Redownloaded everything and started from scratch with latest cvs check and bazaar gnupdf.
I am testing in a windows XP machine with msys and mingw32. I've been told running OK under wine is not actually a requirement, so I've just left it aside and went directly to windows.
Both libcheck and libgnupdf compile smoothly (aside from the trivial issues mentioned in my other mail), and making check in libgnupdf runs ok and, although I get some Failures, there are no segfaults (which there used to be!!).
base/text/pdf-text-new-destroy.c:44:F:pdf_text_new_destroy:pdf_text_new_destroy_001:0: Assertion 'pdf_text_new (&newtext) == PDF_EBADCONTEXT' failed
base/types/pdf-i64-mod.c:58:F:pdf_i64_mod:pdf_i64_mod_001:0: Failure 'dest.low != 1' occured
base/types/pdf-i64-mod.c:120:F:pdf_i64_mod:pdf_i64_mod_003:0: Failure 'dest.low!= 0xFFFFFFFE' occured
base/types/pdf-i64-mod.c:157:F:pdf_i64_mod:pdf_i64_mod_004:0: Failure 'dest.low!= 0xFFFFFFFE' occured
base/stm/pdf-stm-read.c:901:F:pdf_stm_read:pdf_stm_read_014:0: Failure 'ret != PDF_OK' occured
base/stm/pdf-stm-bseek.c:191:F:pdf_stm_bseek:pdf_stm_bseek_003:0: Failure 'ret_char != '3'' occured
base/stm/pdf-stm-bseek.c:266:F:pdf_stm_bseek:pdf_stm_bseek_004:0: Failure 'ret_char != '9'' occured
base/stm/pdf-stm-write.c:840:F:pdf_stm_write:pdf_stm_write_013:0: Failure 'ret != PDF_OK' occured
base/stm/pdf-stm-write.c:918:F:pdf_stm_write:pdf_stm_write_014:0: Failure 'ret != PDF_OK' occured
base/fsys/pdf-fsys-get-free-space.c:56:F:pdf_fsys_get_free_space:pdf_fsys_get_free_space_001:0: Failure 'pdf_i64_cmp(free_space, error) == 0' occured
base/fsys/pdf-fsys-file-open.c:127:F:pdf_fsys_file_open:pdf_fsys_file_open_003:0: Failure 'pdf_fsys_file_open(NULL, path, PDF_FSYS_OPEN_MODE_WRITE, &file) != PDF_EBADPERMS' occured
The last one occurs also in the linux build, so there are 10 new failures.
Even though this seems quite fine, I do get some segfaults when I run the test units that come with check itself, both in windows and running with wine in debian. This may indicate that win32 check is still buggy. I will now communicate the errors to check-devel and see what comes up.
As to the new failures, I could pay a more thorough look at them, as they may contain some good info about what's going on with check.
Any advice or comments?
Pablo.