[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: gnats/181: CHDB command must be issued twice in order to work
From: |
Yngve Svendsen |
Subject: |
Re: gnats/181: CHDB command must be issued twice in order to work |
Date: |
15 Oct 2001 09:04:01 -0000 |
The following reply was made to PR gnats/181; it has been noted by GNATS.
From: Yngve Svendsen <address@hidden>
To: Milan Zamazal <address@hidden>,address@hidden
Cc:
Subject: Re: gnats/181: CHDB command must be issued twice in order to
work
Date: Mon, 15 Oct 2001 11:02:53 +0200
At 22:47 14.10.2001 +0200, Yngve Svendsen wrote:
>Hm. I think I have isolated the problem with the CHDB command to the
>find_chain_entry_nocopy in adm.c. I ran debug sessions on identical
>configs, and on Solaris, this subroutine returned (AdmEntry *) 0x0, and on
>Linux (AdmEntry *) 0x8078fd0.
>
>But I can't quite see why there should be a difference there?
Sorry, my mistake. I am still not quite used to gdb -- seems I read the
value one step too early.
I have, however, found another difference in behaviour:
When I do the CHDB that fails, everything runs the same on the two OSes,
all the way to the call res = findDatabaseEntry (databaseEnt); in database.c
On Linux, res == 0 after this call, but on Solaris, it is nonzero. Thus,
the following
res = loadDatabase (databaseName, databaseEnt, err);
is not run on Solaris.
It does seem that this return value is the problem, because if I manually
set res to 0 after the findDatabaseEntry call on Solaris, the change of
database succeeds.
- Yngve