[GHC] #3637: ./configure doesn't understand Gentoo's build/host/target

[GHC-2]

#3637: ./configure doesn't understand Gentoo's build/host/target
-----------------------------+----------------------------------------------
Reporter:  kolmodin          |          Owner:                  
    Type:  bug               |         Status:  new            
Priority:  normal            |      Component:  Build System    
 Version:  6.12.1 RC1        |       Severity:  minor          
Keywords:                    |       Testcase:                  
      Os:  Unknown/Multiple  |   Architecture:  Unknown/Multiple
-----------------------------+----------------------------------------------
 Apparently there are several styles in how to write your build/host/target
 variables.

 In Gentoo they look like this for my amd64 {{{x86_64-pc-linux-gnu}}} which
 doesn't play well with the current build system.

 Default {{{./configure}}} execution in Gentoo ebuilds will pass the
 following flags:
 {{{
 ./configure --prefix=/usr --host=x86_64-pc-linux-gnu
 --mandir=/usr/share/man \
     --infodir=/usr/share/info --datadir=/usr/share --sysconfdir=/etc \
     --localstatedir=/var/lib --libdir=/usr/lib64 --build=x86_64-pc-linux-
 gnu
 }}}

 Which results in:

 {{{
 checking for gfind... no
 checking for find... /usr/bin/find
 checking for sort... /usr/bin/sort
 checking for GHC version date... given 6.12.0.20091010
 checking for ghc... /usr/bin/ghc
 checking version of ghc... 6.10.4
 Target platform inferred as: x86_64-unknown-linux
 Unknown vendor linux
 }}}

 So we sed it like this:
 {{{
 build=`echo "$build" | sed -e 's/linux-gnu/linux/' -e 's/-pc-/-unknown-/'`
 host=`echo "$host" | sed -e 's/linux-gnu/linux/' -e 's/-pc-/-unknown-/'`
 target=`echo "$target" | sed -e 's/linux-gnu/linux/' -e
 's/-pc-/-unknown-/'`
 }}}
 but of course we would prefer not to have to handle this specially.

 That is, the {{{vendor}}} is {{{pc}}} and the {{{OS}}} is {{{linux-gnu}}}.

--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/3637>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@...
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

[GHC-2]

#3637: ./configure doesn't understand Gentoo's build/host/target
------------------------------+---------------------------------------------
 Reporter:  kolmodin          |          Owner:                  
     Type:  bug               |         Status:  new            
 Priority:  normal            |      Milestone:                  
Component:  Build System      |        Version:  6.12.1 RC1      
 Severity:  minor             |     Resolution:                  
 Keywords:  regression        |       Testcase:                  
       Os:  Unknown/Multiple  |   Architecture:  Unknown/Multiple
------------------------------+---------------------------------------------
Changes (by slyfox):

  * keywords:  => regression
 * cc: slyfox@... (added)

--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/3637#comment:1>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@...
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

[GHC-2]

#3637: ./configure doesn't understand Gentoo's build/host/target
---------------------------------+------------------------------------------
    Reporter:  kolmodin          |        Owner:                  
        Type:  bug               |       Status:  new            
    Priority:  high              |    Milestone:  6.12.1          
   Component:  Build System      |      Version:  6.12.1 RC1      
    Severity:  minor             |   Resolution:                  
    Keywords:  regression        |   Difficulty:  Unknown        
    Testcase:                    |           Os:  Unknown/Multiple
Architecture:  Unknown/Multiple  |  
---------------------------------+------------------------------------------
Changes (by igloo):

  * priority:  normal => high
  * difficulty:  => Unknown
  * milestone:  => 6.12.1

Comment:

 The way the build system now works, the build/host/target values must be
 known to the build system. They're normally extracted from the "ghc +RTS
 --info" output of the bootstrapping compiler, but you can give them as
 flags when necessary (e.g. when bootstrapping).

 So, I think we have 3 options:
  * Add normalisation back. This means we have to keep it up-to-date, so I
 would rather not do this
  * Close this as wont-fix
  * Ignore the `--{host,build,target}` flags, and use
 `--ghc-{host,build,target}` instead

--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/3637#comment:2>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@...
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

[GHC-2]

#3637: ./configure doesn't understand Gentoo's build/host/target
---------------------------------+------------------------------------------
    Reporter:  kolmodin          |        Owner:                  
        Type:  bug               |       Status:  new            
    Priority:  high              |    Milestone:  6.12.1          
   Component:  Build System      |      Version:  6.12.1 RC1      
    Severity:  minor             |   Resolution:                  
    Keywords:  regression        |   Difficulty:  Unknown        
    Testcase:                    |           Os:  Unknown/Multiple
Architecture:  Unknown/Multiple  |  
---------------------------------+------------------------------------------
Comment (by kolmodin):

 Right. Given the three options; I think these are the results from our
 perspective;

  * You do the normalisation, which means work for you to keep up
  * Close as wont-fix means that we have to do it instead. Or, we could
 hack the ebuilds not to use {{{econf}}} but rather call {{{./configure}}}
 ourselves.
  * Using {{{--ghc-{host,build,target}}}} would mean the problem goes away
 as far as I know

 I think option 2 and 3 looks most attractive.
 With option 3, what could the consequences be? Would things break for
 other people?

--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/3637#comment:3>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@...
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs