librelist archives

« back to archive

ET

ET

From:
Raphael Mack
Date:
2013-02-21 @ 22:10
Hi there,

ET is (re)born: http://et.liberty-eiffel.org .

It is far from complete, and needs some CSS but let's get it working and 
than make pretty. So far it just pulls LibertyEiffel/Liberty after each 
commit into upstream and tries to bootstrap the compiler. Which is the 
first failing stage. Actually I do not have a glue why (you may follow the
link to see the output) The first time the compiler is executed from the 
Eiffel-compile it fails with

975	/home/et/Liberty/src/lib/parse/impl/descending/descending_terminal.e
976	/home/et/Liberty/src/lib/parse/atoms/parse_terminal.e
977	/home/et/Liberty/src/lib/parse/grammar/eiffel/eiffel_default_node_factory.e
****** Fatal Error:  Bad assignment. Cannot assign `2147483647'
which is of type INTEGER_32 into `Result' which is of type STRING.

The source lines involved by the message are the following:

Line 76 column 4 in TYPED_INTERNALS 
(/home/et/Liberty/src/lib/kernel/internal/typed_internals.e):
Trying to read file "/home/et/Liberty/src/lib/kernel/internal/typed_internals.e".
   type_attribute_count: INTEGER is
   ^                               
Line 88 column 4 in TYPED_INTERNALS 
(/home/et/Liberty/src/lib/kernel/internal/typed_internals.e):
Trying to read file "/home/et/Liberty/src/lib/kernel/internal/typed_internals.e".
   type_attribute_generating_type (i: INTEGER): STRING is
   ^                                                     
------

On my local machine bootstrapping works fine, so I guess it was not my 
commit which crashed it. The obvious differences are debian(squeeze) 64 
bit vs. Ubunut(12.10) 32 bit.

so, if you have any fast ideas, let me know, otherwise I'll investigate 
into this, before going on with ET implementation, where I think of

- class_check, compile and execute example programs in tutorial
- class_check in src/lib and src/wrappers
- se test (Liberty/test)

Cheers,
Rapha

ET

From:
Cyril Adrian
Date:
2013-03-26 @ 16:13
 Raphael,

I have a request: is it possible to have access to the last finished test?
It would help finding problems without having to wait for ET to be idle :-)

Thanks!




*Cyril ADRIAN** (from office)*
 [image: Google Plus] <http://plus.google.com/100388810006463519079> My
latest G+: 
http://computerhistory....<http://plus.google.com/100388810006463519079/posts/DkqWMSU7bVj>
My G+ <http://plus.google.com/100388810006463519079> -
Posts<http://plus.google.com/100388810006463519079/posts/>- Add
to Circles <http://plus.google.com/100388810006463519079/about/> - 17:00
Feb-26 <http://plus.google.com/100388810006463519079>
  Get this email app!

<http://www.wisestamp.com/apps/plus?utm_source=extension&utm_medium=email&utm_term=plus&utm_campaign=apps>

  <http://fr.linkedin.com/in/cadrian/>
  Get a signature like this.

<http://r1.wisestamp.com/r/landing?promo=35&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_35>
CLICK

HERE.<http://r1.wisestamp.com/r/landing?promo=35&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_35>
​

   <http://twitter.com/cadbart>
Follow @ <http://twitter.com/cadbart>    • https://t.co/4CPuEPnWqu new
germ<http://twitter.com/cadbart/statuses/316505158190256128> 12:01
Mar-26 <http://twitter.com/cadbart/statuses/316505158190256128>

Reply<http://twitter.com/?status=@cadbart&in_reply_to=cadbart&in_reply_to_status_id=316505158190256100>
• https://t.co/uyKeA9y1sn vision is not included in
LibertyEiffel<http://twitter.com/cadbart/statuses/316425076977319937>
06:43
Mar-26 <http://twitter.com/cadbart/statuses/316425076977319937>

Reply<http://twitter.com/?status=@cadbart&in_reply_to=cadbart&in_reply_to_status_id=316425076977319940>
   View
more <http://twitter.com/cadbart>
   Get this email app!

<http://www.wisestamp.com/apps/twitter?utm_source=extension&utm_medium=email&utm_term=twitter&utm_campaign=apps>

Re: [libertyeiffel] ET

From:
Raphael Mack
Date:
2013-03-26 @ 22:03
oh, at the beginning it was that way and old results have just been
overwritten. I found this not very elegant, as it was not clear which of
the results are already new. And also I store the result internally not
structured by the names, but just in numeric directories (e. g.
stage4/sub1/sub4) and then the introduction of a new test directory the
same name could occur twice and I have orphan directories in case a test
directory is removed.
It seems, that it would be best to use names for the substage
directories and mark the old results (in output it may be gray for
example). Additionally a complete view on the last run should be added.

Created #74 for this.

