FPC: runtime error 87 - What is it? 
Author Message
 FPC: runtime error 87 - What is it?


Quote:

> I have a program that runs in BP 7, and when I compile it with the
> win32 FPC I get an I/O error 87.

According to errors.pp of the rtl/win32 dir, this means
"ERROR_INVALID_PARAMETER = 87;" (it seems like a bug in the compiler
and/or RTL if you get this message).

Could you give a small example program where the problem occurs? Also,
what compiler version are you using? If possible, please try with the
latest snapshot (available in the development section of the website).

Jonas



Wed, 05 Dec 2001 03:00:00 GMT  
 FPC: runtime error 87 - What is it?

Quote:

> Explored the FPC site a little more, and noticed the Write bugs were
> fixed in version 99.11 or 99.12.  I downloaded the latest win32
> snapshot and it compiles ok.  Seems to me that 6 months is too long to
> wait for an updated release version.  I didn't try the snapshot right

You're right. 0.99.12 (the next release) should've been released
already quite a few weeks, but it keeps getting pushed back because of
additional bugfixes. It's hard to draw a line and say: this should be
good enough, even if there are some known issues left.

That's why we started with the snapshots. The harsh wording regarding
no support etc is simply to let people know that those aren't
necessarily stable versions of the compiler (so we don't get a bad
name). If you report a bug with a snapshot, it's checked just as
thourough as a bug reported regarding a release version.

Quote:
> away because of the dire warnings and copouts about using a non-
> supported snapshot.  Something as important and basic as a Write
> compiler bug should be fixed right away IMHO and the release version
> updated.

You have a point here. We do keep a "fixes" branch after every release
in which only bugs are fixed, but it's rarely released, unless there
was a really blatant bug (like not being able to compile anything if a
mouse driver is loaded or something like that... No, this hasn't
happened yet :)

I've forwarded your comment to the other developers.

Quote:
> Now that I know how the FPC 'system' works, I will be trying out the
> snapshots when I run into any more problems.

Snapshots are indeed your best friend. These compiler/RTL packages are
at least able to compile the source tree three time in a row, so they
should always be usable.

Jonas



Fri, 07 Dec 2001 03:00:00 GMT  
 FPC: runtime error 87 - What is it?

Quote:
> If I were you guys I would try to get the compiler stable enough for
> ordinary Pascal programs, with records and pointers, and I/O.  That
> way you would get a following from students and medium experienced
> (like myself) Pascal programmers.  When it comes to Borland Objects,
> and Inheritance (the {*filter*} stuff), those fixes could wait a few
> months for fixing, besides probably being harder to fix.

> The Write(ln) bug (for example) makes the students start looking for a
> pirated version of TP, when FPC is a very good product for students.
> Students also make a good user base, and will keep using FPC as it is
> improved.

> Since coming back to BP and FPC to write a quick and dirty dos
> utility, I have seen there are a lot of students and 'old' Pascal
> programmers out there, that want a good Pascal compiler, and FPC fills
> the bill.  Only one problem looming <g>.  Rumours of a BP8 and TP8.  I
> believe there is a market for a $49 dollar TP 32 bit compiler.

hmm... that's not a nice thing to say :(
Besides, does bp8 provide platform independency ? FPC isn't only a dos compiler
!
Also the "the {*filter*} stuff" you are referring to is something most programmers,
like me, find very important.

SnipeZ



Fri, 07 Dec 2001 03:00:00 GMT  
 FPC: runtime error 87 - What is it?
I have absoulutely co complaint with the speed at which the FPC
compilers are being released.

I use the DOS/DMPI , WIN32 and OS/2 versions and so far have had only
minor problems, mostly related to how FPC handles sets.
The IDE however is another story :)

To the FPC developers: Great Job Guys! You're putting Borland/Inprise
to shame!


