1. type test broken
I am feeding a sequence to some code:
data = parse("AbCdEF",{"si",'c'})
global function parse(sequence s, object c)
if sequence(c) then
str = 0 -- this is executed
end if
if atom(c)
then
--code -- not executed
else
-- code -- not executed
end if
So why won't the "if atom(c)... else" code execute?
Kat
2. Re: type test broken
Kat wrote:
>
> I am feeding a sequence to some code:
>
> }}}
<eucode>
> data = parse("AbCdEF",{"si",'c'})
>
> global function parse(sequence s, object c)
>
> if sequence(c) then
> str = 0 -- this is executed
> end if
>
> if atom(c)
> then
> --code -- not executed
> else
> -- code -- not executed
> end if
> </eucode>
{{{
>
> So why won't the "if atom(c)... else" code execute?
>
Could you post the full function (or at least a full demo that exhibits this
behavior)? I've done stuff like this, and found I overlooked a return or
something, or had incorrectly nested if statements.
Matt Lewis
3. Re: type test broken
- Posted by ags <eu at 531pi.co.nz>
Jul 11, 2005
Kat wrote:
>
> I am feeding a sequence to some code:
>
data = parse("AbCdEF",{"si",'c'})
global function parse(sequence s, object c)
if sequence(c) then
str = 0 -- this is executed
end if
if atom(c)
then
--code -- not executed
else
-- code -- not executed
end if
>
> So why won't the "if atom(c)... else" code execute?
>
> Kat
>
>
if "AbCdEF",{"si",'c'} is the actual data you are basing this on then
the "if atom(c)" will fail (since {"si",'c'} is a sequence) but it should at
least _do_ the test (let me guess, this is the program you can't trace? :)
My advice, if you're not being paid for it, sleep on it :)
Gary
4. Re: type test broken
On 11 Jul 2005, at 3:49, Matt Lewis wrote:
>
>
> posted by: Matt Lewis <matthewwalkerlewis at gmail.com>
>
> Kat wrote:
> >
> > I am feeding a sequence to some code:
> >
> > }}}
<eucode>
> > data = parse("AbCdEF",{"si",'c'})
> >
> > global function parse(sequence s, object c)
> >
> > if sequence(c) then
> > str = 0 -- this is executed
> > end if
> >
> > if atom(c)
> > then
> > --code -- not executed
> > else
> > -- code -- not executed
> > end if
> > </eucode>
{{{
> >
> > So why won't the "if atom(c)... else" code execute?
> >
>
> Could you post the full function (or at least a full demo that exhibits this
> behavior)? I've done stuff like this, and found I overlooked a return or
> something, or had incorrectly nested if statements.
I added the "if sequence(c)" to strtok-v2-1.e while trying to find why calling
parse() didn't work properly. I called thusly:
include get.e -- for wait_key()
include wildcard.e -- for upper()
include Strtok-v2-2.e
include file.e
with trace
object junk, data, writefile
trace(1)
data = parse("AbCdEF",{"si",'c'})
writefile = open("H:\\DataMinerCode\\minenews\\newgetter-wget\\wget\\tmp2-
parsed.txt","w")
for loop = 1 to length(data) do
puts(writefile,data[loop]&"\n====================================
====================================\n")
end for
close(writefile)
--junk = wait_key()
I have since rebooted the machine, no joy. I also tried the interpreter in
several project directories, in case one of them had been corrupted, no joy. I
also tried both exw.exe and exw40.exe, no joy.
Kat
5. Re: type test broken
Hi again Kat,
A guess would be that you're modifying one file and your program
is using another different file. Check to see if you have other
files that are named the same in your 'include' directories.
include stringstuff.e --Where exactly is this file? Are there two?
--Is my main program reading the right one or
--using one from my current directory?
Take care,
Al
And, good luck with your Euphoria programming!
My bumper sticker: "I brake for LED's"
6. Re: type test broken
Kat wrote:
>
>
> I added the "if sequence(c)" to strtok-v2-1.e while trying to find why calling
>
> parse() didn't work properly. I called thusly:
>
> }}}
<eucode>
> include get.e -- for wait_key()
> include wildcard.e -- for upper()
> include Strtok-v2-2.e
> include file.e
> with trace
>
> object junk, data, writefile
>
> trace(1)
> data = parse("AbCdEF",{"si",'c'})
>
> writefile = open("H:\\DataMinerCode\\minenews\\newgetter-wget\\wget\\tmp2-
> parsed.txt","w")
>
> for loop = 1 to length(data) do
>
> puts(writefile,data[loop]&"\n====================================
> ====================================\n")
> end for
> close(writefile)
> --junk = wait_key()
> <font color="#330033"></eucode>
{{{
</font>
>
> I have since rebooted the machine, no joy. I also tried the interpreter in
> several project directories, in case one of them had been corrupted, no joy. I
>
> also tried both exw.exe and exw40.exe, no joy.
I only have strtok v2.1 (maybe 2.2 is unreleased, couldn't find any
reference to it other than this post). I modified it a bit by adding
the test for sequence (str isn't here, so I just did a "? 0"). I also
reformatted some of the code in the big if block where you test for an
atom. I think the problem was that the "while" was on the same line
as "else" and the tracer skips right over them. If you move the while
to its own line, you can see that the while condition is evaluated,
and just skips over it.
I did a dissassembly of the il code using ooeu, and the reason for this
I think is that an else doesn't get a STARTLINE opcode, which is what
the tracer looks for, and since it's not a new line, the while must not
generate it. So I think that the code is executing properly, but just
looks like it isn't because of the formatting and the peculiarities with
the tracer.
global function parse(sequence s, object c) -- "object" by kat
integer slen, spt, flag, keep, case
sequence parsed, upperc, uppers
keep = 0
case = 0
upperc = ""
uppers = ""
if sequence(c) then
? 0
end if
if atom(c) then -- kat
c = {c}
else
while equal(c[1],"k") or equal(c[1],"i") do
if equal(c[1],"k") then
keep = 1
c = c[2..length(c)]
end if
if equal(c[1],"i") then
case = 1
c = c[2..length(c)]
end if
end while
end if -- kat
Matt Lewis
7. Re: type test broken
On 11 Jul 2005, at 5:31, Matt Lewis wrote:
>
>
> posted by: Matt Lewis <matthewwalkerlewis at gmail.com>
>
> Kat wrote:
> >
> >
> > I added the "if sequence(c)" to strtok-v2-1.e while trying to find why
> > calling
> > parse() didn't work properly. I called thusly:
> >
> > }}}
<eucode>
> > include get.e -- for wait_key()
> > include wildcard.e -- for upper()
> > include Strtok-v2-2.e
> > include file.e
> > with trace
> >
> > object junk, data, writefile
> >
> > trace(1)
> > data = parse("AbCdEF",{"si",'c'})
> >
> > writefile = open("H:\\DataMinerCode\\minenews\\newgetter-wget\\wget\\tmp2-
> > parsed.txt","w")
> >
> > for loop = 1 to length(data) do
> >
> > puts(writefile,data[loop]&"\n====================================
> > ====================================\n")
> > end for
> > close(writefile)
> > --junk = wait_key()
> > <font color="#330033"></eucode>
{{{
</font>
> >
> > I have since rebooted the machine, no joy. I also tried the interpreter in
> > several project directories, in case one of them had been corrupted, no joy.
> > I
> > also tried both exw.exe and exw40.exe, no joy.
>
> I only have strtok v2.1 (maybe 2.2 is unreleased, couldn't find any
> reference to it other than this post).
It *is* unreleased. I found this bug(?) and cannot get past it.
> I modified it a bit by adding
> the test for sequence (str isn't here, so I just did a "? 0"). I also
> reformatted some of the code in the big if block where you test for an
> atom. I think the problem was that the "while" was on the same line
> as "else" and the tracer skips right over them. If you move the while
> to its own line, you can see that the while condition is evaluated,
> and just skips over it.