Cheers,
Rapha


Am Dienstag, den 26.03.2013, 17:13 +0100 schrieb Cyril ADRIAN: 
> Raphael,
> 
> 
> I have a request: is it possible to have access to the last finished
> test? It would help finding problems without having to wait for ET to
> be idle :-)
> 
> 
> Thanks!
> 
> 
> 
> 
> Cyril ADRIAN
> (from office)
> 
> 
> Google
> Plus
> My
> latest G
> +:
> http://computerhistory.... My G+ - Posts - Add to Circles - 17:00 Feb-26
>   Get this email
> app!    Get a
> signature like
> this. CLICK
> HERE. ​ 
>  
>  
> 
> 
> Follow
> @ 
> •
> https://t.co/4CPuEPnWqu new germ 12:01 Mar-26 Reply 
> •
> https://t.co/uyKeA9y1sn vision is not included in LibertyEiffel 06:43 
Mar-26 Reply 
> View
> more
>   Get this email
> app!   

Re: [libertyeiffel] ET

From:
Cyril Adrian
Date:
2013-02-22 @ 13:54
Hi Raphael,

2013/2/21 Raphael Mack <ramack@raphael-mack.de>

> ET is (re)born: http://et.liberty-eiffel.org .
>

great!


> It is far from complete, and needs some CSS but let's get it working and
> than make pretty.


yes, step by step



> So far it just pulls LibertyEiffel/Liberty after each commit into upstream
> and tries to bootstrap the compiler. Which is the first failing stage.
> Actually I do not have a glue why (you may follow the link to see the
> output) The first time the compiler is executed from the Eiffel-compile it
> fails with (…)
>

wow! what's happened?? I never saw that problem. It looks weird, as if the
compiler mixed unrelated data.
Could you try removing the "target" repository before starting the
bootstrap?


On my local machine bootstrapping works fine, so I guess it was not my
> commit which crashed it. The obvious differences are debian(squeeze) 64 bit
> vs. Ubunut(12.10) 32 bit.
>

I only have 64-bit machines (the last 32-bit was removed last month). I use
both ubuntu (12.04 but with a recent gcc 4.7) and debian (testing).


so, if you have any fast ideas, let me know, otherwise I'll investigate
> into this,


Try rm -rf target


before going on with ET implementation, where I think of
>
> - class_check, compile and execute example programs in tutorial
> - class_check in src/lib and src/wrappers
> - se test (Liberty/test)
>

yes

Cheers,



 *Cyril ADRIAN** (from office)*
 [image: Google Plus] <http://plus.google.com/100388810006463519079> My
latest G+: #golang 's mock framework is simple and beautiful — and powerful
:-) <http://plus.google.com/100388810006463519079/posts/Fr6n37mZLpP>
My G+ <http://plus.google.com/100388810006463519079> -
Posts<http://plus.google.com/100388810006463519079/posts/>- Add
to Circles <http://plus.google.com/100388810006463519079/about/> - 07:59
Jan-25 <http://plus.google.com/100388810006463519079>
  Get this email app!

<http://www.wisestamp.com/apps/plus?utm_source=extension&utm_medium=email&utm_term=plus&utm_campaign=apps>

  <http://fr.linkedin.com/in/cadrian/>
  Get a signature like this.

<http://r1.wisestamp.com/r/landing?promo=35&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_35>
CLICK

HERE.<http://r1.wisestamp.com/r/landing?promo=35&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_35>


   <http://twitter.com/cadbart>  Latest tweet:
Follow @cadbart <http://twitter.com/cadbart>    • https://t.co/27cHWKRD it
starts to work! - a
bit.<http://twitter.com/cadbart/statuses/304340029801652224>
• https://t.co/WyNqZCQj eiffeltest_ng: more
traces<http://twitter.com/cadbart/statuses/303978356913885184>
• https://t.co/uMJEf6SU
oops<http://twitter.com/cadbart/statuses/303978355202613249>    View
more <http://twitter.com/cadbart>
   Get this email app!

<http://www.wisestamp.com/apps/twitter?utm_source=extension&utm_medium=email&utm_term=twitter&utm_campaign=apps>

Re: [libertyeiffel] ET

From:
Raphael Mack
Date:
2013-02-22 @ 21:58
Hi,

Am Freitag, den 22.02.2013, 14:54 +0100 schrieb Cyril ADRIAN: 
>         So far it just pulls LibertyEiffel/Liberty after each commit
>         into upstream and tries to bootstrap the compiler. Which is
>         the first failing stage. Actually I do not have a glue why
>         (you may follow the link to see the output) The first time the
>         compiler is executed from the Eiffel-compile it fails with (…
>         ) 

> wow! what's happened?? I never saw that problem. It looks weird, as if
> the compiler mixed unrelated data.
> Could you try removing the "target" repository before starting the
> bootstrap?

I tried, but no chance. the germ compiler works but the first built from
eiffel source crashed this way.

When using -no_check instead of -boost for the eiffel compilation it
works as expected.

