Re: options.e in win32lib
Mike777 wrote:
>
> CChris wrote:
> >
> > Mike777 wrote:
> > >
> > > Not sure about this, but I was trying to use the getOpt() function and
> > > found
> > > it didn't work. Peeking inside options.e I *think* I found two places
> > > where
> > > changes could be right:
> > >
> > > First --------------------
> > >
> > > 153 if match(pCategory, lData) = 1 then
> > >
> > > should be
> > >
> > > 153 if match(pCategory, lower(lData)) = 1 then
> > >
> > > This is fairly obvious in that the most recently acquired lData (on line
> > > 147)
> > > hasn't yet been "lowered".
> > >
> >
> > Actually, I have changed this part so that several groups may start by the
> > same
> > letters, like [ Group ] and { \tGroup_2 ].
> > So I replaced
> > }}}
<eucode>
> > if match(pCategory, lData) = 1 then
> > lInCategory = w32True
> > end if
> > </eucode>
{{{
> > with
> > }}}
<eucode>
> > if match(pCategory, lower(lData)) = 1 then
> > for i=length(pCategory)+1 to length(lData) do
> > if lData[i]=' ' or lData[i]='\t' then -- keep scanning
> > else
> > lInCategory = (lData[i]=']')
> > exit
> > end if
> > end for
> > end if
> > </eucode>
{{{
>
> I like it. However, where did you "make the change"? I have downloaded the
> 12/15 Win32Lib and it doesn't appear to have been changed there.
>
I released 70.2e before reading your post, so the change isn't in 2e. It will be
in next release, hopefully official 70.3.
>
> > > Second ------------------
> > >
> > > 162 if length(lOptData) >= 2 and lOptData[1] = '"' and
> > > lOptData[length(lOptData)]
> > > = '"' then
> > > lData = lOptData[2..length(lOptData)-1]
> > > else
> > >
> > > might be replaced with:
> > >
> > > 162 if length(lOptData) >= 2 then
> > > if lOptData[1] = '"' and lOptData[length(lOptData)] = '"' then
> > > lData = lOptData[2..length(lOptData)-1]
> > > else
> > > lData = lOptData
> > > end if
> > > else
> > >
> > > This merely ensures that lData is reset to the correct part of lOptData if
> > > its
> > > length is greater than two, whether or not it starts and ends with a
> > > quotation
> > > mark. I know that doing what I've done obliterates the code beneath it
> > > that
> > > attempts to change numbers to text and attempts to change text input (yes,
> > > no,
> > > etc.) to corresponding w32 constants (w32True and w32False). Frankly, I'm
> > > not
> > > sure that is such a good idea, anyway, so I don't think it is much of a
> > > loss.
> > > For example, if I have an option with value "Tre rain in Spain if false."
> > > I
> > > really don't want it to be obliterated in favor of a w32False return
> > > value.
> > >
> > > I have my doubts about anybody actually using this function, since it
> > > appears
> > > to have been non-functional.
> > >
> > > Yes, I'm aware of the various libraries in the Archive that deal with ini
> > > files,
> > > but this seemed like a "simple" place to start. Famous last words, that!
> > >
> > > If somebody thinks I'm off-base on this issue, or how I raised it, please
> > > let
> > > me know (or you are likely to see more of the same in the near future).
> > >
> > > Mike
> >
> > I don't understand what the issue is.
>
> I didn't communicate it clearly enough.
>
> > What appears on the right side of the '=' sign may be either, after
> > collating
> > any continuation lines and stripping comments off:
> > 1/ a double quoted string (so the length is at least 2), in which case the
> > quotes
> > are stripped and the inner part returned;
> > 2/ a sequence of characters that fully converts to a number, in which case
> > the
> > number is returned. Currency signs are ignored.
> > 3/ anything else. In that case, the code tries to identify the sequence as
> > one
> > of special meaning:
> > 3a/ If some is recognised, its associated boolean value is returned
>
> I think 3a is my issue. I have a number of ini files that have long strings
> to the right of the equals sign and they are not enclosed in double quotes.
> If any part of that string has the characters "false" (ignore the double
> quotes
> for this purpose) then the return value is 0.
>
Not true. The relevant code is:
if find(lOptData,{"true","yes","on"}) then
lData = w32True
elsif find(lOptData,{"false","no","off"}) then
lData = w32False
end if
So, if your .ini line is
someVar = yes, I do
, the right hand side of '=' is not any of the litteral constants in the code
above, so "yes, I do" will be returned (outer whitespace is trimmed) as is.
find() and match() are different.
HTH
CChris
> Perhaps there is a standard for ini files somewhere and those who wrote these
> particular ini files are not conforming to the standards as written. No
> matter,
> I guess, now that I know the standard as intended, since I will be in charge
> of the ini files that my application uses, I can live with it.
>
> > 3b/ else lData=lOptData is returned.
>
> > The option you quoted has double quotes around it
>
> That was my miscommunication. I meant to imply that it would be those
> characters,
> without the double quotes. Sorry.
>
> , so the first part will apply.
> > Be sure to use double quotes if the option starts with a digit, but isn't to
> > be treated as a number
> > Could you post an example of behaviour you wouldn't expect or wish, but
> > actually
> > takes place?
>
> I think it will work, now.
>
> Thanks
>
> Mike
|
Not Categorized, Please Help
|
|
