#lisp - Wed 21 Mar 2007 between 00:05 and 00:55

ModiusKind of lisp related as Microsoft's identity columns aren't exposed through CL-SQL. . . . what SQL technique is used in non-MS databases to get the equivalent of identity columns (that is, incrementing unique ID columns)?
rudiproduct-specific techniques outlined in these non-MS database's manuals, I imagine.
Jaskoautoincrement columns
rudiSQL (relational algebra, really) doesn't believe in row identity
Cinis package::something a way to define things inside a package from without?
pkhuongCin: it's a way to access unexported symbols.
kpreidit's a way to refer to non-exported symbols in a package without being in that package
Cinoh, oops
I've got a class in one package, and I've defined a subclass in another using that, e.g. (defclass example-window (spiffy:window) ()) and then defined a method for that subclass
is it possible for the other package to access that in a standardish way?
that method, that is
nyefWell, the method is presumably on a generic function known to the other package, so normal GF dispatch should do you if you pass arguments of the right type.
Cinhm, that's what I thought and tried. the package called the method defined within itself (of the superclass)
I guess I did it wrongly. hehe
rudiCin: what is the output of (apropos "the-method-that-Cin-defined") ? if it prints information about two fbound symbols, you have package confusion
Cinlet's see
it shows, SPIFFY::ON-CREATE and SPIFFY-EXAMPLE::ON-CREATE, both functions
rudithere ya go, then - change (defmethod on-create ...) to (defmethod spiffy:on-create ...)
Cinah, okay :)
rudiand call spiffy:on-create instead of on-create, of course
jsnellwhich of the following do you think looks most readable? http://jsnell.iki.fi/tmp/1.html http://jsnell.iki.fi/tmp/2.html http://jsnell.iki.fi/tmp/3.html or http://jsnell.iki.fi/tmp/4.html
Cinrudi, yeah. so that'll enable it to differentiate between classes and call the right spiffy:on-create, yes?
rudiCin: if I understood you correctly, yes
nyefjsnell: My preferences are (in descending order) 2, 1, 3, 4.
Cinrudi, just confirming
actionrudi votes for 2
rudivotes for 2
nyef(4 is really obnoxious looking.)
rudiindeed, my preferences are the same as nyef's
tritchey_2 looks nice
pkhuongjsnell: did you put the line numbers on their own line on purpose?
nyefpkhuong: I blame your browser.
pkhuongyeah, 2 is really nice.
nyefThe thing about 4 is that the blocks are too small, and they're all broken up.
jsnellok, thanks. that looks pretty unanimous. though not the answer I was expecting
nyefThe thing about 3 is that the padding out to the right edge disrupts the reading field.
The thing about 1 is that the blocks padding out to the -left- edge also disrupts the visual flow.
Which answer were you expecting?
jsnellI was expecting people to like 3 more than 2, due to not having an ugly saw-tooth on the right
nyefNo, the "ugly saw-tooth" is important information for finding the right edge of the code.

Page: 2 9 16 23 30