I also tried an older version, but this failed with
---------------------
Trying to read file "compile_to_c.id".
Previous IDs reloaded (max_id = 1521).
1 /home/et/Liberty_test/src/lib/kernel/any.e
2 /home/et/Liberty_test/src/lib/kernel/tuple.e
****** Fatal Error: Double definition of feature count.

The source lines involved by the message are the following:

Line 16 column 4 in TUPLE
(/home/et/Liberty_test/src/lib/kernel/tuple.e):
Trying to read file "/home/et/Liberty_test/src/lib/kernel/tuple.e".
   count: INTEGER is
   ^                
---------------------

I'm at my wits' end for now and I'll try later. Cyril, maybe I'll ask
you for a public key to give you an ssh login to the ET account, before
I'm too frustrated.

cheers,
Rapha

Re: [libertyeiffel] ET

From:
Cyril Adrian
Date:
2013-02-23 @ 14:41
2013/2/22 Raphael Mack <ramack@raphael-mack.de>

> > wow! what's happened?? I never saw that problem. It looks weird, as if
> > the compiler mixed unrelated data.
> > Could you try removing the "target" repository before starting the
> > bootstrap?
>
> I tried, but no chance. the germ compiler works but the first built from
> eiffel source crashed this way.
>
> When using -no_check instead of -boost for the eiffel compilation it
> works as expected.
>

OK… that gives me another idea. Could you modify ~/.serc/liberty.se and
remove the -Os in the [boost] section? (or modify install.sh)

Thanks,


*Cyril A**DRIAN*

[image: Facebook] <https://www.facebook.com/cyril.adrian> [image:
Twitter]<https://twitter.com/cadbart> [image:
LinkedIn] <http://fr.linkedin.com/in/cadrian/> [image: Google
Plus]<https://plus.google.com/u/0/100388810006463519079>
Contact me: [image: Google Talk] cyril.adrian@gmail.com
   [image: Twitter] <http://twitter.com/cadbart>  Latest tweet:
https://t.co/fI80INx4UV those generated files should not have been
commited Follow
@cadbart <http://twitter.com/cadbart>

Reply<http://twitter.com/?status=@cadbart&in_reply_to=cadbart&in_reply_to_status_id=305321035228643300>
Retweet <http://twitter.com/?status=RT @cadbart: https://t.co/fI80INx4UV
those generated files should not have been commited> 15:19
Feb-23<http://twitter.com/cadbart/statuses/305321035228643328>
  Get this email app!

<http://www.wisestamp.com/apps/twitter?utm_source=extension&utm_medium=email&utm_term=twitter&utm_campaign=apps>

  Free signature tool.

<http://r1.wisestamp.com/r/landing?promo=32&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_32>
CLICK
HERE TO GET 
IT.<http://r1.wisestamp.com/r/landing?promo=32&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_32>

Re: [libertyeiffel] ET

From:
Raphael Mack
Date:
2013-02-23 @ 22:29
Am Samstag, den 23.02.2013, 15:41 +0100 schrieb Cyril ADRIAN: 

> OK… that gives me another idea. Could you modify ~/.serc/liberty.se
> and remove the -Os in the [boost] section? (or modify install.sh)

-O2 and -O3 work fine. So it seems to be related to the gcc version
(Debian 4.4.5-8) 4.4.5. 
Thanks, Cyril for this solution!

Rapha

Re: [libertyeiffel] ET

From:
Cyril Adrian
Date:
2013-02-24 @ 09:37
ok, so I'll stick to -O2 instead of -Os

it seems a good compromise

cheers,

*Cyril A**DRIAN*

[image: Facebook] <https://www.facebook.com/cyril.adrian> [image:
Twitter]<https://twitter.com/cadbart> [image:
LinkedIn] <http://fr.linkedin.com/in/cadrian/> [image: Google
Plus]<https://plus.google.com/u/0/100388810006463519079>
Contact me: [image: Google Talk] cyril.adrian@gmail.com
   [image: Twitter] <http://twitter.com/cadbart>  Latest tweet:
https://t.co/fI80INx4UV those generated files should not have been
commited Follow
@cadbart <http://twitter.com/cadbart>

Reply<http://twitter.com/?status=@cadbart&in_reply_to=cadbart&in_reply_to_status_id=305321035228643300>
Retweet <http://twitter.com/?status=RT @cadbart: https://t.co/fI80INx4UV
those generated files should not have been commited> 15:19
Feb-23<http://twitter.com/cadbart/statuses/305321035228643328>
  Get this email app!

<http://www.wisestamp.com/apps/twitter?utm_source=extension&utm_medium=email&utm_term=twitter&utm_campaign=apps>

  Free signature tool.

