🚀 go-pugleaf

>       Uuu! I used to hack a Soviet PDP-11/34 clone named SM-4.
>       On this box an anonimous hacker hung an ad of
>       "sperm-sucking machine SM-4" :-) :-) :-) oh, poor bulls!
>...

Hilarious... [how was it in Russian, by the way?]

This reminds me... There is a little LAN in midtown Manhattan and 2 386 boxes
on it are named HAMMER and SICKLE (you can guess who put in the LAN and who
christened the nodes). Unfortunately, I ran out of Bolshevitzky paraphernalia
for names after 2: the other nodes are called COGWHEEL, ROLLINGROCK, and other
silly names. (Naming nodes is hard!)

---

Speaking of error messages: there's a very nice program out there called SQL
Server for OS/2. It was developed by Sybase and is distributed by Microsoft.
I've been using it for about a year and I've been collecting the more bizarre
messages from it.

One of SQL server's distinguishing features is that it "optimizes" SQL queries,
that is, if you ask it something, it will think long and hard about how to
compute the answer in the most efficient manner. Sometimes this optimization
gets confused when confronted by a non-trivial query. I've been haunted by
these 2 messages lately:

>Cannot run query--referenced object (name NAME NOT RECOVERABLE) dropped during
>query optimization.
and
>Cannot drop the table 'drop by id' because it is currently in use.

(Explanation: it creates temporary objects, then drops them, then realizes it
still wants them.)

It also deadlocks very easily. Once I got this message:
>Deadlock encountered in log allocation for database 'studies'. This state
>should not be reached.
and all too often I get messages of the sort:
>Your server command (process id 5) was deadlocked with another process and
>has been chosen as deadlock victim.  Re-run your command.

Finally, there's a utility there called DBCC, for database consistency
checking. It always ends with the message:
>DBCC execution completed. If DBCC printed error messages, see your System
>Administrator.
Well, I am the SA and the sole user, so this looks sort of silly to me...
(Also, is it hard to have a flag that would remind it whether or not it printed
any error messages? :)

>DBCC detected internal database inconsistancy; see your System Administrator.
The only way I found to fix it was to erase everything and to reload from tape.

>Number of data pages counted by DBCC differs from count in Sysindexes row.
>Please UPDATE 'dpages' in the Sysindexes row for this object to equal the
>count just printed.

According to the manual, the above condition arises naturally (is "normal"),
and if not corrected, it leads to this when you try to order by a key:

>Sort failed because dpages in the Sysindexes row for table 'stat_u' in
>database 'studies' had an incorrect value. Please run DBCC CHECKTABLE on this
>table to find the correct value and then update sysindexes.dpages. Then retry
>the sort.

The manual gives a long and boring procedure for correcting it (which I
automated, and run unattended once a week, so at least I know how to deal with
this one:).

---

Finally, speaking of Russian error messages: the Russian typing tutor program,
that I wrote and that still floats around the BBS's, displays messages in both
Russian and English. One of them says:
>Press `Esc'' to quit. / Navmite `Esc'', ~toby kon~itx.
and a a few users have remarked that this is ambiguous.

D.V.