Re: [racket-users] "Link: bad variable linkage" when trying to work through tutorial

[Matthias Felleisen]

How did you install Racket? Via the Mac installer from ‘download’? 

[This error happens when you had an older installation and the new one doesn’t 
re-compile the byte code files.]



> On May 13, 2017, at 5:19 PM, onn...@gmail.com wrote:
> 
> Hi all,
> 
> While working through the first bits of the HtDP2e book, in DrRacket 6.9 on 
> Mac, I get the following error whenever I try to run any program in the 
> "Definitions" window that contains (require 2htdp/universe):
> 
> link: bad variable linkage;
> reference to a variable that is not a procedure or structure-type constant 
> across all instantiations
>  reference phase level: 0
>  variable module: "/Applications/Racket 
> v6.9/share/pkgs/string-constants-lib/string-constants/string-constant.rkt"
>  variable phase: 0
>  reference in module: "/Applications/Racket 
> v6.9/share/pkgs/htdp-lib/2htdp/private/world.rkt"
> 
> Googling the error didn't help me out, and I'm afraid I'm too much of a noob 
> right now to understand what's going on. What's more, I'm really sure this 
> worked fine a few days ago. Restarting DrRacket doesn't help either.
> 
> Any suggestions would be most welcome! Thanks in advance :-)
> 
> Best regards,
> 
> 
> Onno
> 
> -- 
> You received this message because you are subscribed to the Google Groups 
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] "Link: bad variable linkage" when trying to work through tutorial

[Dmitry Pavlov]

Onno,

That happens when the .zo files produced by plain command-line racket conflict 
with ones that come from DrRacket. Or the other way around, I am not sure. Or 
something else related to .zo files that are no longer valid for some reason. 
Anyway, try removing the 'compiled' directories from your source tree and 
restarting.

Regards,

Dmitry

On May 14, 2017 12:19:22 AM EET, onn...@gmail.com wrote:
>Hi all,
>
>While working through the first bits of the HtDP2e book, in DrRacket
>6.9 on Mac, I get the following error whenever I try to run any program
>in the "Definitions" window that contains (require 2htdp/universe):
>
>link: bad variable linkage;
>reference to a variable that is not a procedure or structure-type
>constant across all instantiations
>  reference phase level: 0
>variable module: "/Applications/Racket
>v6.9/share/pkgs/string-constants-lib/string-constants/string-constant.rkt"
>  variable phase: 0
>reference in module: "/Applications/Racket
>v6.9/share/pkgs/htdp-lib/2htdp/private/world.rkt"
>
>Googling the error didn't help me out, and I'm afraid I'm too much of a
>noob right now to understand what's going on. What's more, I'm really
>sure this worked fine a few days ago. Restarting DrRacket doesn't help
>either.
>
>Any suggestions would be most welcome! Thanks in advance :-)
>
>Best regards,
>
>
>Onno
>
>-- 
>You received this message because you are subscribed to the Google
>Groups "Racket Users" group.
>To unsubscribe from this group and stop receiving emails from it, send
>an email to racket-users+unsubscr...@googlegroups.com.
>For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] link: bad variable linkage

[WarGrey Gyoudmon Ju]

I have this issue in Racket 6.8.0.1 this morning, the project is compiled
by compile-directory-zos, and all .zos are ensured up-to-date.
Before I gave up, I had wrote a simple tests in DrRacket, lucky to saw the
root cause, a contract error in the body of the newly added typed class,
but the issue was reported with two totally irrelevant modules.

On Sat, Feb 4, 2017 at 9:31 AM, Robby Findler 
wrote:

