1. errors in ex.err
I've been getting less than useful error msgs lately, from an app that has run
for years fairly flawlessly (except for the strange errors i reported and got
laughed at here, and so i restart the app when i detect them), but this one i
don't see how i can detect in advance:
<from ex.err>
================
u:\newsgetter\Eu webgetter tcp4u\Getabcnews4.exw:478
slice lower index is less than 1 (-4)
u:\newsgetter\Eu webgetter tcp4u\Getabcnews4.exw:478 if length(href)
and not equal(href,"") and
href = {47'/'}
===========================
</from>
The entire if-then code is:
if length(href) and not equal(href,"") and
(not match("/REUTERS",upper(href))) and
(not match("/AP",upper(href))) and
(
(not match(".HTML",upper(href[length(href)-5..length(href)]))) and
(not match(".HTM", upper(href[length(href)-4..length(href)])))
)
then href = ""
end if
As you can see, i FIRST test to see if there is anything at all in href, and
that
is where Eu is reporting i made a mistake. Naturally, i am now testing to see
if (length(href) > 5) .
Kat
2. Re: errors in ex.err
Kat wrote:
>
> I've been getting less than useful error msgs lately, from an app that has run
>
> for years fairly flawlessly (except for the strange errors i reported and got
> laughed at here, and so i restart the app when i detect them), but this one i
> don't see how i can detect in advance:
>
> <from ex.err>
> ================
> u:\newsgetter\Eu webgetter tcp4u\Getabcnews4.exw:478
> slice lower index is less than 1 (-4)
>
> u:\newsgetter\Eu webgetter tcp4u\Getabcnews4.exw:478 if length(href)
> and not equal(href,"") and
>
> href = {47'/'}
>
> ===========================
> </from>
>
> The entire if-then code is:
> }}}
<eucode>
> if length(href) and not equal(href,"") and
> (not match("/REUTERS",upper(href))) and
> (not match("/AP",upper(href))) and
> (
> (not match(".HTML",upper(href[length(href)-5..length(href)]))) and
> (not match(".HTM", upper(href[length(href)-4..length(href)])))
> )
> then href = ""
> end if
> </eucode>
{{{
>
> As you can see, i FIRST test to see if there is anything at all in href, and
> that
> is where Eu is reporting i made a mistake. Naturally, i am now testing to see
> if (length(href) > 5) .
Consider mastering regular expressions. If you are going to work
with text (especially in the Internet world), they are more than
worth the effort. Learn PERL compatible regular expressions,
such as, in the PCRE library which Bochert has provided as a
user contribution, in Javascript, in many other apps, and of
course in PERL.
I understand that href, in your code, will always be a
sequence/string.
One if statement using a regular expression match and a pattern
like the following one will work better than your large compound
if statement, and you don't have to worry about the length of
href.
Example regular expression pattern to use:
"(/REUTERS)|(/AP)|(.HTML$)|(.HTM$)"
I would set the case sensitivity option/switch to do being case
insensitive. Then (using Bochert's library) use RGXscan() (or
RGXfind()) with the pattern and RGXcnt in a single if statement.
I know this is not a quick fix/help. But, in the long term, you would
have a much more useful and powerful tool at your disposal. One
that will save your hours of programming time and many headaches.
If you want some help getting started with regular expressions,
email me individually.
Terry Constant
3. Re: errors in ex.err
On 9 Oct 2004, at 12:06, Terry Constant wrote:
>
>
> posted by: Terry Constant <EUforum at terryconstant.com>
<snip>
> Consider mastering regular expressions. If you are going to work
> with text (especially in the Internet world), they are more than
> worth the effort. Learn PERL compatible regular expressions,
> such as, in the PCRE library which Bochert has provided as a
> user contribution, in Javascript, in many other apps, and of
> course in PERL.
But using regexp will always be slower too. I do use regexp in other
languages, such as:
\b(f|ph)[ ]*[ucs][ ]*[cu][ ]*[k<]\b
and
\bm+(o+|u+|0+)t+h+(a+|e+|3+)(r+|)[\s]*(p+l+|p+h+|p+|f+|h+|)(a+|u+|o+|0+)(c+k
+|w+k+|r+k+|k+|c+)[\w]*
Like i said, the code i gave to explain the ex.err has been running for years,
and little oddities are creeping up, and it's a tangle of if-then since we have
no clean goto, and i am rewriting it. Until i am done, the old code is called.
even if we had a goto, it wouldn't help the ex.err reporting that error occurred
on that line, as that line did no subscripting.
