Tom Lane [Fri, 26 Jan 2001 18:23:12 +0000 (18:23 +0000)]
Special case in ProcSleep() wasn't sufficiently general: must check to
see if we shouldn't block whenever we insert ourselves anywhere before
the end of the queue, not only at the front.
Vadim B. Mikheev [Fri, 26 Jan 2001 01:24:31 +0000 (01:24 +0000)]
First step in attempt to fix tree at runtime: create upper levels
and new root page if old root one was splitted but new root page
wasn't created.
New code is protected by FixBTree bool flag setted to FALSE, so
nothing should be affected by this untested approach.
Peter Mount [Thu, 25 Jan 2001 09:16:36 +0000 (09:16 +0000)]
Added an alternative constructor to PGSQLException so that debugging
some more osteric bugs is easier. If only 1 arg is supplied and it's
of type Exception, then that Exception's stacktrace is now included.
This was done as there's been a report of an unusual bug during connection.
This will make this sort of bug hunting easier from now on.
Bruce Momjian [Thu, 25 Jan 2001 03:28:27 +0000 (03:28 +0000)]
Further to the previous ODBC patches I posted today, I found a couple of
problems with char array sizes having set a couple of constants to 0 for
unlimited query length and row length. This additional patch cleans those
problems up by defining a new constant (STD_STATEMENT_LEN) to 65536 and
using that in place of MAX_STATEMENT_LEN.
Another constant (MAX_MESSAGE_LEN) was defined as 2*BLCKSZ, but is now
65536. This is used to define the length of the message buffer in a number
of places and as I understand it (probably not that well!) therefore also
places a limit on the query length. Fixing this properly is beyond my
capabilities but 65536 should hopefully be large enough for most people.
Apologies for being over-enthusiastic and posting 3 patches in one day
rather than 1 better tested one!
Bruce Momjian [Thu, 25 Jan 2001 03:27:05 +0000 (03:27 +0000)]
I decided to give this a go after all :-) The attached patch does the
following but it does *not* check whether the user is connected to
PostgreSQL 7.0.x or 7.1 first (as would be required for some of the
features) - the driver doesn't do this at all afaik and it's beyond my
capabilities to implement such checking in code that doesn't look like it
was written by my 1 year old daughter!
1) The driver now reports no maximum query length (SQL_MAX_QUERY_SIZE).
2) The driver now reports no maximum row length (SQL_MAX_ROW_SIZE).
3) The driver now reports that Outer Joins are supported (SQL_OUTER_JOINS),
but still does not report oj capabilities (SQL_OJ_CAPABILITIES).
4) The version number has been incremented to 7.1.0000 in psqlodbc.h *and*
psqlodbc.rc
Bruce Momjian [Thu, 25 Jan 2001 00:02:58 +0000 (00:02 +0000)]
This patch fixes an arrayindexoutofbounds exception that was just
introduced into the code. The fix is a fix to
org.postgresql.core.ByteArrayDim1.java.
Bruce Momjian [Wed, 24 Jan 2001 23:41:04 +0000 (23:41 +0000)]
Attached is a revised patch that removes the static SimpleDateFormat
objects that Thomas pointed out might be a problem.
PPS. I have included and updated the comments from the original patch
request to reflect the changes made in this revised patch.
> Attached is a set of patches for a couple of bugs dealing with
> timestamps in JDBC.
>
> Bug#1) Incorrect timestamp stored in DB if client timezone different
> than DB.
> The buggy implementation of setTimestamp() in PreparedStatement simply
> used the toString() method of the java.sql.Timestamp object to convert
> to a string to send to the database. The format of this is yyyy-MM-dd
> hh:mm:ss.SSS which doesn't include any timezone information. Therefore
> the DB assumes its timezone since none is specified. That is OK if the
> timezone of the client and server are the same, however if they are
> different the wrong timestamp is received by the server. For example if
> the client is running in timezone GMT and wants to send the timestamp
> for noon to a server running in PST (GMT-8 hours), then the server will
> receive 2000-01-12 12:00:00.0 and interprete it as 2000-01-12
> 12:00:00-08 which is 2000-01-12 04:00:00 in GMT. The fix is to send a
> format to the server that includes the timezone offset. For simplicity
> sake the fix uses a SimpleDateFormat object with its timezone set to GMT
> so that '+00' can be used as the timezone for postgresql. This is done
> as SimpleDateFormat doesn't support formating timezones in the way
> postgresql expects.