> This is the commit:
>
> https://github.com/racket/drracket/commit/81d74e484c306c3584030cf8ede298
> 25c0e8be1a
>
> I pushed it around the time 6.5 was released and it is the kind of
> thing I would do right after a release branch has been created, so
> probably it was only 6.6 and later.
>
> Robby
>
>
>
> On Fri, Feb 3, 2017 at 7:15 PM, Dan Liebgold
>  wrote:
> > On Saturday, March 26, 2016 at 4:53:51 PM UTC-7, Robby Findler wrote:
> >> Matthew and I have figured out one way in which DrRacket could go
> >> wrong here and implemented a better strategy. The problem we
> >> identified doesn't explain the symptoms expressed here in this thread
> >> exactly, but it is a complex system and maybe there was some piece
> >> missing from the explanations that I didn't think to ask about.
> >> Finger's crossed that it helps!
> >>
> >
> > I was getting this error and variations on it somewhat often in version
> 6.5. Do you think upgrading to 6.8 will help?
> >
> > Thanks,
> > Dan
> >
> > --
> > You received this message because you are subscribed to the Google
> Groups "Racket Users" group.
> > To unsubscribe from this group and stop receiving emails from it, send
> an email to racket-users+unsubscr...@googlegroups.com.
> > For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google Groups
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] link: bad variable linkage

[Robby Findler]

This is the commit:

https://github.com/racket/drracket/commit/81d74e484c306c3584030cf8ede29825c0e8be1a

I pushed it around the time 6.5 was released and it is the kind of
thing I would do right after a release branch has been created, so
probably it was only 6.6 and later.

Robby



On Fri, Feb 3, 2017 at 7:15 PM, Dan Liebgold
 wrote:
> On Saturday, March 26, 2016 at 4:53:51 PM UTC-7, Robby Findler wrote:
>> Matthew and I have figured out one way in which DrRacket could go
>> wrong here and implemented a better strategy. The problem we
>> identified doesn't explain the symptoms expressed here in this thread
>> exactly, but it is a complex system and maybe there was some piece
>> missing from the explanations that I didn't think to ask about.
>> Finger's crossed that it helps!
>>
>
> I was getting this error and variations on it somewhat often in version 6.5. 
> Do you think upgrading to 6.8 will help?
>
> Thanks,
> Dan
>
> --
> You received this message because you are subscribed to the Google Groups 
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] link: bad variable linkage

[Dan Liebgold]

On Saturday, March 26, 2016 at 4:53:51 PM UTC-7, Robby Findler wrote:
> Matthew and I have figured out one way in which DrRacket could go
> wrong here and implemented a better strategy. The problem we
> identified doesn't explain the symptoms expressed here in this thread
> exactly, but it is a complex system and maybe there was some piece
> missing from the explanations that I didn't think to ask about.
> Finger's crossed that it helps!
> 

I was getting this error and variations on it somewhat often in version 6.5. Do 
you think upgrading to 6.8 will help?

Thanks,
Dan

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] link: bad variable linkage

[jon stenerson]

Sorry I missed this. Apparently I'm several months behind on the racket 
list.  The problem does seem fixed however. Thank you!


Jon


On 10/23/2016 11:34 AM, Laurent wrote:
As of now, I haven't experienced any more problem of this sort since 
then. Admittedly I haven't tried hard either to reproduce this behavior.


So thank you so much for the time you took to look into this! Very 
appreciated.



On Sat, Mar 26, 2016 at 11:53 PM, Robby Findler 
> wrote:


Matthew and I have figured out one way in which DrRacket could go
wrong here and implemented a better strategy. The problem we
identified doesn't explain the symptoms expressed here in this thread
exactly, but it is a complex system and maybe there was some piece
missing from the explanations that I didn't think to ask about.
Finger's crossed that it helps!

In any case, if you are still experiencing the bad symptom after you
get the latest code, please let me know.

