ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mariusz Nowostawski <mari...@marni.otago.ac.nz>
Subject Re: long command lines from execon with parallel=true
Date Tue, 23 Jan 2001 11:05:52 GMT
> Actually, as Nico pointed out, we've decided to use a very radical
> limit of 4096 bytes that POSIX defines and that works across all
> platforms (as far as we know) in the jikes/jvc case. We didn't want to
> support a map of ${os.name} to acceptable length of command lines, I
> still think this has been the right decision.
[...]
> > Add yet another argument, let say 'enablesplit' with default value
> > "true" or "false" which controls b) and a) respectively in case the
> > argument list exceeded a limit for the given OS.
> 
> This is getting quite complicated, isn't it 8-).

uhm... I do not like it that much either. I think I would settle for
4096 limit, with 'parallel' gone and 'enablesplit' default false.

> We'd have three attributes that influence each other - quite difficult
> to explain. If we want to follow that route, we'd better deprecate
> parallel in favour of limit="1" and limit="NONE".

or maybe something like wieviel="1" and wieviel="ALLES"   ;o)
agreed, parallel and limit do not make much sense to be used together, and
as 'parallel' is a subset of 'limit' in terms of semantics, parallel
should be deprecated.

> enablesplit should be set to false by default IMHO - keep things
> explicit, i.e. apply magic only if the user asks for it.

agreed,

> > Also, this length limit applies only to a limited number of OSes,
> 
> Not true at all. It applies to (almost?) every OS, the limit is just
> different.

right, mea culpa, I had in mind overcoming a particular limit value for
particular <execon> case with particular OS with the lowest limit value.


thanks
Mariusz


Mime
View raw message