Re: trouble compiling

[Daniel J Sebald]

On 05/18/2012 01:43 PM, Doug Stewart wrote:
> On Fri, May 18, 2012 at 2:17 PM, Jordi Gutiérrez Hermoso
> <address@hidden <mailto:address@hidden>> wrote:
>
>     On 18 May 2012 13:50, Daniel J Sebald <address@hidden
>     <mailto:address@hidden>> wrote:
>     >  Apparently I need to reconstruct lex.cc, so why wasn't this flagged
>     >  as an error?  Any reason?  Or just an oversight in a rather big
>     >  config process?
>
>     Oh, by the way, the reason that error message about flex is phrased
>     that way because it's primarily aimed at people who work from
>     tarballs. In a tarball, lex.cc is supplied, hence the word
>     "reconstruct". The build system has a few extra steps when building
>     from hg. I'm not sure how to document more prominently that more work
>     needs to be done to compile from hg than a tarball. I guess expanding
>     the HACKING file?
>
>     - Jordi G. H.
>     x
>
>
>
>
>
> I did extensive testing and I can prove that
>   hg st -un0 | xargs -0 rm
> is the command that fixed my problems.
> I don't understand exactly why,  or where the problem came from, but
> this fixes it.

OK, thanks Doug.  These past two emails have motivated me to go back and 
run "make clean".  After doing so, I see that lex.cc is still present. 
So that is where a problem lies.  lex.cc isn't being reconstructed after 
the first bad building.  The solution would be to add removal of lex.cc 
from the clean/configure process, but then won't that cause a problem 
for those who've gotten the code from a tarball?

Dan