Thanks,
Robby

On Mon, Feb 1, 2016 at 9:03 AM, Laurent > wrote:
> I would be so happy if this kind of error could be resolved!
(even when all
> files are compiled by DrRacket)
> Usually, I just turn off "populate 'compiled' directories", but then
> DrRacket needs to compile the file on each F5, which can take
some precious
> time.
> Thank you both then :)
>
> On Sun, Jan 31, 2016 at 11:19 PM, jon stenerson
>
> wrote:
>>
>>
>> The simplest thing I can find: I have three files in two
directories.
>> The first directory, called junk, contains a.rkt and b.rkt.
>> The second directory is a sibling to junk called local-libs and
contains
>> c.rkt. (local-libs is known to the package manager; it was
installed via the
>> DrRacket interface).
>> If I open all three files in DrRacket and try running a.rkt,
sometimes it
>> works, sometimes not. By modifiying a, b, c in some order (in
the DrRacket
>> editor) I can toggle the work/not work flag. As you say, it's a
little
>> non-deterministic.
>>
>> Thanks,
>> Jon
>>
>>
>> a.rkt:
>> #lang racket
>> (require "b.rkt")
>>
>> b.rkt
>> #lang racket
>> (require "../local-libs/c.rkt")
>> (s 1)
>>
>> c.rkt
>> #lang racket
>> (provide (struct-out s) )
>> (struct s (p))
>>
>>
>>
>>
>> On 1/31/2016 4:03 PM, Robby Findler wrote:
>>>
>>> Then it would be helpful to us if you could provide some
(hopefully
>>> small) program and instructions to reproduce the problem.
>>>
>>> Thanks,
>>> Robby
>>>
>>>
>>> On Sun, Jan 31, 2016 at 4:51 PM, jon stenerson
>
>>> wrote:

 Just using DrRacket 6.3 on Win 10.

 Jon


 On 1/31/2016 3:22 PM, Robby Findler wrote:
>
> The compilation of racket is (not yet) deterministic so
things like
> that can throw off internal tables in the .zo files (and a
.zo files
> is loaded without checking dependencies, optimistically
hoping that
> the files are not changed).
>
> So: are you using "raco make" or somehow mangaging .zo files
yourself,
> or are you seeing this with the .zo files that DrRacket
maintains for
> your automatically? If the former, then you need to adapt your
> workflow to either not use .zo files or to keep them up to
date. If
> the latter, then it is a bug.
>
> Robby
>
>
>
> On Sun, Jan 31, 2016 at 4:17 PM, jon stenerson
> >
> wrote:
>>
>> I have some .rkt files and, depending on which file was
edited most
>> recently, either it works fine or it gives a compilation
error. (By
>> "edited"
>> I mean just adding or subtracting a blank line to change the
>> timestamp).
>> I
>> can submit a simplified example if nobody else has seen
this problem.
>> If
>> this is well-known, what is the fix?
>>
>> link: bad variable linkage;
>>reference to a variable that is not a procedure or
structure-type
>> constant
>> across all instantiations
>> reference phase level: 0
>> variable module: 
>> variable phase: 0
>> reference in module:  in: 
>>
>> --
>> You received this message because you are subscribed to the
Google
>> Groups
>> "Racket Users" group.
>> To unsubscribe from this group 

Re: [racket-users] link: bad variable linkage

[Laurent]

As of now, I haven't experienced any more problem of this sort since then.
Admittedly I haven't tried hard either to reproduce this behavior.

So thank you so much for the time you took to look into this! Very
appreciated.


On Sat, Mar 26, 2016 at 11:53 PM, Robby Findler  wrote:

