Wireshark · Ethereal-dev: Re: [Ethereal-dev] Did anyone succeed in cleanly building Ethereal withMSVC now?

["Olivier Biot" <ethereal@xxxxxxxxxx>]

From: Olivier Biot

| Hi list,
|
| If you perform a "nmake -f Makefile.nmake distclean", then a svn
| update, then a normal (n)make, did someone succeed in building
| Ethereal? I always get problems at the linkage step in epan as the
| adler32.obj and other top-level directory objects cannot be found
(see
| yasterday's mails for more details). Some time later, linkage fails
| again because of missing symbols in libethereal.def.

Following patch already helps, but it does not solve all issues:

$ svn diff epan/Makefile.nmake epan/libethereal.def
Index: epan/Makefile.nmake
===================================================================
--- epan/Makefile.nmake (revision 11556)
+++ epan/Makefile.nmake (working copy)
@@ -89,7 +89,7 @@
                /DEF:libethereal.def /OUT:libethereal.dll \
                /IMPLIB:libethereal.lib $(OBJECTS) $(libethereal_LIBS)
\
                ..\image\libethereal.res  dissectors\register.obj \
-                $(DISSECTOR_SUPPORT_OBJECTS:../=) $(EXTRA_OBJECTS)
+                $(DISSECTOR_SUPPORT_OBJECTS) $(EXTRA_OBJECTS)

 ethereal.lib   : ..\config.h $(OBJECTS)
        link /lib /out:ethereal.lib $(OBJECTS) $(EXTRA_OBJECTS)
Index: epan/libethereal.def
===================================================================
--- epan/libethereal.def        (revision 11556)
+++ epan/libethereal.def        (working copy)
@@ -86,7 +86,6 @@
 decode_enumerated_bitfield_shifted
 decode_numeric_bitfield
 deletefile
-destroy_print_stream
 dfilter_apply_edt
 dfilter_compile
 dfilter_dump
@@ -298,7 +297,6 @@
 mtp3_addr_to_str_buf
 mtp3_service_indicator_code_short_vals DATA
 new_create_dissector_handle
-new_page
 new_register_dissector
 nt_cmd_vals                     DATA
 num_tap_filters                 DATA
@@ -320,15 +318,10 @@
 prefs_register_string_preference
 prefs_register_uint_preference
 prefs_set_pref
-print_bookmark
 print_finale
 print_hex_data
 print_line
 print_preamble
-print_stream_ps_new
-print_stream_ps_stdio_new
-print_stream_text_new
-print_stream_text_stdio_new
 protocols_module                DATA
 proto_can_match_selected
 proto_can_toggle_protocol


I still need to find out why some objects get generated both in
top-level and in epan (xmlstub.obj, print.obj and ps.obj). Once I
found out, I'll check in a patch.

Best regards,

Olivier