[CLN-list] Result of building cln-1.1.13 on pc (Windows XP) with cygwin

[Richard Haney]

Hi,

--- Sheplyakov Alexei <varg at theor.jinr.ru> wrote:

> On Sat, Aug 12, 2006 at 04:41:07PM -0700, Richard Haney wrote:
> > Building cln-1.1.13 has a few nice improvements, but there were
> > a few build problems that required adjustments and reruns.
> 
> I'm sorry, but your message lacks any relevant info. PLEASE, post
> your config.log file. 

I'm sorry my message lacked any information relevant to your particular
interests, but I would think that some users might find much of that
information very useful.

For example, the fact that the CLN configure command took only about
two minutes could make a _huge_ difference in how someone goes about
building CLN or even in deciding to build CLN at all (say, for a
complete "newbie" only tentatively interested in cln).  The GiNaC
configure script takes about 30 minutes on the same computer.  The time
taken by cln configure made a difference for me in quickly deciding to
rerun configure with the added INSTALL=... argument.  It was an easy,
simple decision; I didn't have to weigh pros and cons.  And I seem to
recall that the configure commands took an extremely long time to run
on a previous computer a few years ago with about 40 percent of the
rated processor speed and somewhat less RAM (probably 256 MB).

Also, I would think the information I provided about the missing .exe
extensions for executable targets in the Makefiles would be extremely
useful for anyone wanting to fix that particular problem I described.

Since I suppose you are interested in the config.log file for the first
configure run after I changed the order of directories in my PATH
variable (and running "make distclean" and doing manual cleanup, as I
recall), I am attaching essentially a copy of that file, with slight
editing as noted for "host name".  (I'm not sure whether it could be a
significant point of vulnerability to Internet mischief if made
public.)  I'm also deleting the last directory name listed for PATH
because it seems irrelevant and could possibility be misconstrued as to
being of any general significance by sufficiently ignorant people.  I
suppose the information after "checking for suffix of executables" and
at "ac_cv_exeext=.exe" may be of particular interest in determining
where the autoconf/configure process may be in error in regard to the
missing .exe extensions for executable targets in Makefiles.

I hope this helps.

> > So even if DLL-building were supported in this
> > environment, very likely only one object module per C++ source
> > module would need to be built.)
> 
> Life is not that easy, unfortunatelly... First of all, the major
> issue is multiply defined symbols due to MAYBE_INLINE hack. This
> makes the linker fail, so a2dll will fail too. Secondly, to make
> any use of DLL one needs to insert a lot of __declspec stuff
> into the headers anyway.

As the a2dll documentation points out, there is often or typically
considerable work involved in converting a static library to a DLL;
it's not a simple matter of running a2dll with a static library as
input to get a DLL as output.  That documentation and your description
of the difficulties are why I used the phrase "industrious enough" to
describe someone who attempts to create a cln DLL from the static
library.

Hopefully, the config.log file will also help in determining why the
default value for the Makefile INSTALL variable causes a crash in the
doc directory for "make install".

Best regards,

Richard

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: config.log--edited--2006-08-10, 2..56 pm.txt
Url: http://www.cebix.net/pipermail/cln-list/attachments/20060814/293aabd2/config.log--edited--2006-08-102..56pm-0001.txt