> Matthew and I have figured out one way in which DrRacket could go
> wrong here and implemented a better strategy. The problem we
> identified doesn't explain the symptoms expressed here in this thread
> exactly, but it is a complex system and maybe there was some piece
> missing from the explanations that I didn't think to ask about.
> Finger's crossed that it helps!
>
> In any case, if you are still experiencing the bad symptom after you
> get the latest code, please let me know.
>
> Thanks,
> Robby
>
> On Mon, Feb 1, 2016 at 9:03 AM, Laurent  wrote:
> > I would be so happy if this kind of error could be resolved! (even when
> all
> > files are compiled by DrRacket)
> > Usually, I just turn off "populate 'compiled' directories", but then
> > DrRacket needs to compile the file on each F5, which can take some
> precious
> > time.
> > Thank you both then :)
> >
> > On Sun, Jan 31, 2016 at 11:19 PM, jon stenerson <
> jonstener...@comcast.net>
> > wrote:
> >>
> >>
> >> The simplest thing I can find: I have three files in two directories.
> >> The first directory, called junk, contains a.rkt and b.rkt.
> >> The second directory is a sibling to junk called local-libs and contains
> >> c.rkt. (local-libs is known to the package manager; it was installed
> via the
> >> DrRacket interface).
> >> If I open all three files in DrRacket and try running a.rkt, sometimes
> it
> >> works, sometimes not. By modifiying a, b, c in some order (in the
> DrRacket
> >> editor) I can toggle the work/not work flag. As you say, it's a little
> >> non-deterministic.
> >>
> >> Thanks,
> >> Jon
> >>
> >>
> >> a.rkt:
> >> #lang racket
> >> (require "b.rkt")
> >>
> >> b.rkt
> >> #lang racket
> >> (require "../local-libs/c.rkt")
> >> (s 1)
> >>
> >> c.rkt
> >> #lang racket
> >> (provide (struct-out s) )
> >> (struct s (p))
> >>
> >>
> >>
> >>
> >> On 1/31/2016 4:03 PM, Robby Findler wrote:
> >>>
> >>> Then it would be helpful to us if you could provide some (hopefully
> >>> small) program and instructions to reproduce the problem.
> >>>
> >>> Thanks,
> >>> Robby
> >>>
> >>>
> >>> On Sun, Jan 31, 2016 at 4:51 PM, jon stenerson <
> jonstener...@comcast.net>
> >>> wrote:
> 
>  Just using DrRacket 6.3 on Win 10.
> 
>  Jon
> 
> 
>  On 1/31/2016 3:22 PM, Robby Findler wrote:
> >
> > The compilation of racket is (not yet) deterministic so things like
> > that can throw off internal tables in the .zo files (and a .zo files
> > is loaded without checking dependencies, optimistically hoping that
> > the files are not changed).
> >
> > So: are you using "raco make" or somehow mangaging .zo files
> yourself,
> > or are you seeing this with the .zo files that DrRacket maintains for
> > your automatically? If the former, then you need to adapt your
> > workflow to either not use .zo files or to keep them up to date. If
> > the latter, then it is a bug.
> >
> > Robby
> >
> >
> >
> > On Sun, Jan 31, 2016 at 4:17 PM, jon stenerson
> > 
> > wrote:
> >>
> >> I have some .rkt files and, depending on which file was edited most
> >> recently, either it works fine or it gives a compilation error. (By
> >> "edited"
> >> I mean just adding or subtracting a blank line to change the
> >> timestamp).
> >> I
> >> can submit a simplified example if nobody else has seen this
> problem.
> >> If
> >> this is well-known, what is the fix?
> >>
> >> link: bad variable linkage;
> >>reference to a variable that is not a procedure or structure-type
> >> constant
> >> across all instantiations
> >> reference phase level: 0
> >> variable module: 
> >> variable phase: 0
> >> reference in module:  in: 
> >>
> >> --
> >> You received this message because you are subscribed to the Google
> >> Groups
> >> "Racket Users" group.
> >> To unsubscribe from this group and stop receiving emails from it,
> send
> >> an
> >> email to racket-users+unsubscr...@googlegroups.com.
> >> For more options, visit https://groups.google.com/d/optout.
> 
> 
>  --
>  You received this message because you are subscribed to the Google
>  Groups
>  "Racket Users" group.
>  To unsubscribe from this group and stop receiving emails from it, send
>  an
>  email to racket-users+unsubscr...@googlegroups.com.
>  For more options, visit https://groups.google.com/d/optout.
> >>
> >>
> >> --
> >> You received this message because you are subscribed to the Google
> Groups
> >> "Racket 

Re: [racket-users] link: bad variable linkage

[Robby Findler]

Matthew and I have figured out one way in which DrRacket could go
wrong here and implemented a better strategy. The problem we
identified doesn't explain the symptoms expressed here in this thread
exactly, but it is a complex system and maybe there was some piece
missing from the explanations that I didn't think to ask about.
Finger's crossed that it helps!

In any case, if you are still experiencing the bad symptom after you
get the latest code, please let me know.