Kat
4. Re: errors in ex.err
- Posted by Terry Constant <EUforum at terryconstant.com>
Oct 09, 2004
-
Last edited Oct 10, 2004
Kat wrote:
>it wouldn't help the ex.err reporting that error occurred
>on that line, as that line did no subscripting.
Probably, the error points to the if statement and not to the line.
The if statement starts on that line, even though the subscripting
occurs on another text line in a single if statement. That is, doesn't the
error info just point to the EU statement where an error occurs and not
to the specific text line within a statement that traverses multiple
lines?
Terry Constant
5. Re: errors in ex.err
- Posted by Robert Craig <rds at RapidEuphoria.com>
Oct 09, 2004
-
Last edited Oct 10, 2004
Terry Constant wrote:
> Probably, the error points to the if statement and not to the line.
> The if statement starts on that line, even though the subscripting
> occurs on another text line in a single if statement. That is, doesn't the
> error info just point to the EU statement where an error occurs and not
> to the specific text line within a statement that traverses multiple
> lines?
That's correct.
I don't keep enough information around at run-time
to be able to point to a specific line within a
multi-line statement. On the other hand, with
compile-time (syntax) errors, I can point at the
specific line and the approximate position within the line.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
6. Re: errors in ex.err
- Posted by Pete Lomax <petelomax at blueyonder.co.uk>
Oct 09, 2004
-
Last edited Oct 10, 2004
On Sat, 9 Oct 2004 12:44:01 -0500, Kat <gertie at visionsix.com> wrote:
>but this one i don't see how i can detect in advance:
Can I first knock one thing on the head:
>}}}
<eucode>
> if length(href) and not equal(href,"") and
Is that not tautology, testing the same thing twice?
(which in itself is tautological 
) Anyway, that off my chest.
Can I secondly assume that the error message missing the actual line
of the sub-expression is perhaps misleading, but not quite the point.
<aside>
To be fair, I've been flummoxed before by:
if equal(a[x..y],stuff) then <blah>..[p..q] end if
and been equally baffled that x and y *are* correct, before
finally noticing that it is one of p or q that is wrong.
These things happen.
</aside>
>Naturally, i am now testing to see if (length(href) > 5) .
Am I right in guessing that what you really meant to say was that
changing the test that way stops it crashing, but allows stuff like
"/APxx" and "b.htm" pass by?
Regards,
Pete
7. Re: errors in ex.err
- Posted by "Juergen Luethje" <j.lue at gmx.de>
Oct 09, 2004
-
Last edited Oct 10, 2004
Pete Lomax wrote:
> On Sat, 9 Oct 2004 12:44:01 -0500, Kat <gertie at visionsix.com> wrote:
>
>> but this one i don't see how i can detect in advance:
>
> Can I first knock one thing on the head:
>> }}}
<eucode>
>> if length(href) and not equal(href,"") and
>
> Is that not tautology, testing the same thing twice?
> (which in itself is tautological ) Anyway, that off my chest.
Of course, it is.
<snip>
Regards,
Juergen
8. Re: errors in ex.err
Kat wrote:
>
> I've been getting less than useful error msgs lately, from an app that has run
>
> for years fairly flawlessly (except for the strange errors i reported and got
> laughed at here, and so i restart the app when i detect them), but this one i
> don't see how i can detect in advance:
>
> <from ex.err>
> ================
> u:\newsgetter\Eu webgetter tcp4u\Getabcnews4.exw:478
> slice lower index is less than 1 (-4)
>
> u:\newsgetter\Eu webgetter tcp4u\Getabcnews4.exw:478 if length(href)
> and not equal(href,"") and
>
> href = {47'/'}
>
> ===========================
> </from>
>
> The entire if-then code is:
> }}}
<eucode>
> if length(href) and not equal(href,"") and
> (not match("/REUTERS",upper(href))) and
> (not match("/AP",upper(href))) and
> (
> (not match(".HTML",upper(href[length(href)-5..length(href)]))) and
> (not match(".HTM", upper(href[length(href)-4..length(href)])))
> )
> then href = ""
> end if
> </eucode>
{{{
I assume you are testing for URLs that contain either /Reuters or /AP,
and end in either .HTML or .HTM, and you want to do it quickly, and
you want ex.err to report the more relevant line.
If so, this might also meet your requirements...
if length(href) < 7 then
href = ""
else
HREF = upper(href)
if match("/REUTERS",HREF) = 0 or match("/AP",HREF) = 0 then
href = ""
else
-- Just examine last 7 chars.
FERH = reverse(HREF[length(HREF)-6..length(HREF)])
if match("LMTH.",FREH) != 1 and match("MTH.", FREH) != 1 then
href = ""
end if
end if
end if
--
Derek Parnell
Melbourne, Australia
