[MOAB-dev] make dist and cgm2moab.man

Steve Jackson sjackson at cae.wisc.edu
Thu Feb 25 16:05:26 CST 2010


(Jason's getting this message twice because I forgot to send it to the list last time.  Sorry!)

On Feb 25, 2010, at 11:25 , Jason Kraftcheck wrote:
> 
> The handling of the man page in the Makefile.am looks strange.  Why is it
> 'dist_man1_MANS' rather than just 'man1_MANS'?  I don't even know what
> 'dist_man1_MANS' is supposed to do.

The line in question reads: 'dist_man1_MANS = cgm2moab.man'

This line was added to MOAB by jakraft (==Jason?) in 2006.  The commit message was "Include man pages in tarball created with 'dist' target."  So it seems to be trying to do the right thing, but because the line has since been moved into the 'if HAVE_CGM block,' I guess it doesn't always trigger when a dist tarball is being made.

>  I'd try the following changes, in this
> order until things work correctly:
>  a) s/dist_man1_MANS/man1_MANS/
>  b) move man1_MANS outside of the conditional block.
>     E.g. conditionally define $(cgm2moabman) and then unconditionally
>     define: man1_MANS = $(cgm2moabman)

Neither of these solutions work.  I tried moving the dist_man1_MANS line outside the conditional block.  That does work, but it means that cgm2moab.1 gets installed under share/man/man1 even if cgm is disabled and cgm2moab itself does not exist.  That's probably not what we want either.  So we might have to fall back on your third suggestion.

On the other hand, now that MOAB has a ReadCGM module, CGM files can be processed with mbconvert.  Is there any need for the cgm2moab program at all?  Even if there is, the man page doesn't contain much more than the output of cgm2moab -h.  Perhaps the best solution to this (admittedly minor) issue would be to remove cgm2moab's man page, or (if no one needs it any more) remove cgm2moab altogether.
~S




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/moab-dev/attachments/20100225/20425d33/attachment.htm>


More information about the moab-dev mailing list