<http://r1.wisestamp.com/r/landing?promo=32&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_32>
CLICK
HERE TO GET 
IT.<http://r1.wisestamp.com/r/landing?promo=32&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_32>



2013/2/23 Raphael Mack <ramack@raphael-mack.de>

> Am Samstag, den 23.02.2013, 15:41 +0100 schrieb Cyril ADRIAN:
>
> > OK… that gives me another idea. Could you modify ~/.serc/liberty.se
> > and remove the -Os in the [boost] section? (or modify install.sh)
>
> -O2 and -O3 work fine. So it seems to be related to the gcc version
> (Debian 4.4.5-8) 4.4.5.
> Thanks, Cyril for this solution!
>
> Rapha
>
>

Re: [libertyeiffel] ET

From:
H. Zwakenberg | Ocean Consulting GmbH
Date:
2013-02-24 @ 11:09
 which begs the question whether the C-code emitted is LINT-free?

 cheers
 Hans


 Am Sonntag, den 24.02.2013, 10:37 +0100 schrieb Cyril ADRIAN 
 <cyril.adrian@gmail.com>:
> ok, so I'll stick to -O2 instead of -Os
>
> it seems a good compromise
>
> cheers,
>
> CYRIL ADRIAN
>
>  [1]  [2]  [3]  [4]
>  Contact me:   cyril.adrian@gmail.com [5]
>
>  		  [6]
>  		 Latest tweet: https://t.co/fI80INx4UV [7] those generated files
> should not have been commited  Follow @cadbart [8] Reply [9] Retweet
> [10] 15:19 Feb-23 [11]
>
>    Get this email app!  [12] 
>   Free signature tool.  [13]  CLICK HERE TO GET IT. [14] ​
>
> 2013/2/23 Raphael Mack
>  Am Samstag, den 23.02.2013, 15:41 +0100 schrieb Cyril ADRIAN:
>
>  > OK… that gives me another idea. Could you modify
> ~/.serc/liberty.se [16]
>  > and remove the -Os in the [boost] section? (or modify install.sh)
>
>  -O2 and -O3 work fine. So it seems to be related to the gcc version
>  (Debian 4.4.5-8) 4.4.5.
>  Thanks, Cyril for this solution!
>
>  Rapha
>
>
>
> Links:
> ------
> [1] https://www.facebook.com/cyril.adrian
> [2] https://twitter.com/cadbart
> [3] http://fr.linkedin.com/in/cadrian/
> [4] https://plus.google.com/u/0/100388810006463519079
> [5] mailto:cyril.adrian@gmail.com
> [6] http://twitter.com/cadbart
> [7] https://t.co/fI80INx4UV
> [8] http://twitter.com/cadbart
> [9]
> 
> 
http://twitter.com/?status=@cadbart&in_reply_to=cadbart&in_reply_to_status_id=305321035228643300
> [10] http://twitter.com/?status=RT @cadbart: https://t.co/fI80INx4UV
> those generated files should not have been commited
> [11] http://twitter.com/cadbart/statuses/305321035228643328
> [12]
> 
> 
http://www.wisestamp.com/apps/twitter?utm_source=extension&utm_medium=email&utm_term=twitter&utm_campaign=apps
> [13]
> 
> 
http://r1.wisestamp.com/r/landing?promo=32&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_32
> [14]
> 
> 
http://r1.wisestamp.com/r/landing?promo=32&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_32
> [15] mailto:ramack@raphael-mack.de
> [16] http://liberty.se

Re: [libertyeiffel] ET

From:
Cyril Adrian
Date:
2013-02-25 @ 06:32
2013/2/24 H. Zwakenberg | Ocean Consulting GmbH <hz@ocean-consulting.de>

>  which begs the question whether the C-code emitted is LINT-free?
>

I don't know. Lint is a tool that facilitates the life of C developers but
sometimes takes shortcuts that are unfortunate with generated code. It's a
good indicator though, taken with a grain of salt.

But more fundamentally, is the emitted code conform to the C
specifications? That I don't know. Recent problems indicate there may be
some issues.

Cheers,

*Cyril A**DRIAN*

[image: Facebook] <https://www.facebook.com/cyril.adrian> [image:
Twitter]<https://twitter.com/cadbart> [image:
LinkedIn] <http://fr.linkedin.com/in/cadrian/> [image: Google
Plus]<https://plus.google.com/u/0/100388810006463519079>
Contact me: [image: Google Talk] cyril.adrian@gmail.com
   [image: Twitter] <http://twitter.com/cadbart>  Latest tweet:
https://t.co/60L7bTmtyy Merge branch 'master' of
http://t.co/IcRBj5g3xu:LibertyEiffel/Liberty Follow
@cadbart<http://twitter.com/cadbart>

Reply<http://twitter.com/?status=@cadbart&in_reply_to=cadbart&in_reply_to_status_id=305925587275509760>
Retweet <http://twitter.com/?status=RT @cadbart: https://t.co/60L7bTmtyy
Merge branch 'master' of http://t.co/IcRBj5g3xu:LibertyEiffel/Liberty> 07:21
Feb-25 <http://twitter.com/cadbart/statuses/305925587275509760>     Get
this email app!