Thanks,
Robby

On Mon, Feb 1, 2016 at 9:03 AM, Laurent  wrote:
> I would be so happy if this kind of error could be resolved! (even when all
> files are compiled by DrRacket)
> Usually, I just turn off "populate 'compiled' directories", but then
> DrRacket needs to compile the file on each F5, which can take some precious
> time.
> Thank you both then :)
>
> On Sun, Jan 31, 2016 at 11:19 PM, jon stenerson 
> wrote:
>>
>>
>> The simplest thing I can find: I have three files in two directories.
>> The first directory, called junk, contains a.rkt and b.rkt.
>> The second directory is a sibling to junk called local-libs and contains
>> c.rkt. (local-libs is known to the package manager; it was installed via the
>> DrRacket interface).
>> If I open all three files in DrRacket and try running a.rkt, sometimes it
>> works, sometimes not. By modifiying a, b, c in some order (in the DrRacket
>> editor) I can toggle the work/not work flag. As you say, it's a little
>> non-deterministic.
>>
>> Thanks,
>> Jon
>>
>>
>> a.rkt:
>> #lang racket
>> (require "b.rkt")
>>
>> b.rkt
>> #lang racket
>> (require "../local-libs/c.rkt")
>> (s 1)
>>
>> c.rkt
>> #lang racket
>> (provide (struct-out s) )
>> (struct s (p))
>>
>>
>>
>>
>> On 1/31/2016 4:03 PM, Robby Findler wrote:
>>>
>>> Then it would be helpful to us if you could provide some (hopefully
>>> small) program and instructions to reproduce the problem.
>>>
>>> Thanks,
>>> Robby
>>>
>>>
>>> On Sun, Jan 31, 2016 at 4:51 PM, jon stenerson 
>>> wrote:

 Just using DrRacket 6.3 on Win 10.

 Jon


 On 1/31/2016 3:22 PM, Robby Findler wrote:
>
> The compilation of racket is (not yet) deterministic so things like
> that can throw off internal tables in the .zo files (and a .zo files
> is loaded without checking dependencies, optimistically hoping that
> the files are not changed).
>
> So: are you using "raco make" or somehow mangaging .zo files yourself,
> or are you seeing this with the .zo files that DrRacket maintains for
> your automatically? If the former, then you need to adapt your
> workflow to either not use .zo files or to keep them up to date. If
> the latter, then it is a bug.
>
> Robby
>
>
>
> On Sun, Jan 31, 2016 at 4:17 PM, jon stenerson
> 
> wrote:
>>
>> I have some .rkt files and, depending on which file was edited most
>> recently, either it works fine or it gives a compilation error. (By
>> "edited"
>> I mean just adding or subtracting a blank line to change the
>> timestamp).
>> I
>> can submit a simplified example if nobody else has seen this problem.
>> If
>> this is well-known, what is the fix?
>>
>> link: bad variable linkage;
>>reference to a variable that is not a procedure or structure-type
>> constant
>> across all instantiations
>> reference phase level: 0
>> variable module: 
>> variable phase: 0
>> reference in module:  in: 
>>
>> --
>> You received this message because you are subscribed to the Google
>> Groups
>> "Racket Users" group.
>> To unsubscribe from this group and stop receiving emails from it, send
>> an
>> email to racket-users+unsubscr...@googlegroups.com.
>> For more options, visit https://groups.google.com/d/optout.


 --
 You received this message because you are subscribed to the Google
 Groups
 "Racket Users" group.
 To unsubscribe from this group and stop receiving emails from it, send
 an
 email to racket-users+unsubscr...@googlegroups.com.
 For more options, visit https://groups.google.com/d/optout.
>>
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "Racket Users" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to racket-users+unsubscr...@googlegroups.com.
>> For more options, visit https://groups.google.com/d/optout.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" 

Re: [racket-users] link: bad variable linkage

[Laurent]

I would be so happy if this kind of error could be resolved! (even when all
files are compiled by DrRacket)
Usually, I just turn off "populate 'compiled' directories", but then
DrRacket needs to compile the file on each F5, which can take some precious
time.
Thank you both then :)

On Sun, Jan 31, 2016 at 11:19 PM, jon stenerson 
wrote:

