<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix"><tt>I'll be compiling it myself anyway
but with this patch added:<br>
<a class="moz-txt-link-freetext" href="http://hackage.haskell.org/trac/ghc/ticket/3858">http://hackage.haskell.org/trac/ghc/ticket/3858</a><br>
... if there is an interest having the patch there.<br>
<br>
I'm typically adding two more patches (but I can skip these for
you if you<br>
do not like it):<br>
* add "breakTick" id to break point list and the information
about ignored<br>
breakpoint; this is interesting when you want to find out
which <br>
breakpoint was ignored; it is very dirty workaround for<br>
<a class="moz-txt-link-freetext" href="http://hackage.haskell.org/trac/ghc/ticket/2950">http://hackage.haskell.org/trac/ghc/ticket/2950</a><br>
I did not have time to do a proper patch.<br>
* I'm adding 0.1s delay when printing ghci prompt. Another dirty
way to get<br>
proper interleaving of stderr/stdout when stderr is coloured
using<br>
something like <a class="moz-txt-link-freetext" href="http://www.hck.sk/users/peter/pub/colorize.c">http://www.hck.sk/users/peter/pub/colorize.c</a><br>
I use it to get all stderr in red (stdout is default black),
and still<br>
achieve proper interleaving of stderr text with stdout text.<br>
It is quite possible this patch is not needed any more but it
was needed<br>
when haskeline was introduced the first time. I do not know
why, ghci was<br>
the only thing which showed any issues.<br>
<br>
I have 16GiB, so if the only problem is the memory, then I'll be
quick.<br>
<br>
Otherwise, I do not mind if you use ghc from [extra]. There is
no special<br>
reason not to use [extra] version except extra maintainers can
break<br>
our dependencies if they do unexpected recompile probably with
some<br>
significant patches. Although every recompile can break
dependencies,<br>
mostly it does not happen.<br>
You can just place copy of the [extra] version in [haskell] and
there<br>
should not be any problem at all since you would be in charge
of any<br>
ghc update. We would mirror only the ghc package in [haskell].<br>
<br>
Peter.<br>
</tt><br>
<br>
On 06/15/2012 09:54 PM, Magnus Therning wrote:<br>
</div>
<blockquote cite="mid:20120615195431.GA16404@ohann" type="cite">
<pre wrap="">I've been working on upgrading [haskell] to Ghc 7.4.2 for a few days
now. So far no luck :(
The build always fails on creating documentation. Yesterday I
resorted to taking the source package from [extra], which is known to
build fine... but still it failed on generating the docs. I now
suspect this is due to the limited memory available on the machine I
use to build (1.5GB).
For the moment I've given up on building Ghc especially for use in
[haskell]. In order to move on I'll instead make use of Ghc out of
[extra]. If anyone has any objections to that, then please speak up.
/M
</pre>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
arch-haskell mailing list
<a class="moz-txt-link-abbreviated" href="mailto:arch-haskell@haskell.org">arch-haskell@haskell.org</a>
<a class="moz-txt-link-freetext" href="http://www.haskell.org/mailman/listinfo/arch-haskell">http://www.haskell.org/mailman/listinfo/arch-haskell</a>
</pre>
</blockquote>
<br>
<br>
</body>
</html>