If you watch the variable contents in the trace window, str should have been
set to 1 in the ..else.. code (see below where i added it). It isn't. Ergo, it
isn't
executing. I'll paste the entire parse code to this email (below), and the code
which calls parse() again.
> I did a dissassembly of the il code using ooeu, and the reason for this
> I think is that an else doesn't get a STARTLINE opcode, which is what
> the tracer looks for, and since it's not a new line, the while must not
> generate it. So I think that the code is executing properly, but just
> looks like it isn't because of the formatting and the peculiarities with
> the tracer.
> }}}
<eucode>
> global function parse(sequence s, object c) -- "object" by kat
> integer slen, spt, flag, keep, case
> sequence parsed, upperc, uppers
> keep = 0
> case = 0
> upperc = ""
> uppers = ""
>
> if sequence(c) then
> ? 0
> end if
>
> if atom(c) then -- kat
>
> c = {c}
> else
> while equal(c[1],"k") or equal(c[1],"i") do
put this code where it's obvious to be:
else while equal(c[1],"k") or equal(c[1],"i") or equal(c[1],"s")do
if equal(c[1],"s") then
str = 1
thestring = c[2]
lenc = length(thestring)
c = c[2..length(c)]
end if
> if equal(c[1],"k") then
> keep = 1
> c = c[2..length(c)]
> end if
> if equal(c[1],"i") then
> case = 1
> c = c[2..length(c)]
> end if
> end while
> end if -- kat
> </eucode>
{{{
--Andy Serpas Turbo version 3X faster
-- modified by Kat for case and keep-separator
-- getting modified more by Kat July 11 2005
-- yes, this is a mess, i am trying to add my parses() to parse()
-- and keep backwards compatability
global function parse(sequence s, object c) -- "object" by kat
integer slen, spt, flag, keep, case, str, lenc
sequence parsed, upperc, uppers, thestring, temp
keep = 0
case = 0
str = 0
upperc = ""
uppers = ""
parsed = {}
if sequence(c) then -- i added this *only* because the atom() failed
str = 0 -- if the trace window stops here, it's a seq
end if -- if a seq, then the else right below should run too
if atom(c) -- kat
then
c = {c}
temp = c
else while equal(c[1],"k") or equal(c[1],"i") or equal(c[1],"s")do
if equal(c[1],"s") then
str = 1 -- this does NOT happen, watch the variable dump!
thestring = c[2]
lenc = length(thestring)
c = c[2..length(c)]
end if
if equal(c[1],"k") then
keep = 1
c = c[2..length(c)]
end if
if equal(c[1],"i") then
case = 1
c = c[2..length(c)]
end if
end while
end if -- kat
if str then
if case
then temp = -lenc & find_alls(upper(thestring),upper(s)) & length(s)
else temp = -lenc & find_alls(thestring,s) & length(s)
end if
for loop = 1 to length(temp)-1 do
parsed &= {s[temp[loop]+lenc+1..temp[loop+1]-1]}
end for
return parsed
end if
if atom(c) -- kat
then c = {c}
end if
parsed = {}
slen = length(s)
spt = 1
flag = 0
if case = 0 then
for i = 1 to slen do
--if s[i] = c then -- kat
if find(s[i],c) then
if flag = 1 then
parsed = append(parsed,s[spt..i-1])
if keep then
parsed = append(parsed,{s[i]})
end if
flag = 0
spt = i+1
else
spt += 1
end if
else
flag = 1
end if
end for
if flag = 1 then
parsed = append(parsed,s[spt..slen])
end if
end if
if case = 1 then -- kat
upperc = upper(c)
uppers = upper(s)
for i = 1 to slen do
--if s[i] = c then -- kat
if find(uppers[i],upperc) then
if flag = 1 then
parsed = append(parsed,s[spt..i-1])
if keep then
parsed = append(parsed,{s[i]})
end if
flag = 0
spt = i+1
else
spt += 1
end if
else
flag = 1
end if
end for
if flag = 1 then
parsed = append(parsed,s[spt..slen])
end if
end if
return parsed
end function
include get.e -- for wait_key()
include wildcard.e -- for upper()
include Strtok-v2-2.e
include file.e
with trace
object junk, data, writefile
trace(1)
data = parse("AbCdEF",{"si",'c'})
writefile = open("H:\\DataMinerCode\\minenews\\newgetter-wget\\wget\\tmp2-
parsed.txt","w")
for loop = 1 to length(data) do
puts(writefile,data[loop]&"\n====================================
====================================\n")
end for
close(writefile)
--junk = wait_key()
8. Re: type test broken
On 11 Jul 2005, at 5:47, Al Getz wrote:
>
>
> posted by: Al Getz <Xaxo at aol.com>
>
> Hi again Kat,
>
>
> A guess would be that you're modifying one file and your program
> is using another different file. Check to see if you have other
> files that are named the same in your 'include' directories.