>
> The simplest thing I can find: I have three files in two directories.
> The first directory, called junk, contains a.rkt and b.rkt.
> The second directory is a sibling to junk called local-libs and contains
> c.rkt. (local-libs is known to the package manager; it was installed via
> the DrRacket interface).
> If I open all three files in DrRacket and try running a.rkt, sometimes it
> works, sometimes not. By modifiying a, b, c in some order (in the DrRacket
> editor) I can toggle the work/not work flag. As you say, it's a little
> non-deterministic.
>
> Thanks,
> Jon
>
>
> a.rkt:
> #lang racket
> (require "b.rkt")
>
> b.rkt
> #lang racket
> (require "../local-libs/c.rkt")
> (s 1)
>
> c.rkt
> #lang racket
> (provide (struct-out s) )
> (struct s (p))
>
>
>
>
> On 1/31/2016 4:03 PM, Robby Findler wrote:
>
>> Then it would be helpful to us if you could provide some (hopefully
>> small) program and instructions to reproduce the problem.
>>
>> Thanks,
>> Robby
>>
>>
>> On Sun, Jan 31, 2016 at 4:51 PM, jon stenerson 
>> wrote:
>>
>>> Just using DrRacket 6.3 on Win 10.
>>>
>>> Jon
>>>
>>>
>>> On 1/31/2016 3:22 PM, Robby Findler wrote:
>>>
 The compilation of racket is (not yet) deterministic so things like
 that can throw off internal tables in the .zo files (and a .zo files
 is loaded without checking dependencies, optimistically hoping that
 the files are not changed).

 So: are you using "raco make" or somehow mangaging .zo files yourself,
 or are you seeing this with the .zo files that DrRacket maintains for
 your automatically? If the former, then you need to adapt your
 workflow to either not use .zo files or to keep them up to date. If
 the latter, then it is a bug.

 Robby



 On Sun, Jan 31, 2016 at 4:17 PM, jon stenerson <
 jonstener...@comcast.net>
 wrote:

> I have some .rkt files and, depending on which file was edited most
> recently, either it works fine or it gives a compilation error. (By
> "edited"
> I mean just adding or subtracting a blank line to change the
> timestamp).
> I
> can submit a simplified example if nobody else has seen this problem.
> If
> this is well-known, what is the fix?
>
> link: bad variable linkage;
>reference to a variable that is not a procedure or structure-type
> constant
> across all instantiations
> reference phase level: 0
> variable module: 
> variable phase: 0
> reference in module:  in: 
>
> --
> You received this message because you are subscribed to the Google
> Groups
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

>>> --
>>> You received this message because you are subscribed to the Google Groups
>>> "Racket Users" group.
>>> To unsubscribe from this group and stop receiving emails from it, send an
>>> email to racket-users+unsubscr...@googlegroups.com.
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>
> --
> You received this message because you are subscribed to the Google Groups
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] link: bad variable linkage

[Robby Findler]

Then it would be helpful to us if you could provide some (hopefully
small) program and instructions to reproduce the problem.

Thanks,
Robby


On Sun, Jan 31, 2016 at 4:51 PM, jon stenerson  wrote:
> Just using DrRacket 6.3 on Win 10.
>
> Jon
>
>
> On 1/31/2016 3:22 PM, Robby Findler wrote:
>>
>> The compilation of racket is (not yet) deterministic so things like
>> that can throw off internal tables in the .zo files (and a .zo files
>> is loaded without checking dependencies, optimistically hoping that
>> the files are not changed).
>>
>> So: are you using "raco make" or somehow mangaging .zo files yourself,
>> or are you seeing this with the .zo files that DrRacket maintains for
>> your automatically? If the former, then you need to adapt your
>> workflow to either not use .zo files or to keep them up to date. If
>> the latter, then it is a bug.
>>
>> Robby
>>
>>
>>
>> On Sun, Jan 31, 2016 at 4:17 PM, jon stenerson 
>> wrote:
>>>
>>> I have some .rkt files and, depending on which file was edited most
>>> recently, either it works fine or it gives a compilation error. (By
>>> "edited"
>>> I mean just adding or subtracting a blank line to change the timestamp).
>>> I
>>> can submit a simplified example if nobody else has seen this problem.  If
>>> this is well-known, what is the fix?
>>>
>>> link: bad variable linkage;
>>>   reference to a variable that is not a procedure or structure-type
>>> constant
>>> across all instantiations
>>>reference phase level: 0
>>>variable module: 
>>>variable phase: 0
>>>reference in module:  in: 
>>>
>>> --
>>> You received this message because you are subscribed to the Google Groups
>>> "Racket Users" group.
>>> To unsubscribe from this group and stop receiving emails from it, send an
>>> email to racket-users+unsubscr...@googlegroups.com.
>>> For more options, visit https://groups.google.com/d/optout.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] link: bad variable linkage

