So, as usual, I made a _minor_ modification to a working file and suddenly all
Hell broke loose.

At the end of the file I added a text object, and everything was fine except for
the orientation.

So I added two simple vector declarations in my camera section,
and then some calculations on those vectors just before the text object.

Suddenly, in the untouched MIDDLE of the file, POV-Ray's telling me that I'm
using > version 3.5 and I need a semicolon.   Great - but it's highlighting a
line of code that doesn't have anything to do with that (the #end of a for loop
in the #declare AllStrands union)

Huh?

So I comment out the vectors declarations, move them to the bottom of the scene
file, run the scene in 3.8-alpha, 3.71-beta-whatever, 3.7 official, ... same
thing, only now it's complaining about a perfectly formed #declare identifier =
value; a few lines down that's worked perfectly for dozens of test renders.
[though 3.7 official complained about a trailing comma in my sor{} defined by a
loop]

(#declare Ymax = 0;)
                  ^ LOOK, you confounded Parser!  There's your precious
semicolon!

I'm perplexed, confused, and possibly cursed.

It was late, I had a long, sweltering day, and the eyelids were drooping, so I
decided not to stay up until midnight hunting phantoms.

Perhaps someone else could just run the code to confirm this bizarre behaviour
and move some lines (at the end near the text block) around to see if the error
notification changes.  #declare Loc and LA were what I added at the top and then
moved to the bottom.

Pay no attention to the unattributed stolen texture and hdr lighting code behind
the curtain...

Post a reply to this message

Attachments:
Download 'bellhenge.pov.txt' (12 KB)

"Bald Eagle" <cre### [at] netscapenet> wrote:
>
> Suddenly, in the untouched MIDDLE of the file, POV-Ray's telling me that I'm
> using > version 3.5 and I need a semicolon.   Great - but it's highlighting a
> line of code that doesn't have anything to do with that (the #end of a for loop
> in the #declare AllStrands union)
>
> So I comment out the vectors declarations, move them to the bottom of the scene
> file, run the scene in 3.8-alpha, 3.71-beta-whatever, 3.7 official, ... same
> thing, only now it's complaining about a perfectly formed #declare identifier =
> value; a few lines down that's worked perfectly for dozens of test renders.
> [though 3.7 official complained about a trailing comma in my sor{} defined by a
> loop]
>
> (#declare Ymax = 0;)
>                   ^ LOOK, you confounded Parser!  There's your precious
> semicolon!
>
> I'm perplexed, confused, and possibly cursed.
>
> It was late, I had a long, sweltering day, and the eyelids were drooping, so I
> decided not to stay up until midnight hunting phantoms.
>
> Perhaps someone else could just run the code to confirm this bizarre behaviour
> and move some lines (at the end near the text block) around to see if the error
> notification changes.  #declare Loc and LA were what I added at the top and then
> moved to the bottom.

Sorry that I can't check into this anymore right now but I did find that same
thing about the sor object getting an error on the Newspline, (comma, no ending
}).
Which is you probably know got a change for the latest versions to allow that.

After removing that comma the memory usage goes "haywire" too, making for a
difficult time forcing a close of POV-Ray 3.7.0. Have not tried 3.8 alpha.

Did see that version message anyhow. Parser sure doesn't like something there!

Post a reply to this message

On 26/09/2017 21:49, omniverse wrote:
> Sorry that I can't check into this anymore right now but I did find that same
> thing about the sor object getting an error on the Newspline, (comma, no ending
> }).


Me too.

-- 

Regards
     Stephen

Post a reply to this message

"omniverse" <omn### [at] charternet> wrote:

> Sorry that I can't check into this anymore right now but I did find that same
> thing about the sor object getting an error on the Newspline, (comma, no ending
> }).
> Which is you probably know got a change for the latest versions to allow that.

Right, I figured it must have been a fix that got added post 3.7-official.

> After removing that comma the memory usage goes "haywire" too, making for a
> difficult time forcing a close of POV-Ray 3.7.0. Have not tried 3.8 alpha.

That could be a result of the undulating spheres I use for the clapper rope.
Ridiculous, but it was a quick cut & paste solution.
I'm running on 16 gig, but before I trimmed it down far enough I crashed it a
few times too.

> Did see that version message anyhow. Parser sure doesn't like something there!

Well, I wish I could have looked it over during the day, but the workbench with
the laptop is buried under piles of - you name it.

Got back home, did a full file search for "#declare" and found 2 lines without
semicolons (419 & 420).  Not sure why that would trigger an error higher up in
the file, but yes, something weird was going on.
Anyway, I'm back on track trying to hammer out something suitable.

Thanks for taking that quick look  :)

Post a reply to this message