>
> Bug#2) Incorrect handling of partial seconds in getting timestamps from
> the DB
>
> When the SimpleDateFormat object parses a string with a format like
> yyyy-MM-dd hh:mm:ss.SS it expects the fractional seconds to be three
> decimal places (time precision in java is miliseconds = three decimal
> places). This seems like a bug in java to me, but it is unlikely to be
> fixed anytime soon, so the postgresql code needed modification to
> support the java behaviour. So for example a string of '2000-01-12
> 12:00:00.12-08' coming from the database was being converted to a
> timestamp object with a value of 2000-01-12 12:00:00.012GMT-08:00. The
> fix was to check for a '.' in the string and if one is found append on
> an extra zero to the fractional seconds part.
>
>
> I also did some cleanup in ResultSet.getTimestamp(). This method has
> had multiple patches applied some of which resulted in code that was no
> longer needed. For example the ISO timestamp format that postgresql
> uses specifies the timezone as an offset like '-08'. Code was added at
> one point to convert the postgresql format to the java one which is
> GMT-08:00, however the old code was left around which did nothing. So
> there was code that looked for yyyy-MM-dd hh:mm:sszzzzzzzzz and
> yyyy-MM-dd hh:mm:sszzz. This second format would never be encountered
> because zzz (i.e. -08) would be converted into the former (also note
> that the SimpleDateFormat object treats zzzzzzzzz and zzz the same, the
> number of z's does not matter).
>
>
> There was another problem/fix mentioned on the email lists today by
> mcannon@internet.com which is also fixed by this patch:
>
> Bug#3) Fractional seconds lost when getting timestamp from the DB
> A patch by Jan Thomea handled the case of yyyy-MM-dd hh:mm:sszzzzzzzzz
> but not the fractional seconds version yyyy-MM-dd hh:mm:ss.SSzzzzzzzzz.
> The code is fixed to handle this case as well.
Bruce Momjian [Wed, 24 Jan 2001 15:19:36 +0000 (15:19 +0000)]
It looks Ok, but it has one unnecessary step. There is no need to do the "mv
privkey.pem cert.pem.pw" if you just use "privkey.pem" in the following
openssl command (e.g. openssl rsa -in privkey.pem -out cert.pem".
But there is nothing wrong with it as it is now, as far as I can see.
Bruce Momjian [Wed, 24 Jan 2001 03:50:06 +0000 (03:50 +0000)]
attached is take-2 of a patch which fixes a bug related
to the use of getpwuid when running in standalone mode.
this patch allocates some persistent storage (using
strdup) to store the username obtained with getpwuid
in src/backend/main/main.c. this is necessary because
later on, getpwuid is called again (in ValidateBinary).
the man pages for getpwuid on SCO OpenServer, FreeBSD,
and Darwin all have words to this effect (this is from
the SCO OpenServer man page):
Note
====
All information is contained in a static area, so it must
be copied if it is to be saved. Otherwise, it may be
overwritten on subsequent calls to these routines.
in particular, on my platform, the storage used to hold
the pw_name from the first call is overwritten such that
it looks like an empty username. this causes a problem
later on in SetSessionUserIdFromUserName.
i'd assume this isn't a problem on most platforms because
getpwuid is called with the same UID both times, and the
same thing ends up happening to that static storage each
time. however, that's not guaranteed, and is _not_ what
happens on my platform (at least :).
this is for the version of 7.1 available via anon cvs as
of Tue Jan 23 15:14:00 2001 PST:
.../src/backend/main/main.c,v 1.37 2000/12/31 18:04:35 tgl Exp
Bruce Momjian [Wed, 24 Jan 2001 03:46:16 +0000 (03:46 +0000)]
I would like to do a interface change in pgcrypto. (Good
timing, I know :)) At the moment the digest() function returns
hexadecimal coded hash, but I want it to return pure binary. I
have also included functions encode() and decode() which support
'base64' and 'hex' encodings, so if anyone needs digest() in hex
he can do encode(digest(...), 'hex').