>
> include stringstuff.e --Where exactly is this file? Are there two?
> --Is my main program reading the right one or
> --using one from my current directory?
I can see the changed code in the Eu trace screen when i run it. There is
only one edition of the strtok*.2 file.
Kat
9. Re: type test broken
Kat wrote:
>
> On 11 Jul 2005, at 5:47, Al Getz wrote:
>
> >
> > posted by: Al Getz <Xaxo at aol.com>
> >
> > Hi again Kat,
> >
> >
> > A guess would be that you're modifying one file and your program
> > is using another different file. Check to see if you have other
> > files that are named the same in your 'include' directories.
> >
> > include stringstuff.e --Where exactly is this file? Are there two?
> > --Is my main program reading the right one or
> > --using one from my current directory?
>
> I can see the changed code in the Eu trace screen when i run it. There is
> only one edition of the strtok*.2 file.
>
> Kat
>
>
Hi Kat,
Oh ok, well i ran your code the way you posted it and it worked.
I printed ?1 if the test passed, and ?2 if it failed, and it
printed '2' because it wasnt an atom, so i was looking for
possibilities.
Take care,
Al
And, good luck with your Euphoria programming!
My bumper sticker: "I brake for LED's"
10. Re: type test broken
Kat wrote:
>
> If you watch the variable contents in the trace window, str should have been
> set to 1 in the ..else.. code (see below where i added it). It isn't. Ergo, it
> isn't
> executing. I'll paste the entire parse code to this email (below), and the
> code
> which calls parse() again.
<snip>
else while equal(c[1],"k") or equal(c[1],"i") or equal(c[1],"s")do
data = parse("AbCdEF",{"si",'c'})
Here's the problem. You've either constructed 'c' incorrectly, or your
tests are incorrect, because c[1] = "si". Perhaps it should be:
else while equal(c[1],"k") or match("i",c[1]) or match("s",c[1]) do
If you move the start of the while loop off of the same line as the else,
you can see that it does go into the else block, but the while tests fail.
Matt Lewis
11. Re: type test broken
On 11 Jul 2005, at 9:49, Matt Lewis wrote:
>
>
> posted by: Matt Lewis <matthewwalkerlewis at gmail.com>
>
> Kat wrote:
> >
> > If you watch the variable contents in the trace window, str should have been
> > set to 1 in the ..else.. code (see below where i added it). It isn't. Ergo,
> > it
> > isn't executing. I'll paste the entire parse code to this email (below), and
> > the code which calls parse() again.
>
> <snip>
>
> }}}
<eucode>
> else while equal(c[1],"k") or equal(c[1],"i") or equal(c[1],"s")do
>
> data = parse("AbCdEF",{"si",'c'})
> </eucode>
{{{
>
> Here's the problem. You've either constructed 'c' incorrectly, or your
> tests are incorrect, because c[1] = "si". Perhaps it should be:
> }}}
<eucode>
> else while equal(c[1],"k") or match("i",c[1]) or match("s",c[1]) do
> </eucode>
{{{
>
> If you move the start of the while loop off of the same line as the else,
> you can see that it does go into the else block, but the while tests fail.
Yes, i noticed that this morning too. It's another case of compounding the
error before submitting the question here. I'm sorry.
Kat
looking sheepish.