<http://www.wisestamp.com/apps/twitter?utm_source=extension&utm_medium=email&utm_term=twitter&utm_campaign=apps>

  Free signature tool.

<http://r1.wisestamp.com/r/landing?promo=32&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_32>
CLICK
HERE TO GET 
IT.<http://r1.wisestamp.com/r/landing?promo=32&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_32>

Re: [libertyeiffel] ET

From:
H. Zwakenberg | Ocean Consulting GmbH
Date:
2013-02-25 @ 08:25
> But more fundamentally, is the emitted code conform to the C
> specifications? That I don't know. Recent problems indicate there may
> be some issues.

 I suppose nowadays the greatest common denominator is C99.  C11 is 
 already supported by some compilers but I guess for now C99 would be a 
 safer choice for crossplatform retargetting.

Re: [libertyeiffel] ET

From:
Raphael Mack
Date:
2013-02-25 @ 20:59
Hello Hans, hello others!

First of all welcome here on this list. - It is nice to see some
interest in Liberty.

Am Montag, den 25.02.2013, 09:25 +0100 schrieb H. Zwakenberg | Ocean
Consulting GmbH: 
> > But more fundamentally, is the emitted code conform to the C
> > specifications? That I don't know. Recent problems indicate there may
> > be some issues.
> 
>  I suppose nowadays the greatest common denominator is C99.  C11 is 
>  already supported by some compilers but I guess for now C99 would be a 
>  safer choice for crossplatform retargetting.

For the C standard I'm on a line with you. C99 can be expected to be
available.

About the "language-restricting" "tools" like lint, MISRA or whatever
you like, I think they are great to help writing of portable code and to
enforce some really useful guidelines. On the other side my experience
is, that changing code just to satisfy any of these static checkers has
higher chance to introduce bugs, than to remove existing bugs. So, I'd
claim that theses tools are useful, but mostly when applied during
development. "Clean" (from lint's perspective) code is easier to
maintain, so it may be worth some effort but not every change suggested
by the tool should be realized blindly.

For Liberty I think the way could be:
- remove C-compiler warnings (use gccs -Wall)
- analyze generated C code with lint or MISRA
- decide which of the changes really improve the code and fix them

If you'd like to investigate into this you are very welcome to
contribute!

Best regards,
Rapha

Re: [libertyeiffel] ET

From:
Raphael Mack
Date:
2013-03-08 @ 18:39
Hello,

after Paolos last commit I again get crashes on both of my machine. But
I don't think he broke it, instead I believe that
- the "-Os" problem of this thread
- what I currently discover with -O2 (without -Ox it compiles fine)
- #47

is the same effect. And I stopped to believe it being a gcc bug. At
least my current crashes happen during mark phase of garbage collection.

So, I'd be interested about how the current repository content
bootstraps on your sites...

Regards,
Rapha

Am Montag, den 25.02.2013, 21:59 +0100 schrieb Raphael Mack: 
> Hello Hans, hello others!
> 
> First of all welcome here on this list. - It is nice to see some
> interest in Liberty.
> 
> Am Montag, den 25.02.2013, 09:25 +0100 schrieb H. Zwakenberg | Ocean
> Consulting GmbH: 
> > > But more fundamentally, is the emitted code conform to the C
> > > specifications? That I don't know. Recent problems indicate there may
> > > be some issues.
> > 
> >  I suppose nowadays the greatest common denominator is C99.  C11 is 
> >  already supported by some compilers but I guess for now C99 would be a 
> >  safer choice for crossplatform retargetting.
> 
> For the C standard I'm on a line with you. C99 can be expected to be
> available.
> 
> About the "language-restricting" "tools" like lint, MISRA or whatever
> you like, I think they are great to help writing of portable code and to
> enforce some really useful guidelines. On the other side my experience
> is, that changing code just to satisfy any of these static checkers has
> higher chance to introduce bugs, than to remove existing bugs. So, I'd
> claim that theses tools are useful, but mostly when applied during
> development. "Clean" (from lint's perspective) code is easier to
> maintain, so it may be worth some effort but not every change suggested
> by the tool should be realized blindly.
> 
> For Liberty I think the way could be:
> - remove C-compiler warnings (use gccs -Wall)
> - analyze generated C code with lint or MISRA
> - decide which of the changes really improve the code and fix them
> 
> If you'd like to investigate into this you are very welcome to
> contribute!
> 
> Best regards,
> Rapha
> 

Re: [libertyeiffel] ET

From:
Cyril Adrian
Date:
2013-03-11 @ 16:00
Hi,

2013/3/8 Raphael Mack <ramack@raphael-mack.de>

> At
> least my current crashes happen during mark phase of garbage collection.
>


Please Raphael, could you try using that patch:
https://github.com/cadrian/Liberty/commit/6fed299c48a2d93e1ea5f5526639cfe131dc5a3a

thanks :-)

Cyril

Re: [libertyeiffel] ET

From:
Raphael Mack
Date:
2013-03-11 @ 20:26
Hi,

bad news, it does not change anything. - At least not the observable
behavior, the amount of used RAM is certainly influenced ;-)

compile_to_c without GC was not possible, as I run out of memory (if I
remember correctly)

Running the compiler with gc is good, as it detects bugs in the GC (if
it is there) I'm quite optimistic, that we sooner or later find it.



