bug-libtool
[Top][All Lists]
Advanced

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

sys_max_cmd_len for BSDs (was: (no subject))


From: Ralf Wildenhues
Subject: sys_max_cmd_len for BSDs (was: (no subject))
Date: Wed, 30 Mar 2005 10:53:24 +0200
User-agent: Mutt/1.4.1i

Hi Mike,

* Mike Stump wrote on Wed, Mar 30, 2005 at 12:57:59AM CEST:
> How about this one...  The /4 comes from the old code in ltconfig  
> that did:
> 
> max = compute
> headroom = m /4
> max = max - headroom
> 
> In the end it was 3/4 the original value..

Yes, I looked around in the old code a bit, and think you are right
about this.  (16384 seems ok for C++ to me.)  The patch to the current
code[1] bears no obvious argument towards 0.25 vs 0.75. 

> That seems safe enough  for me, and allows all of gcc's libjava to
> link in one step, instead  of two extra steps, thus speeding it up by
> what I think was around 17  minutes for just one of the extra link
> steps.

Just curious: how long do the other two links take (the final link step;
and the only link step with your patch applied)?

> Let me know if sounds ok and if this goes into the tree, I want to
> pull it into gcc after you ok it.

Unless anybody has arguments against it, I will put it in this weekend
(and notify you then).

I wonder if we should adjust the safety factor for the default test as
well (it divides by two as of now)?

Just curious: which Libtool version did gcc originally branch its
version off of?  Is there an easy way to extract all Libtool-related
patches in gcc which differ from (i.e., did not go into) Libtool proper?

Regards,
Ralf

[1] http://lists.gnu.org/archive/html/libtool-patches/2004-04/msg00114.html

> 2005-03-29  Mike Stump  <address@hidden>
> 
>       * libtool.m4: Use 3/4 of argmax, instead of 1/4.
> 
> Doing diffs in .:
> --- ./libtool.m4.~1~  2005-02-12 04:18:34.000000000 -0800
> +++ ./libtool.m4      2005-03-29 14:29:18.000000000 -0800
> @@ -674,6 +674,7 @@ AC_CACHE_VAL([lt_cv_sys_max_cmd_len], [d
>      fi
>      # And add a safety zone
>      lt_cv_sys_max_cmd_len=`expr $lt_cv_sys_max_cmd_len \/ 4`
> +    lt_cv_sys_max_cmd_len=`expr $lt_cv_sys_max_cmd_len \* 3`
>      ;;
>    osf*)
>      # Dr. Hans Ekkehard Plesser reports seeing a kernel panic running 
> configure





reply via email to

[Prev in Thread] Current Thread [Next in Thread]