Re: i guess we're frozen & stuff

[Ken Raeburn]

On Sep 16, 2009, at 15:00, Andy Wingo wrote:
> Hi Ken,
>
> On Tue 11 Aug 2009 15:59, Ken Raeburn <address@hidden> writes:
>
> > Perhaps I'm building [Guile] in ways that are unusual for the other
> > developers (build dir != src dir, libgmp and guile-1.8 installed in
> > the same place, libgmp and libunistring installed in different
> > nonstandard directories)?
>
> I think I fixed the case in which $prefix was not in LD_LIBRARY_PATH  
> or
> the like. Can you update us on the status of your build issues?

Frequently failing. :-(
On the bright side, using the --with-foo-prefix options, I haven't  
seen indications yet of the include- and load-path problems I saw  
before.

Today (f60e892), it can't compile i18n.c at line 799 on my Mac because  
it tries to deference scm_t_locale which points to an incomplete type.

Backing up to bcccf04, it builds okay, but lots of regexp tests fail  
with "illegal byte sequence" errors.  With "make -k check", that's the  
only error reported.

With f60e892, if I disable the defining of USE_GNU_LOCALE_API, I still  
get the regexp test errors, and srfi-18.test fails, but only sometimes:

Running srfi-18.test
WARNING: (srfi srfi-18): imported module (srfi srfi-34) overrides core  
binding `raise'
throw from within critical section.
error key: foo
/bin/sh: line 1: 74711 Abort trap              ${dir}$tst
FAIL: check-guile

I haven't checked if srfi-18.test has intermittent failures in the  
older revision.  It might also matter that I was sometimes piping the  
test output through "more", which could've conceivably blocked process  
output for a while, changing the timing of events between threads.

On GNU/Linux (RHEL 4), I had to download, build and install the GC  
library, and took the stable one linked to from the web page for the  
package.  But that didn't declare GC_dump, so I had to disable  
warnings in the build.  After that, building, installing and testing  
*seemed* to go just fine.

I'm seeing another weird effect though -- sometimes some of the tests  
outputs are getting reported twice.  In another test I just did on my  
Mac, the repetition isn't starting on line boundaries, though, so it  
looks like it's actually the output getting corrupted, vs re-running  
or re-reporting tests for no obvious reason:

Running socket.test
WARNING: (srfi srfi-18): imported module (srfi srfi-34) overrides core  
binding `raise'
WARNING: (test-srfi-69): imported module (srfi srfi-69) overrides core  
binding `make-hash-table'
WARNING: (test-srfi-69): imported module (srfi srfi-69) overrides core  
binding `hash-table?'
exp" "illegal byte sequence" #f ("a<C7>")))
ERROR: regexp.test: regexp-quote: regexp/extended: (string "aX" 200 # 
\<C8> "a<C8>" "a<C8>") - arguments:
 ((regular-expression-syntax "make-regexp" "illegal byte sequence" #f  
("a<C8>")))

I've seen the duplicated output on GNU/Linux too.  In reproducing  
this, again, it seems to help to pipe the output through "more", and  
follow it slowly (i.e., so presumably the pipe will fill and output  
will occasionally be blocked).

Ken