Am Montag, den 11.03.2013, 17:00 +0100 schrieb Cyril ADRIAN: 
> Hi,
> 
> 2013/3/8 Raphael Mack <ramack@raphael-mack.de>
>         At
>         least my current crashes happen during mark phase of garbage
>         collection. 
> 
> 
> 
> 
> Please Raphael, could you try using that
> patch: 
https://github.com/cadrian/Liberty/commit/6fed299c48a2d93e1ea5f5526639cfe131dc5a3a
> 
> 
> thanks :-)
> 
> 
> Cyril
>  

Re: [libertyeiffel] ET

From:
Cyril Adrian
Date:
2013-03-11 @ 20:31
Hi,

2013/3/11 Raphael Mack <ramack@raphael-mack.de>

> bad news, it does not change anything. - At least not the observable
> behavior, the amount of used RAM is certainly influenced ;-)
>

That's a pity. Oh, well. I guess it fixes other problems since I already
saw pointers with value 1 (FSOH_MARKED) and I found this strange union
while reading the code. I still don't really understand it.


compile_to_c without GC was not possible, as I run out of memory (if I
> remember correctly)
>

Yes, I tries after suggesting that and had the same result.


Running the compiler with gc is good, as it detects bugs in the GC (if
> it is there) I'm quite optimistic, that we sooner or later find it.
>

We agree. The hunt is on :-)

Cheers,

*Cyril A**DRIAN*

[image: Facebook] <https://www.facebook.com/cyril.adrian> [image:
Twitter]<https://twitter.com/cadbart> [image:
LinkedIn] <http://fr.linkedin.com/in/cadrian/> [image: Google
Plus]<https://plus.google.com/u/0/100388810006463519079>
Contact me: [image: Google Talk] cyril.adrian@gmail.com
   [image: Twitter] <http://twitter.com/cadbart>  Latest tweet:
https://t.co/0T121F4tZ8 eiffeltest_ng is not ripe enough, keep eiffeltest
as default test command Follow @cadbart <http://twitter.com/cadbart>

Reply<http://twitter.com/?status=@cadbart&in_reply_to=cadbart&in_reply_to_status_id=311148547762434050>
Retweet <http://twitter.com/?status=RT @cadbart: https://t.co/0T121F4tZ8
eiffeltest_ng is not ripe enough, keep eiffeltest as default test
command> 17:16
Mar-11 <http://twitter.com/cadbart/statuses/311148547762434048>     Get
this email app!

<http://www.wisestamp.com/apps/twitter?utm_source=extension&utm_medium=email&utm_term=twitter&utm_campaign=apps>

  Free signature tool.

<http://r1.wisestamp.com/r/landing?promo=32&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_32>
CLICK
HERE TO GET 
IT.<http://r1.wisestamp.com/r/landing?promo=32&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_32>
​

Re: [libertyeiffel] ET

From:
Cyril Adrian
Date:
2013-03-11 @ 08:26
Hi,

2013/3/8 Raphael Mack <ramack@raphael-mack.de>

> I stopped to believe it being a gcc bug. At
> least my current crashes happen during mark phase of garbage collection.
>
> So, I'd be interested about how the current repository content
> bootstraps on your sites...
>

They do fine, thanks :-)

You are right, the GC is a good candidate. Your message makes me remember
that I already saw strange inexplicable artifacts when the GC is activated.

A temporary solution is to deactivate the GC for compile_to_c. SmartEiffel
tools were compiled without GC anyway (Dominique was very clever with
manual memory management).

A long term solution is to find the bug in the GC… I'd like the compiler to
use its own GC (eat your own dog food) ;-)

Cheers,




 *Cyril ADRIAN** (from office)*
 [image: Google Plus] <http://plus.google.com/100388810006463519079> My
latest G+: 
http://computerhistory....<http://plus.google.com/100388810006463519079/posts/DkqWMSU7bVj>
My G+ <http://plus.google.com/100388810006463519079> -
Posts<http://plus.google.com/100388810006463519079/posts/>- Add
to Circles <http://plus.google.com/100388810006463519079/about/> - 17:00
Feb-26 <http://plus.google.com/100388810006463519079>
  Get this email app!

<http://www.wisestamp.com/apps/plus?utm_source=extension&utm_medium=email&utm_term=plus&utm_campaign=apps>

  <http://fr.linkedin.com/in/cadrian/>
  Get a signature like this.

