liberty-eiffel
[Top][All Lists]
Advanced

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

Re: [Liberty-eiffel] Puzzled


From: Raphael Mack
Subject: Re: [Liberty-eiffel] Puzzled
Date: Wed, 02 Mar 2016 22:04:01 +0100

Hi,

we have already discussed it, but for the interested reader and the
archives:

It seems to be the good old issue in the se_gc. Paolo, you could try
with -bdw_gc option (or "collect (bdw)" in ace)

Cheers,
Rapha

Am Montag, den 29.02.2016, 12:47 +0100 schrieb Paolo Redaelli:
> I'm quite puzzled as I have been "battling" against my own
> wrappers_generator to make it digest libxml2.
> 
> In fact it keeps "misteriously" crashing wrapping structure
> xmlParserCtxt when compiled with
> https://github.com/tybor/Liberty/blob/master/src/tools/wrappers-generator/wrappers_generator.ace
> while it happily process everything when it is compiled with
> https://github.com/tybor/Liberty/blob/master/src/tools/wrappers-generator/debug.ace
> which produces wrappers_generator.debug
> 
> What can it cause this most peculiar behaviour?
> I have double checked and I'm pretty sure the queries in the pre- and
> post-conditions should NOT have any side-effects: I tried hard to avoid
> side effects at all, knowing that sometimes in the future the compiler
> would forbit them.
> 
> I have commited those changes to my repo on github
> https://github.com/tybor/Liberty/
> 
> Since we're going to release ASAP I would rather compile
> wrappers_generator using a non-boost mode making this bug manifest.
> Currently the debug ace file - the one that makes the "working"
> wrappers_generator -  compiles the lib and tools clusters in boost mode
> and wrappers_generator with all assertions enabled. All in all it's not
> *that* slow. I will try lowering assertion level to discover the lowest one.
> 
> Meanwhile I'll turn my attention to SQLITE wrappers that turned out to
> be the current show-stoppers.
> 
> Oh, how much the user of those library would benefit from the convert
> clause!
> 
> More on this this evening.
> 
>  Paolo
> 





reply via email to

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