[jon stenerson]

The simplest thing I can find: I have three files in two directories.
The first directory, called junk, contains a.rkt and b.rkt.
The second directory is a sibling to junk called local-libs and contains 
c.rkt. (local-libs is known to the package manager; it was installed via 
the DrRacket interface).
If I open all three files in DrRacket and try running a.rkt, sometimes 
it works, sometimes not. By modifiying a, b, c in some order (in the 
DrRacket editor) I can toggle the work/not work flag. As you say, it's a 
little non-deterministic.


Thanks,
Jon


a.rkt:
#lang racket
(require "b.rkt")

b.rkt
#lang racket
(require "../local-libs/c.rkt")
(s 1)

c.rkt
#lang racket
(provide (struct-out s) )
(struct s (p))



On 1/31/2016 4:03 PM, Robby Findler wrote:

Then it would be helpful to us if you could provide some (hopefully
small) program and instructions to reproduce the problem.

Thanks,
Robby


On Sun, Jan 31, 2016 at 4:51 PM, jon stenerson  wrote:

Just using DrRacket 6.3 on Win 10.

Jon


On 1/31/2016 3:22 PM, Robby Findler wrote:

The compilation of racket is (not yet) deterministic so things like
that can throw off internal tables in the .zo files (and a .zo files
is loaded without checking dependencies, optimistically hoping that
the files are not changed).

So: are you using "raco make" or somehow mangaging .zo files yourself,
or are you seeing this with the .zo files that DrRacket maintains for
your automatically? If the former, then you need to adapt your
workflow to either not use .zo files or to keep them up to date. If
the latter, then it is a bug.

Robby



On Sun, Jan 31, 2016 at 4:17 PM, jon stenerson 
wrote:

I have some .rkt files and, depending on which file was edited most
recently, either it works fine or it gives a compilation error. (By
"edited"
I mean just adding or subtracting a blank line to change the timestamp).
I
can submit a simplified example if nobody else has seen this problem.  If
this is well-known, what is the fix?

link: bad variable linkage;
   reference to a variable that is not a procedure or structure-type
constant
across all instantiations
reference phase level: 0
variable module: 
variable phase: 0
reference in module:  in: 

--
You received this message because you are subscribed to the Google Groups
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


--
You received this message because you are subscribed to the Google Groups
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


--
You received this message because you are subscribed to the Google Groups "Racket 
Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] link: bad variable linkage

[jon stenerson]

Just using DrRacket 6.3 on Win 10.

Jon

On 1/31/2016 3:22 PM, Robby Findler wrote:

The compilation of racket is (not yet) deterministic so things like
that can throw off internal tables in the .zo files (and a .zo files
is loaded without checking dependencies, optimistically hoping that
the files are not changed).

So: are you using "raco make" or somehow mangaging .zo files yourself,
or are you seeing this with the .zo files that DrRacket maintains for
your automatically? If the former, then you need to adapt your
workflow to either not use .zo files or to keep them up to date. If
the latter, then it is a bug.

Robby



On Sun, Jan 31, 2016 at 4:17 PM, jon stenerson  wrote:

I have some .rkt files and, depending on which file was edited most
recently, either it works fine or it gives a compilation error. (By "edited"
I mean just adding or subtracting a blank line to change the timestamp). I
can submit a simplified example if nobody else has seen this problem.  If
this is well-known, what is the fix?

link: bad variable linkage;
  reference to a variable that is not a procedure or structure-type constant
across all instantiations
   reference phase level: 0
   variable module: 
   variable phase: 0
   reference in module:  in: 

--
You received this message because you are subscribed to the Google Groups
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


--
You received this message because you are subscribed to the Google Groups "Racket 
Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.