From attila.lendvai at gmail.com Tue Apr 17 19:54:58 2007 From: attila.lendvai at gmail.com (Attila Lendvai) Date: Tue, 17 Apr 2007 21:54:58 +0200 Subject: [fetter-devel] carriage returns In-Reply-To: References: Message-ID: dear list, i'm experimenting with Verrazano and it looks very interesting and useful! but some of the files have carriage return chars in them (0x0d) which makes life hard when not on windows (and even there). i've got a handy script on common-lisp.net in ~alendvai/bin/delete-carriage-returns.sh **/* that removes them. i can offer help on this, but this offer is obviously worthless without commit rights... so it would be very helpful if someone could run it on a clean checkout and commit the changes. on a different note: i've made some changes to use trivial-shell and echo the gccxml output to the repl which greatly simplifies life. but right now a cvs diff results in a huge patch due to the carriage returns. and finally a suggestion from a newbie perspective (i've just passed about an hour with Verrazano): the way errors are catched and different errors are rethrown is not too helpful. after all, all the users will be programmers with a debugger at hand. if i were to decide i would let them reach the debugger as is, but with the echoed gccxml output the situation is already much better here. if the trivial-shell change is desirable then i can send a diff of my changes. hth, -- attila From davidhuebel at gmail.com Thu Apr 19 01:58:57 2007 From: davidhuebel at gmail.com (David Huebel) Date: Wed, 18 Apr 2007 20:58:57 -0500 Subject: [fetter-devel] carriage returns In-Reply-To: References: Message-ID: <50c6b7b60704181858x27c4fafard0868851987beb1d@mail.gmail.com> On 4/17/07, Attila Lendvai wrote: > i'm experimenting with Verrazano and it looks very interesting and useful! > > but some of the files have carriage return chars in them (0x0d) which > makes life hard when not on windows (and even there). My fault for not handling this when migrating from CVS. CVS translates line endings by default, but Subversion only does so when specifically instructed. Subversion handles end-of-line translations on files whose svn:eol-style property has the value "native". I will add this property to all the text files in the repository (this weekend at the latest; probably not before then). Unfortunately, the Subversion repository can't be configured to set this property automatically on newly added files. I believe the svn client can be configured to set it automatically, but I haven't tried it. -David From attila.lendvai at gmail.com Fri Apr 20 09:13:16 2007 From: attila.lendvai at gmail.com (Attila Lendvai) Date: Fri, 20 Apr 2007 11:13:16 +0200 Subject: [fetter-devel] carriage returns In-Reply-To: <50c6b7b60704181858x27c4fafard0868851987beb1d@mail.gmail.com> References: <50c6b7b60704181858x27c4fafard0868851987beb1d@mail.gmail.com> Message-ID: > My fault for not handling this when migrating from CVS. CVS translates line > endings by default, but Subversion only does so when specifically just to clarify, i've only checked out the CVS repo and i've seen the 0x0d chars there (didn't even know about the SVN repo before your previous mail. hint: the cl.net site advertises a one and only CVS repo). but either way, thanks for the help, -- attila From attila.lendvai at gmail.com Mon Apr 23 09:54:15 2007 From: attila.lendvai at gmail.com (Attila Lendvai) Date: Mon, 23 Apr 2007 11:54:15 +0200 Subject: [fetter-devel] Objections or comments on migrating to subversion? In-Reply-To: <50c6b7b60612141656i5cddb219o7b57963b9eb57202@mail.gmail.com> References: <50c6b7b60612141656i5cddb219o7b57963b9eb57202@mail.gmail.com> Message-ID: > I'll do the migration on Monday if I don't hear any objections. > Rayiner, you're the latest committer on all the files in the > repository -- let me know if this bothers you. i know it's a bit late, but FWIW i'd vote for darcs. i use both svn and darcs regularily and i find darcs to be the better one and it also has a trac plugin (http://progetti.arstecnica.it/trac+darcs/). svn is just a better written cvs, while darcs is a completly different design. i've made tailor converted versions of the cvs repo (keeping the commit history): darcs get http://www.common-lisp.net/project/cl-wdim/darcs/verrazano/ darcs get http://www.common-lisp.net/project/cl-wdim/darcs/verrazano-support/ i don't have a single patch in Verrazano yet, so don't take it as a heavy-weight opinion. i just wanted to send a pointer to the darcs repos for your consideration. on a different note: i've checked the svn repo and it does not have the comitted carriage returns that cvs had. so, as we seem to have obsoleted cvs, this is not an issue anymore. but whichever repo is the final word, the site should be updated not to advertise cvs. hth, -- attila From attila.lendvai at gmail.com Thu Apr 26 21:53:35 2007 From: attila.lendvai at gmail.com (Attila Lendvai) Date: Thu, 26 Apr 2007 23:53:35 +0200 Subject: [fetter-devel] policy on lib dependencies Message-ID: hi, i've successfully generated a cffi mapping to imagemagick and i must say that other then the relatively big pain to set up everything, it was a matter of minutes until i could resize an image from the lisp repl. wow, great work! while i was patching this and that, i felt the need for two additional dependencies: iterate and cl-ppcre. what do the maintainers think about depending on these quasi-standard libs? as another possibility it may be worth depending on alexandria (which is aiming to be "the cl utils lib") and eliminate some local utilities, but i didn't scan the code yet to actually suggest it. i'm planning to implement: - camel-case to hyphen conversion - automatic generation of cffi:with-foreign-string for (const ?) char * arguments. ideas? - porting to cffi-newtypes - get rid of the foo.binding interface and create a functional one - using trivial-shell to echo the gccxml output to the repl and assert for the return code (done) and sending the various small fixes i had to do. comments on the above plans are very welcome! happy hacking, -- attila From luismbo at gmail.com Thu Apr 26 23:05:16 2007 From: luismbo at gmail.com (=?ISO-8859-1?Q?Lu=EDs_Oliveira?=) Date: Fri, 27 Apr 2007 00:05:16 +0100 Subject: [fetter-devel] policy on lib dependencies In-Reply-To: References: Message-ID: <391f79580704261605s1b734c31p44de1a929402a84b@mail.gmail.com> On 26/04/07, Attila Lendvai wrote: > - automatic generation of cffi:with-foreign-string for (const ?) char > * arguments. ideas? If you want to do this, you can simply use the :string type. -- Lu?s Oliveira http://student.dei.uc.pt/~lmoliv/