<http://r1.wisestamp.com/r/landing?promo=35&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_35>
CLICK

HERE.<http://r1.wisestamp.com/r/landing?promo=35&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_35>
​

   <http://twitter.com/cadbart>
Follow @ <http://twitter.com/cadbart>    • https://t.co/k96N5eFNAa
oops bis<http://twitter.com/cadbart/statuses/308814926087081984> 06:43
Mar-05 <http://twitter.com/cadbart/statuses/308814926087081984>

Reply<http://twitter.com/?status=@cadbart&in_reply_to=cadbart&in_reply_to_status_id=308814926087082000>
• https://t.co/jlFJ9djb9p
oops<http://twitter.com/cadbart/statuses/308814513271107584> 06:41
Mar-05 <http://twitter.com/cadbart/statuses/308814513271107584>

Reply<http://twitter.com/?status=@cadbart&in_reply_to=cadbart&in_reply_to_status_id=308814513271107600>
• https://t.co/oWpzvW1k4v agent bug: closed operands did not use inline
dispatch information<http://twitter.com/cadbart/statuses/308814053948661760>
06:39
Mar-05 <http://twitter.com/cadbart/statuses/308814053948661760>

Reply<http://twitter.com/?status=@cadbart&in_reply_to=cadbart&in_reply_to_status_id=308814053948661760>
   View
more <http://twitter.com/cadbart>
   Get this email app!

<http://www.wisestamp.com/apps/twitter?utm_source=extension&utm_medium=email&utm_term=twitter&utm_campaign=apps>

Re: [libertyeiffel] ET

From:
Paolo Redaelli
Date:
2013-02-22 @ 08:37
Wow!
Please excuse me, Raphael, I left my laptop on last night while working on
the netbook.
Bootstrap does fail because currently is tries to build everything even
regenerating all low-level wrappers and other things I don't remember....
When I run it locally it actually and correctly build the compiler.


2013/2/21 Raphael Mack <ramack@raphael-mack.de>

> Hi there,
>
> ET is (re)born: http://et.liberty-eiffel.org .
>
> It is far from complete, and needs some CSS but let's get it working and
> than make pretty. So far it just pulls LibertyEiffel/Liberty after each
> commit into upstream and tries to bootstrap the compiler. Which is the
> first failing stage. Actually I do not have a glue why (you may follow the
> link to see the output) The first time the compiler is executed from the
> Eiffel-compile it fails with
>
> 975
> /home/et/Liberty/src/lib/parse/impl/descending/descending_terminal.e
> 976     /home/et/Liberty/src/lib/parse/atoms/parse_terminal.e
> 977
> /home/et/Liberty/src/lib/parse/grammar/eiffel/eiffel_default_node_factory.e
> ****** Fatal Error:  Bad assignment. Cannot assign `2147483647'
> which is of type INTEGER_32 into `Result' which is of type STRING.
>
> The source lines involved by the message are the following:
>
> Line 76 column 4 in TYPED_INTERNALS
> (/home/et/Liberty/src/lib/kernel/internal/typed_internals.e):
> Trying to read file
> "/home/et/Liberty/src/lib/kernel/internal/typed_internals.e".
>    type_attribute_count: INTEGER is
>    ^
> Line 88 column 4 in TYPED_INTERNALS
> (/home/et/Liberty/src/lib/kernel/internal/typed_internals.e):
> Trying to read file
> "/home/et/Liberty/src/lib/kernel/internal/typed_internals.e".
>    type_attribute_generating_type (i: INTEGER): STRING is
>    ^
> ------
>
> On my local machine bootstrapping works fine, so I guess it was not my
> commit which crashed it. The obvious differences are debian(squeeze) 64 bit
> vs. Ubunut(12.10) 32 bit.
>
> so, if you have any fast ideas, let me know, otherwise I'll investigate
> into this, before going on with ET implementation, where I think of
>
> - class_check, compile and execute example programs in tutorial
> - class_check in src/lib and src/wrappers
> - se test (Liberty/test)
>
> Cheers,
> Rapha
>
>

Re: [libertyeiffel] ET

From:
Cyril Adrian
Date:
2013-02-22 @ 15:10
2013/2/22 Paolo Redaelli <paolo.redaelli@gmail.com>

> Bootstrap does fail because currently is tries to build everything even
> regenerating all low-level wrappers and other things I don't remember....
>

That's strange: I disabled most features, currently install.sh
-bootstrapshould only build the tools…

Cheers,



*Cyril ADRIAN** (from office)*
 [image: Google Plus] <http://plus.google.com/100388810006463519079> My
latest G+: #golang 's mock framework is simple and beautiful — and powerful
:-) <http://plus.google.com/100388810006463519079/posts/Fr6n37mZLpP>
My G+ <http://plus.google.com/100388810006463519079> -
Posts<http://plus.google.com/100388810006463519079/posts/>- Add
to Circles <http://plus.google.com/100388810006463519079/about/> - 07:59
Jan-25 <http://plus.google.com/100388810006463519079>
  Get this email app!

