Details regarding Bug #1383
joel.sherrill at OARcorp.com
Mon Mar 9 09:21:22 CDT 2009
Santosh vattam wrote:
> I see that this bug was detected on account of another bug - bug #1376
> - which says the lseek handler returns an int whereas lseek returns
> off_t. Are we waiting for bug #1383 to be fixed in order to fix bug
> #1376? Because from the bug report I see that the problem has been
> tackled, if not fixed. I have tried to fix the bug #1376 and have
> generated a patch which I have attached here. If we are not waiting
> for bug #1383 to be fixed, then I request you to kindly review my
> patch and tell me if there are mistakes in it.
This is the tip of the patch. This is where the per file
type handler prototype is defined. That makes the
code in libcsupport correct but doesn't address the
actual implementations of the lseek handlers in
the various filesystems. For example, in the IMFS:
This is a very good example of how a simple appearing
patch can be more than meets the eye.
+ When you generate a diff use -u
+ I put you patch in my tree, compiled and looked at
the new warnings. Then I used that to generate the
+ 4.10 includes pipe/fifo and has enough changes to
libio.h where my 4.10 patch did not apply to 4.9.
+ I created a 4.9 patch and tried it on 4.8. It again
did not apply 100% clean because 4.8 does not have the
NFS client. So I created a 4.8 patch.
+ I attached all three patches to the PR, committed
and closed the PR.
+ But I still had to update the 4.9 and 4.8 release
notes pages on the wiki.
I am not faulting you and appreciate the kick in the tush
to fix this. Since this is Summer of Code season, this just
turned out to be a great example. Your patch was correct
And I wanted to get a new release out today or tomorrow
so fixing this for 4.9.2 is important. :-D
Joel Sherrill, Ph.D. Director of Research & Development
joel.sherrill at OARcorp.com On-Line Applications Research
Ask me about RTEMS: a free RTOS Huntsville AL 35805
Support Available (256) 722-9985
More information about the rtems-users