[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

first inconsistencies



On Sat, Mar 08, 2003 at 10:06:43PM +0100, Remi Vanicat wrote:
> - mlgtk/lablgtk/lablgtk2 have modules that have the same name.
> - labltk and shell have a Shell module
> - labltk and camlimage have a Image module

Well the "same name problem" is a problem only if the modules with the
same name are supposed to be linked together.

- lablgtk vs lablgtk2 is pretty clear and seems not to be a problem:
  they definitely aren't supposed to be linked together

- labltk and shell again are a no problem. Looking at labltk META file
  we can see that only labltk.cma is supposed to be linked with the
  final application. This cma doesn't have references to any Shell
  interface name

  So probably Shell module @ lablgtk is an internal one, the real
  question is: why we ship modules that aren't supposed to be useful?
  After some more tests I discovered that labltk ships also a lot more
  of modules that aren't linked in the cma: Baloon, Fileselect, Frx_*,
  Jbp_font and more.

- labltk and camlimage inconsistency is a real problem:

  zack@lordsoth:~$ cat a.ml
  raise Image.Out_of_image;;
  zack@lordsoth:~$ ocamlfind ocamlc -package camlimages,labltk -linkpkg a.ml 
  findlib: [WARNING] Interface image.cmi occurs in several directories: /usr/lib/ocaml/3.06/camlimages, /usr/lib/ocaml/3.06/labltk
  Files a.cmo and /usr/lib/ocaml/3.06//labltk/labltk.cma(Image)
  make inconsistent assumptions over interface Image
  zack@lordsoth:~$ 

  Actually applications that use an Image module can't use both
  camlimages and labltk.

  We should probably report this problem to the ocaml and camliamges
  teams ...

> - then I don't understand this one :
> Inconsistent assumptions over Pa_extend:
>   Objects that believe Pa_extend has md5sum 6f6267d8155ea57dd34369c019b58e69:
>     /usr/lib/ocaml/3.06/camlp4/pa_extfold.cmo
>   Objects that believe Pa_extend has md5sum 3546e1510c5b78fe76214f0c974dac46:
>     /usr/lib/ocaml/3.06/camlp4/pa_extend_m.cmo
>     /usr/lib/ocaml/3.06/camlp4/pa_extend.cmo

This could be a problem indeed, we should ask camlp4 author ...

Cheers.

-- 
Stefano Zacchiroli  -  Undergraduate Student of CS @ Uni. Bologna, Italy
zack@{cs.unibo.it,debian.org,bononia.it}  -  http://www.bononia.it/zack/
"  I know you believe you understood what you think I said, but I am not
sure you realize that what you heard is not what I meant!  " -- G.Romney

Attachment: pgp_LSt9BmDNX.pgp
Description: PGP signature


Reply to: