[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Chicken-users] Inch by inch, row by row
From: |
Brandon J. Van Every |
Subject: |
Re: [Chicken-users] Inch by inch, row by row |
Date: |
Sun, 16 Jul 2006 21:43:53 -0700 |
User-agent: |
Thunderbird 1.5.0.4 (Windows/20060516) |
John Cowan wrote:
Okay. I'm now building on Cygwin using both autotools and
cmake in parallel (but only installing one at a time, obviously),
and here's the current state of progress:
0) Cmake doesn't have a "make uninstall", so I wrote a rather
drastic script called "zap-chicken" that attempts to zap
all currently installed traces of Chicken. Obviously, don't
put it in /usr/local/bin! I've attached it.
By design, CMake doesn't provide an uninstall. In the FAQ, they state
that uninstalls are dangerous. True, but so's breathing. I suppose we
could conjure an uninstall, but it would need to be written as a CMake
script, not a Unix shell script. At this moment it's not my priority.
I'm in the middle of a major refactoring because of the Windows pathname
issues. That is the only issue I consider important right now.
1) Both builds work correctly using the current darcs head.
2) "(use srfi-1)" fails in both; it is trying to load
/usr/local/bin/cygchicken.dll, but CMake is building
/usr/local/bin/libchicken.dll, whereas autotools is still
building /usr/local/bin/cygchicken-0.dll.
Felix wuz due to expunge cygchicken.dll. After which, I'd expect both
builds to work.
3) Using "ln" to link libchicken or cygchicken-0 as the case
may be to cygchicken eliminates the "No such file or directory" error,
but leaves us with this error instead:
#;1> (use srfi-1)
; loading library srfi-1 ...
nursery is too small - try higher setting using the `-:s' option - execution
terminated
Well, this is what partially motivated me to implement nursery support
over the past 2 days. The ability to poke at the nursery is there. But
is that really the problem? I have no idea. Again, I'm not going to
work on that right now. I only care about Windows pathnames... which
might in turn solve other problems.
4) Setting any -:s value, even -:s1000000, does not change the above.
I don't even know what it does or how it does. Maybe it has no effect
at all, the size of the stack is hardwired at build time, and the
message is a legacy. For all I know.
Grumble. Back to 2.315.
It is difficult to keep in synch with other developers when fixing bugs.
Cheers,
Brandon Van Every