Transformers:
* report:
- chekin done
- must be online after a rereading
* Wiki:
- update done
* Slides of the fourth stratego users day:
- must be fixed by Robert (he's awar of it)
- must be sent to E.Visser
--
Nicolas Desprès
nicolas.despres(a)lrde.epita.fr
Martin Bravenboer <martin(a)cs.uu.nl> writes:
> make: No pp entry found for: (1,["\"sdf-2.1\""]). This is our fault
> and probably does not occur in 0.9.5. Meanwhile, the pretty-printer
> has been improved considerably and has been upgraded to SDF as it
> appears in pgen 2.0. I've just fixed the latest pp-sdf to handle the
> old SDF constructs as well. The new ones are preferred. I don't think
> this results in problems.
I am trying to update. But many tools are not available with this new
grammar like "sdf-cons".
--
Valentin David
valentin(a)lrde.epita.fr
Martin Bravenboer <martin(a)cs.uu.nl> writes:
> Hey Transformers,
>
> > Transformers 0.1 is now released.
>
> Congrats!
>
> Do you want to announce this to stratego-announce as well? If you want
> to do this, then it would nice if the announcement contains some more
> information on the project and what's offered by the Transformers
> package.
>
> I've just installed the release on the latest sources of
> StrategoXT. This resulted in a few minor problems at installation
> time. After a succesful installation, I cannot figure out how to parse
> C++ files. I'm using the latest StrategoXT sources. Is that a problem?
Yes, it is. There so many changes in StrategoXT, that we have to
change a lot of things in transformers to follow.
>
> ---------------------------------------------------------------------
> configure: help2man is required, it wasn't available by default on my
> SuSE 9.0 system. Maybe it is a good idea to include the man page in a
> distribution?
We have discussed of that, and we decided to not distribute any
generated files. But, if help2man is not so popular on every
distribution, it will make us change our opinion.
>
> ---------------------------------------------------------------------
> configure: "stlport" wasn't available at my system as well. Might be
> useful to mention this requirement explicitly.
Sorry, i forgot. It was added in the requirement of cxx-grammar, but
not in the bundle.
>
> ---------------------------------------------------------------------
> make: No pp entry found for: (1,["\"sdf-2.1\""]). This is our fault
> and probably does not occur in 0.9.5. Meanwhile, the pretty-printer
> has been improved considerably and has been upgraded to SDF as it
> appears in pgen 2.0. I've just fixed the latest pp-sdf to handle the
> old SDF constructs as well. The new ones are preferred. I don't think
> this results in problems.
>
> I think we should add the transformers packages to our buildfarm. We
> are immediately warned if we break some of your tools then :) . We
> could also dailybuild tarballs and RPMs of Transformers.
We used to developped using the svn version of StrategoXT. But for
this release we used StrategoXT 0.9.5.
>
> ---------------------------------------------------------------------
> - make check after succesful installation:
>
> Is this ok?
> ...
> XFAIL: warn/implicit-typename3
> FAIL: warn/incomplete1
> FAIL: warn/inline1
> PASS: warn/noreturn-1
> FAIL: warn/oldcast1
> PASS: warn/pedantic1
> FAIL: warn/return-reference
> PASS: warn/weak1
> ===========================================
> 281 of 654 tests failed
> Please report to transformers(a)lrde.epita.fr
> ===========================================
> make[3]: *** [check-TESTS] Error 1
> make[3]: Leaving directory
> `/home/martin/wc/transformers-0.1/cxx-grammar/test/gcc'
Argh. It does not work at all.
> ---------------------------------------------------------------------
> A test by hand:
> ------------------------------------------
> int main() {
> }
> ------------------------------------------
> ./parse-cxx -i donothing.cpp
> ------------------------------------------
> sglr: error in /tmp/TransformerOwrMHo, line 1, col 7: character `i'
> (\x69) unexpected
> ------------------------------------------
Normally, it must work. The generated parse table must be wrong. What
"pp-aterm -i /pkg/transformers/2004-04-20-16-38/share/cxx-grammar/Cxx.tbl"
says? It is very strange that sglr stop at the 7th charater, which is
not so special.
>
> A more complex one:
> ------------------------------------------
> #include <iostream>
>
> int main() {
> std::cout << "Hello world!\n";
> }
This one is not supposed to work because we have still some troubles
with iostream, it is the aim of the next release.
> ------------------------------------------
> ./parse-cxx -i hello.cpp -of ast --verbose 10
> ------------------------------------------
> run sglr: FILE("/tmp/StrategoXTz3XkqY")
> xtc-find: Cxx.tbl
> xtc-load: Cxx.tbl
> xtc-loaded: Cxx.tbl
> xtc-find: /pkg/transformers/2004-04-20-16-38/share/cxx-grammar/Cxx.tbl
> xtc-find: sglr
> xtc-load: sglr
> xtc-loaded: sglr
> xtc-find: /home/martin/.nix-profile/bin/sglr
> sglr: opening parse table
> /pkg/transformers/2004-04-20-16-38/share/cxx-grammar/Cxx.tbl
> sglr: parsing file /tmp/StrategoXTz3XkqY (208 tokens)
> sglr: shifting [= ] / 7/208 ( 3%).
