 > 
 > > While I know that generally VM is not supported on strange M$
 > > systems, some of us have to live it.  So, I have a few suggestions
 > > that I think will be helpful, but aren't too M$ specific.  This is
 > > for Windows-NT and MKS and Emacs 21.1.
 > 
 > > First, I tried to download the vm.elc file from the Wonderworks
 > > download page.  Until later, I forgot that the ELC format for Emacs
 > > is different than XEmacs.  You might want to rearrange the download
 > > page to make that more obvious so that people don't download the
 > > wrong version of the ELC.

I will remove the vm.elc file, since its existence is confusing.
There's no common elc format anymore.  At some point Emacs
stopped being backward compatible with v19 compiled elc files, so
there's no longer any way to build an elc that works with all
Emacs and XEmacs versions.

 > > Second, I also wanted the Info file(s) for VM so that I have the
 > > manual to refer when needed.  This wasn't directly on the download
 > > page, but was in the "browse".  You might want to mention this on
 > > the download page as well.

Noted.

 > > BTW, you might also include the VM version number *in* the Info
 > > manual so that people know which version of VM the manual applies
 > > to.  That's useful when you're installing things piecemeal to ensure
 > > that you don't get two versions of VM mixed up together.

Done.

 > > I found that bbdb-insinuate-vm (BBDB 2.32) expects to find
 > > vm-autoloads, so I had to download the sources to VM after all.  You
 > > might want to let the BBDB people know if there is a better way of
 > > loading (and insinuating into) VM.

I have no idea what actions BBDB presently takes, so I couldn't
help them without further guidance.

 > > After downloading the sources, the first step is to modify the
 > > Makefile.  I'd suggest moving the SHELL definition above the user
 > > configurable line as sometimes the shell that people want to use
 > > (even on UNIX systems) is not /bin/sh.

If you're building someplace that doesn't have /bin/sh, you
should be worrying about a lot more than editing below the "no
user servicable parts" line.  If /bin/sh is available, then
that's what should be used.  Otherwise, you're really on your
own.

 > > I'd also suggest putting in a commented out setting for CC
 > > and CFLAGS above the user configurable line so that people
 > > know they *may* need to modify it.  I'd also put into the
 > > Makefile an EXE macro that will be defined to "" on UNIX
 > > and ".exe" on Windows.  With that you could change the
 > > definition of the UTILS to include the EXE macro as in
 > > "qp-decode$(EXE)" (also any appropriate Makefile targets).
 > > [...]

This is more than I am willing to do to support Windows.  This is a
point of principle and self-interest more than of labor.  Inasmuch
as Microsoft is trying to destroy the open source community and the
open standards upon which I base my livelihood, it isn't in my
self-interest to invest much time in making it easier to use their
products.

 > > Finally, in the target to build vm-autoload.elc, you do a "tr -d" to
 > > remove CRs from the file, but then you add the "provide" line to the
 > > end of the file.  On my system, the echo terminated each line with
 > > CRLF, so this line cause a premature EOF in the parsing.  You need
 > > to move the "tr" command.

Thanks.  I've fixed this for VM 7.15.