Re: The 0.18.0 release and Repository Rules

On Wednesday, January 29, 2003, at 03:34  AM, Stefano Corsi wrote:

> Alle 04:48, mercoledì 29 gennaio 2003, David Hakim ha scritto:
>> Well, it looks like the different sub-projects of the 0.18.0 release
>> are coming to a close. If everyone working on a sub-project could 
>> email
>> me where things stand I'll make a decision as to whether to declare a
>> freeze wednesday night and start tagging and building betas. If all
>> goes well we should be able to start putting out betas by thursday or
>> friday.
>
> - Comma operator [committed]
> - Inline "for" declaration [committed]
> - "-O" flag bug fix [committed]
> - Inline "for" declaration in loop without increment part [committed]
> - Array initialization [sent patch]->[received your change]
> This last subproject is in draft state. The tests works, but as I 
> wrote you,
> there are some dirty stuff to clean and some optimization to do. I 
> will now
> integrate your changes to moto.y and start cleaning up a bit. I have 
> created
> a branch sc_array_init_dev to work on this. We can then clean it up 
> later on.
>
> About type verification:
>
> This gets an error (right behaviour):
>
>    String [] ints = {"1", "2", 3};
>    for (i = 0; i < length(ints); i++) {
>     print ints[i] + " ";
>    }
>
> tests/array_init.moto:54: Variable of type <String[]> cannot store type
> <int[]>
>
> This dies with a MotoException (not handled, so no output)

Actually it is handled (or at least caught :) ) the fact that there is 
no output is the bug :) I saw that in motofn.c the other day ... Ah 
well, another one for the bug list.

> when it  tries to
> setVarVal the second initializer.
>
>    int strings[] = {1,"2",3};
>    for (i = 0; i < length(strings); i++) {
>       print strings[i] + " ";
>    }

I sent you an email about these issues just a minute ago ... I should 
have read this one first :) I've merged your branch to the main branch 
and I believe I have taken care of the casting issues in motoi and 
motov. I've left motoc as an exercise for the reader :)

	Its come to my attention that the repository hasn't been in the best
>> of shape in the last few days. This is probably my fault :) .
>
> I have noticed a lot of test fail (memory) after merging with your 
> latest
> changes in the repo. Is it a known behaviour or is it a mistake from 
> my side?
> Probably you are working hard on something "inner" during these 
> hours...
>
Sorry :) This should be all taken care of at this point.

>> 	2) Make sure that everything builds on your local machine (after a
>> distclean all the way through to make install) before you commit. If
>> you have a workarea on webcodex, you might try to test compile your
>> changes there as well. Webcodex uses gcc 2.96, an older version that
>> isn't as ... forgiving :) about certain things.
>
> How can I configure moto on webcodex? On my machine I have installed
> everything as "root" with prefix=/usr and then I compile as "devel" 
> with the
> same prefix without installing. Then I run make test as "devel". How 
> can I do
> the same on webcodex?
>
> ./configure --prefix=?????
>
Configuring --prefix to somewhere in either your home directory or 
workarea (/usr/local/wa/<username>) should work. However 'make test' 
should run correctly even if make install has never been done.

-Dave
> Stefano
>
>