[ ghc-Bugs-742918 ] link error from large pattern bindings

[SourceForge.net]

Bugs item #742918, was opened at 2003-05-24 22:51
Message generated for change (Comment added) made by simonpj
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=108032&aid=742918&group_id=8032

Category: Compiler
Group: 5.04.3
>Status: Closed
>Resolution: Fixed
Priority: 5
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: link error from large pattern bindings

Initial Comment:
When compiling a file where a pattern at the top level
binds more than 62 names compilation dies on a link
error, complaining about an undefined symbol :

test.o(.text+0x5ff): In function `s1MS_1_alt':
: undefined reference to `DataziTuple_Z71T_con_info'

module Main where
[a0,a1,a2,a3,a4,a5,a6,a7,a8,a9,a10,a11,a12,a13,a14,a15,a16,a17,
 a18,a19,a20,a21,a22,a23,a24,a25,a26,a27,a28,a29,a30,a31,a32,
 a33,a34,a35,a36,a37,a38,a39,a40,a41,a42,a43,a44,a45,a46,a47,
 a48,a49,a50,a51,a52,a53,a54,a55,a56,a57,a58,a59,a60,a61,a62] = 
   [0..62]

main = print a62

Compiling this without optimizations will die in the
linker with GHC 5.04 or the nightly build of 5.05 from
May 18 2003. Also try putting the definitions in a
separate module, which seems to always fail.

This needs a better error message, from the compiler
rather than linker, 
something like "Toplevel pattern binds too many names".
It took me a while to guess my problem from the weird
core name.

Brandon Moore
brandon at caltech.edu

----------------------------------------------------------------------

>Comment By: Simon Peyton Jones (simonpj)
Date: 2003-06-10 13:46

Message:
Logged In: YES 
user_id=50165

This long-standing bug is now fixed. The programmer still
can't use very big tuples, but the compiler does not now 
introduce them itself.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=108032&aid=742918&group_id=8032