Main reason for it is "to do one thing and do it well" :)
Another reason is if someone needs really lot of digesting, in
the end he wants to store the binary not the hexadecimal result.
It is really silly to convert it to hex then back to binary
again. As I said if someone needs hex he can get it.
Well, and the real reason that I am doing encrypt()/decrypt()
functions and _they_ return binary. For testing I like to see
it in hex occasionally, but it is really wrong to let them
return hex. Only now it caught my eye that hex-coding in
digest() is wrong. When doing digest() I thought about 'common
case' but hacking with psql is probably _not_ the common case :)
Bruce Momjian [Wed, 24 Jan 2001 03:42:38 +0000 (03:42 +0000)]
Here is a patch to make the current snapshot compile on Win32 (native, libpq
and psql) again. Changes are:
1) psql requires the includes of "io.h" and "fcntl.h" in command.c in order
to make a call to open() work (io.h for _open(), fcntl.h for the O_xxx)
2) PG_VERSION is no longer defined in version.h[.in], but in configure.in.
Since we don't do configure on native win32, we need to put it in
config.h.win32 :-(
3) Added define of SYSCONFDIR to config.h.win32 - libpq won't compile
without it. This functionality is *NOT* tested - it's just defined as "" for
now. May work, may not.
4) DEF_PGPORT renamed to DEF_PGPORT_STR
I have done the "basic tests" on it - it connects to a database, and I can
run queries. Haven't tested any of the fancier functions (yet).
However, I stepped on a much bigger problem when fixing psql to work. It no
longer works when linked against the .DLL version of libpq (which the
Makefile does for it). I have left it linked against this version anyway,
pending the comments I get on this mail :-)
The problem is that there are strings being allocated from libpq.dll using
PQExpBuffers (for example, initPQExpBuffer() on line 92 of input.c). These
are being allocated using the malloc function used by libpq.dll. This
function *may* be different from the malloc function used by psql.exe - only
the resulting pointer must be valid. And with the default linking methods,
it *WILL* be different. Later, psql.exe tries to free() this string, at
which point it crashes because the free() function can't find the allocated
block (it's on the allocated blocks list used by the runtime lib of
libpq.dll).
Shouldn't the right thing to do be to have psql call termPQExpBuffer() on
the data instead? As it is now, gets_fromFile() will just return the pointer
received from the PQExpBuffer.data (this may well be present at several
places - this is the one I was bitten by so far). Isn't that kind of
"accessing the internals of the PQExpBuffer structure" wrong? Instead,
perhaps it shuold make a copy of the string, adn then termPQExpBuffer() it?
In that case, the string will have been allocated from within the same
library as the free() is called.
I can get it to work just fine by doing this - changing from (around line
100 of input.c):
and the same a bit further down in the same function.
But, as I said above, this may be at more places in the code? Perhaps
someone more familiar to it could comment on that?
What do you think shuld be done about this? Personally, I go by the "If you
allocate a piece of memory using an interface, use the same interface to
free it", but the question is how to make it work :-)
Also, AFAIK this only affects psql.exe, so the changes made to the libpq
this patch are required no matter how the other issue is handled.
Tom Lane [Wed, 24 Jan 2001 00:06:07 +0000 (00:06 +0000)]
Make functional index copy attstorage from the column data type, rather
than forcing 'plain'. This probably does not matter right now, but I
think it needs to be consistent with the regular (not-functional) index
case, where attstorage is copied from the underlying table. Clean up
some other dead and infelicitous code too.
Tom Lane [Tue, 23 Jan 2001 22:39:08 +0000 (22:39 +0000)]
Give 'a_expr ::= a_expr Op' production a slightly lower precedence than
Op, so that the sequence 'a_expr Op Op a_expr' will be parsed as
a_expr Op (Op a_expr) not (a_expr Op) Op a_expr as formerly. In other
words, prefer treating user-defined operators as prefix operators to
treating them as postfix operators, when there is an ambiguity.
Also clean up a couple of other infelicities in production priority
assignment --- for example, BETWEEN wasn't being given the intended
priority, but that of AND.
Bruce Momjian [Tue, 23 Jan 2001 20:36:30 +0000 (20:36 +0000)]
Subject: Bug in SQLForeignKeys()
Query used for checking foreign key triggers
returns too many results when there're more than one foreign
key in a table. It happens because only table's oid is used to
link between pg_trigger with INSERT check and pg_trigger with
UPDATE/DELETE check.
I think there should be enough to add following conditions
into WHERE clause of that query:
AND pt.tgconstrname = pg_trigger.tgconstrname
AND pt.tgconstrname = pg_trigger_1.tgconstrname
Tom Lane [Tue, 23 Jan 2001 04:32:23 +0000 (04:32 +0000)]
Fix all the places that called heap_update() and heap_delete() without
bothering to check the return value --- which meant that in case the
update or delete failed because of a concurrent update, you'd not find
out about it, except by observing later that the transaction produced
the wrong outcome. There are now subroutines simple_heap_update and
simple_heap_delete that should be used anyplace that you're not prepared
to do the full nine yards of coping with concurrent updates. In
practice, that seems to mean absolutely everywhere but the executor,
because *noplace* else was checking.
Tom Lane [Tue, 23 Jan 2001 02:32:26 +0000 (02:32 +0000)]
Remove no-longer-needed restriction against referencing system
attributes in a FieldSelect node --- all the places that manipulate
these work just fine with system attribute numbers. OK, it's a new
feature, so shoot me ...
Peter Eisentraut [Mon, 22 Jan 2001 23:34:33 +0000 (23:34 +0000)]
Rip out table expression section from SQL syntax chapter and develop it
into new chapter on query (SELECT) syntax. In the end this should become
a narrative and example-filled counterpart to the SELECT reference page.
Tom Lane [Mon, 22 Jan 2001 23:28:52 +0000 (23:28 +0000)]
Remove rangechecks on errno; just call strerror unconditionally. This
eliminates a raft of portability issues, including whether sys_nerr
exists, whether the platform has any valid negative errnos, etc. The
downside is minimal: errno shouldn't ever contain an invalid value anyway,
and if it does, reasonably modern versions of strerror will not choke.
This rangecheck idea seemed good at the time, but it's clearly a net loss,
and I apologize to all concerned for having ever put it in.
Tom Lane [Mon, 22 Jan 2001 22:30:06 +0000 (22:30 +0000)]
Clean up lockmanager data structures some more, in preparation for planned
rewrite of deadlock checking. Lock holder objects are now reachable from
the associated LOCK as well as from the owning PROC. This makes it
practical to find all the processes holding a lock, as well as all those
waiting on the lock. Also, clean up some of the grottier aspects of the
SHMQueue API, and cause the waitProcs list to be stored in the intuitive
direction instead of the nonintuitive one. (Bet you didn't know that
the code followed the 'prev' link to get to the next waiting process,
instead of the 'next' link. It doesn't do that anymore.)
Peter Eisentraut [Sun, 21 Jan 2001 22:02:01 +0000 (22:02 +0000)]
Description of table expressions, including join syntax, from Robert B.
Easter <reaster@comptechnews.com>, heavily massaged by me. Also cleaned up
value expressions a bit.
Bruce Momjian [Sun, 21 Jan 2001 03:49:14 +0000 (03:49 +0000)]
Hello,
here is the patch attached which do check in each BLOB operation, if we are
in transaction, and raise an error otherwise. This will prevent such mistakes.
Tom Lane [Sun, 21 Jan 2001 00:59:26 +0000 (00:59 +0000)]
Deal with C++ incompatibility of sys_nerr declaration by taking it out
of c.h altogether, and putting it into the only places that use it
(elog.c and exc.c), instead. Modify these routines to check for a
NULL or empty-string return from strerror, too, since some platforms
define strerror to return empty string for unknown errors (what a useless
definition that is ...). Clean up some cruft in ExcPrint while at it.
Tom Lane [Sat, 20 Jan 2001 23:07:27 +0000 (23:07 +0000)]
Get rid of sunos4-only strerror() macro, and arrange to use the
implementation in backend/port/strerror.c if configure finds no strerror
in libc, same as we do for snprintf and inet_aton.
Tom Lane [Sat, 20 Jan 2001 22:09:24 +0000 (22:09 +0000)]
Get rid of initdb -t bugs by the simple expedient of getting rid of
initdb -t. This option is obsoleted by 7.1's ability to drop and
recreate template1 during normal operation.