The Samba-Bugzilla – Bug 2225
Compilation errors in samba 3.0.10 on DGUX R4.20MU04 (aviion, data general)
Last modified: 2006-04-14 14:54:16 UTC
first of all a big THANK YOU for this product and for the certainly huge amount
of work and time you are spending on it!
Compiling samba 3.0.10 on a DGUX machine I got a number of compiler errors. If
possible, I would like a feedback if the corrections might cause problems (this
is my third or fourth attempt to get a samba v3 running on that maching in order
to be able to work with XP SP2 PCs).
- File smbd/vfs.c, line 889: the 'pstring' type specification for the first
parameter in the function call is obviously wrong. I have simply deleted it.
- File libsmb/clireadwrite.c, lines 54 and 301: The compiler warns that the
right shift count >= width of type (for offset). I have now typecasted it into
- client/client.c: In line 699 and 1075 I get warnings that the fourth argument
(&size) is of incompatible pointer type (it is size_t and expected is
BIG_SMB_UINT). In line 2563 the PATH_MAX is undefined. I have now encapsulated
the two lines there into an '#ifdef PATH_MAX' directive.
- client/clitar.c, line 1487: Again, PATH_MAX is undefined. I have defiend it
here with a value of 512 which seems to be a POSIX minimum value.
Besides that there are about a warnings that 'const' or volatile' attributes of
variables are discarded in argument passing, returns, or assignments, and some
warning about incompatible pointer types. If you are interested in these, I
could certainly re-compile making a log of the make/gcc output.
please retest against 3.0.11 since the changeset is rather large. If all
of the issues have not been addressed, please reopen the bug report and
reset the version. Thanks.
All problems are still there in 3.0.14a.
Changed line numbers in the cource code:
- smbd/vfs.c: parse error in 897
- libsmb/clireadwrite: too large shift in 55 and 308
- client/client.c: Warnings in 767 and 1158; undefined PATH_MAX in 2922
- client/clitar.c: undefined PATH_MAX in 1476
The vfs.c bug seems to be exactly the one listed in bug report #2560.
Breaking dependcy on 2560 as it's fixed and I want to close it.
Please retest against a current release and reopen if the issue is still present.