Quote:


>:  > If I were you guys I would try to get the compiler stable enough for
>:  > ordinary Pascal programs, with records and pointers, and I/O.  That
>:  > way you would get a following from students and medium experienced
>:  > (like myself) Pascal programmers.  When it comes to Borland Objects,
>:  > and Inheritance (the {*filter*} stuff), those fixes could wait a few
>:  > months for fixing, besides probably being harder to fix.
>:  >
>:  > The Write(ln) bug (for example) makes the students start looking for a
>:  > pirated version of TP, when FPC is a very good product for students.
>:  > Students also make a good user base, and will keep using FPC as it is
>:  > improved.
>:  >
>:  > Since coming back to BP and FPC to write a quick and dirty dos
>:  > utility, I have seen there are a lot of students and 'old' Pascal
>:  > programmers out there, that want a good Pascal compiler, and FPC fills
>:  > the bill.  Only one problem looming <g>.  Rumours of a BP8 and TP8.  I
>:  > believe there is a market for a $49 dollar TP 32 bit compiler.
>:  
>:  hmm... that's not a nice thing to say :(
>:  Besides, does bp8 provide platform independency ? FPC isn't only a dos compiler
>:  !
>:  Also the "the {*filter*} stuff" you are referring to is something most programmers,
>:  like me, find very important.
>:  
>:  SnipeZ
>:  

>I wasn't trying to say it wasn't important, but FPC should be
>original Pascal complient (no bugs), and the 'new' {*filter*} stuff takes
>longer to fix and implement.  I am referring to the original Pascal as
>defined by Nicky, and the first TP versions to make it useful.  This
>is what is taught in the first few semesters of Pascal programming (or
>used to be).  The basics should all work all the time IMHO in any
>version of Pascal.  If a basic function or procedure is broken, the
>first priority should be to fix that immediately.  The Object stuff
>can wait to make sure it is fixed correctly, which because of its
>nature is probably much harder to fix.  Oh.. ALso I don't think 'most'
>programmers find it (objects) that important.  I would say a large
>'minority' but not most.



Sat, 08 Dec 2001 03:00:00 GMT  
 FPC: runtime error 87 - What is it?

Quote:

> If I were you guys I would try to get the compiler stable enough for
> ordinary Pascal programs, with records and pointers, and I/O.  That

Normally, it should be stable for that. But that really isn't enough.
People are using FPC to write a replacement for Delphi under Linux and
they really need a lot of the {*filter*} stuff.

Quote:
> way you would get a following from students and medium experienced
> (like myself) Pascal programmers.  When it comes to Borland Objects,
> and Inheritance (the {*filter*} stuff), those fixes could wait a few
> months for fixing, besides probably being harder to fix.

As explained above, that's not true. This would cause a lot of people
to get angry, and besides, some of us (not me) are heavily into Delphi
programming and work almost exclusively on the Delphi compatibility.
Since we are all volunteers doing this in our spare time, we all
enhance the parts that we think that is important.

It's not that we're working against eachother of course and there are
common goals: make sure it stays Pascal what we support, as little
dirty hacks as possible, try to maintain TP/Delphi compatibility as
much as possible as long as it's not too limiting. Yet, in the end we
all work on it because we like to do it.

Quote:
> The Write(ln) bug (for example) makes the students start looking for a
> pirated version of TP, when FPC is a very good product for students.  
> Students also make a good user base, and will keep using FPC as it is
> improved.

I agree the writeln bug should have resulted in a re-release after it
had been fixed.

Quote:
> Since coming back to BP and FPC to write a quick and dirty dos
> utility, I have seen there are a lot of students and 'old' Pascal
> programmers out there, that want a good Pascal compiler, and FPC fills
> the bill.  Only one problem looming <g>.  Rumours of a BP8 and TP8.  I
> believe there is a market for a $49 dollar TP 32 bit compiler.

It's a long time ago I heard rumors of TP8. Regarding the $49 market:
we won't tap into it, that's for someone else to exploit :) (if they're
willing)

Jonas



Sat, 08 Dec 2001 03:00:00 GMT  
 
 [ 5 post ] 

 Relevant Pages 

1. FPC: runtime error 215 - lack in documentation....

2. Source of EM86/87

3. fast '87 functions

4. *87 built-in matrix operations?

5. *87 built-in matrix operations?

6. Accessing dbase II+ and/or Clipper (Summer 87) files

7. Heap Overflow Error - runtime error 203

8. FPC linking error under Linux ... cannot find GLU

9. error message in FPC

10. FPC:problem with printing from fpc programs

11. FPC: translate execvp from c to FPC

12. Confusing runtime error porting to TPW

 

 
Powered by phpBB® Forum Software