> sglr: writing error output to /tmp/StrategoXTvuPF5I
> sglr: error in /tmp/StrategoXTz3XkqY, line 1, col 13: character `d'
> (\x64) unexpected
> run sglr: FILE("/tmp/StrategoXTVcIeYt")
> ------------------------------------------
> Line 1 is:
> namespace std { }
> ------------------------------------------
>
> I suppose I'm doing something wrong, but I'm not sure about the direction
I think it is because you used the svn version of StrategoXT, well, i hope.
--
Valentin David
valentin(a)lrde.epita.fr
Anonymous class and namespace are authorized by this two rules. But it
is a problem for the filters. Make 2 filters, one to put a name and
another to delete it.
ClassKey Identifier? BaseClause? -> ClassHead
"namespace" "{" NamespaceBody "}" -> UnnamedNamespaceDefinition
You can put a invalid name to find them easily when deleting them
(e.g. using '<concat-strings>["#",<new>]' and
'where(explode-string;[35|id])').
Tell me when you are starting it.
--
Valentin David
valentin(a)lrde.epita.fr
Martin Bravenboer <martin(a)cs.uu.nl> wrote
> So we really should add the Transformers packages to our buildfarm. Then
> we can fix our stuff immediately if we break your tools.
You can access our subversion repository:
http://svn.lrde.epita.fr/svn/transformers/trunk
The directories are independant sub-projects, and the "bundle"
sub-project builds the whole transformers tarball.
> I've applied the necessary fixes to the latest sources of StrategoXT and
> 0.1 now runs fine :) . The check is ok and examples that don't use
> iostream can be parsed with parse-cxx. Looks promising!
We really appreciate the feedback.
Transformers 0.1 was released in order to get more people interested in
the project. Hopefully 0.2 will provide a more usable parser.
Thanks :)
--
Clement Vasseur -o)
[ nitro :: EPITA CSI 2005 ] /\\
"Programming is about being lazy." _\_V
Hey,
For some reason the archive link doesn't work at the listinfo page:
https://www.lrde.epita.fr/mailman/listinfo/transformers
The link is:
http://lrde.epita.fr/pipermail/transformers/
Adding www doesn't help.
Cheers,
--
Martin Bravenboer
---------------------------------------------------------------------
Center for Software Technology
Institute of Information and Computing Sciences
Utrecht University
------------------ Début de Rapport SpamAssassin ---------------------
Ce message est probablement du SPAM (message non sollicité envoyé en
masse, publicité, escroquerie...).
Cette notice a été ajoutée par le système d'analyse "SpamAssassin" sur
votre serveur de courrier "kualalumpur.lrde.epita.fr", pour vous
aider à identifier ce type de messages.
Le système SpamAssassin ajoute un en-tête "X-Spam-Flag: YES" aux
messages qu'il considère comme étant probablement du Spam.
Vous pouvez si vous le souhaitez utiliser cette caractéristique
pour régler un filtre dans votre logiciel de lecture de courrier,
afin de détruire ou de classer à part ce type de message.
Si ce robot a classifié incorrectement un message qui vous était
destiné, ou pour toute question, veuillez contacter l'administrateur
du système par e-mail à the administrator of that system .
Voir http://spamassassin.org/tag/ pour plus de détails (en anglais).
Détails de l'analyse du message: (13.2 points, 5.0 requis)
1.0 FROM_ENDS_IN_NUMS L'en-tête From: se termine par un nombre
2.7 RATWARE_OE_MALFORMED En-tête X-Mailer indique No de version Outlook Express malformé
0.5 RISK_FREE BODY: Propose une offre "sans aucun risque" (en anglais)
0.7 DATE_IN_PAST_12_24 Date: est 12 à 24 heures avant la date de l'en-tête Received:
1.5 RCVD_IN_BL_SPAMCOP_NET RBL: Relais listé dans http://spamcop.net/bl.shtml
[Blocked - see <http://www.spamcop.net/bl.shtml?212.247.106.210>]
0.7 NIGERIAN_BODY2 Corps ressemble à une escroquerie nigérienne 2+
2.0 NIGERIAN_BODY1 Corps ressemble à une escroquerie nigérienne 1+
0.7 NIGERIAN_BODY4 Corps ressemble à une escroquerie nigérienne 4+
0.7 NIGERIAN_BODY3 Corps ressemble à une escroquerie nigérienne 3+
2.6 FORGED_MUA_OUTLOOK Message falsifié prétendant provenir du logiciel MS Outlook
-------------------- Fin de Rapport SpamAssassin ---------------------
Hi everyone,
Transformers 0.1 is now released.
This release is a bundle of tools containing SDF tools for the C++
grammar, a C++ parser, and some examples of C++ rewriting tools using
Stratego/XT environment.
To use it, you will need Stratego/XT 0.9.5 available at
http://www.stratego-language.org/
Of course, Transformers is available at http://transformers.lrde.epita.fr/
Have a good night,
--
Valentin David
valentin(a)lrde.epita.fr