<http://www.wisestamp.com/apps/plus?utm_source=extension&utm_medium=email&utm_term=plus&utm_campaign=apps>

  <http://fr.linkedin.com/in/cadrian/>
  Get a signature like this.

<http://r1.wisestamp.com/r/landing?promo=35&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_35>
CLICK

HERE.<http://r1.wisestamp.com/r/landing?promo=35&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_35>


   <http://twitter.com/cadbart>  Latest tweet:
Follow @cadbart <http://twitter.com/cadbart>    • https://t.co/27cHWKRD it
starts to work! - a
bit.<http://twitter.com/cadbart/statuses/304340029801652224>
• https://t.co/WyNqZCQj eiffeltest_ng: more
traces<http://twitter.com/cadbart/statuses/303978356913885184>
• https://t.co/uMJEf6SU
oops<http://twitter.com/cadbart/statuses/303978355202613249>    View
more <http://twitter.com/cadbart>
   Get this email app!

<http://www.wisestamp.com/apps/twitter?utm_source=extension&utm_medium=email&utm_term=twitter&utm_campaign=apps>

Re: [libertyeiffel] ET

From:
Raphael Mack
Date:
2013-02-23 @ 10:13
Hi,

Am Freitag, den 22.02.2013, 16:10 +0100 schrieb Cyril ADRIAN: 
> 2013/2/22 Paolo Redaelli <paolo.redaelli@gmail.com>
>         Bootstrap does fail because currently is tries to build
>         everything even regenerating all low-level wrappers and other
>         things I don't remember....  

> That's strange: I disabled most features, currently install.sh
> -bootstrap should only build the tools…

cyril is right. boootstrap fails for the compiler for ET, not at the
wrappers stage (which was the case for #24 - which I also consider
closed)


regards,
Rapha 

Re: [libertyeiffel] ET

From:
Paolo Redaelli
Date:
2013-02-22 @ 08:45
As of wrappers I suggest to check only common, posix, ffi and readline.
I found that recent Gtk headers are not easily digested by gccxml...

I see you have both 32 and 64 machines like me. (let's continue in the
other thread)


2013/2/22 Paolo Redaelli <paolo.redaelli@gmail.com>

> Wow!
> Please excuse me, Raphael, I left my laptop on last night while working on
> the netbook.
> Bootstrap does fail because currently is tries to build everything even
> regenerating all low-level wrappers and other things I don't remember....
> When I run it locally it actually and correctly build the compiler.
>
>
> 2013/2/21 Raphael Mack <ramack@raphael-mack.de>
>
>> Hi there,
>>
>> ET is (re)born: http://et.liberty-eiffel.org .
>>
>> It is far from complete, and needs some CSS but let's get it working and
>> than make pretty. So far it just pulls LibertyEiffel/Liberty after each
>> commit into upstream and tries to bootstrap the compiler. Which is the
>> first failing stage. Actually I do not have a glue why (you may follow the
>> link to see the output) The first time the compiler is executed from the
>> Eiffel-compile it fails with
>>
>> 975
>> /home/et/Liberty/src/lib/parse/impl/descending/descending_terminal.e
>> 976     /home/et/Liberty/src/lib/parse/atoms/parse_terminal.e
>> 977
>> /home/et/Liberty/src/lib/parse/grammar/eiffel/eiffel_default_node_factory.e
>> ****** Fatal Error:  Bad assignment. Cannot assign `2147483647'
>> which is of type INTEGER_32 into `Result' which is of type STRING.
>>
>> The source lines involved by the message are the following:
>>
>> Line 76 column 4 in TYPED_INTERNALS
>> (/home/et/Liberty/src/lib/kernel/internal/typed_internals.e):
>> Trying to read file
>> "/home/et/Liberty/src/lib/kernel/internal/typed_internals.e".
>>    type_attribute_count: INTEGER is
>>    ^
>> Line 88 column 4 in TYPED_INTERNALS
>> (/home/et/Liberty/src/lib/kernel/internal/typed_internals.e):
>> Trying to read file
>> "/home/et/Liberty/src/lib/kernel/internal/typed_internals.e".
>>    type_attribute_generating_type (i: INTEGER): STRING is
>>    ^
>> ------
>>
>> On my local machine bootstrapping works fine, so I guess it was not my
>> commit which crashed it. The obvious differences are debian(squeeze) 64 bit
>> vs. Ubunut(12.10) 32 bit.
>>
>> so, if you have any fast ideas, let me know, otherwise I'll investigate
>> into this, before going on with ET implementation, where I think of
>>
>> - class_check, compile and execute example programs in tutorial
>> - class_check in src/lib and src/wrappers
>> - se test (Liberty/test)
>>
>> Cheers,
>> Rapha
>>
>>
>