--- - branch: MAIN date: Wed Nov 7 00:50:06 UTC 2012 files: - new: '1.31' old: '1.30' path: pkgsrc/devel/guile-www/Makefile pathrev: pkgsrc/devel/guile-www/Makefile@1.31 type: modified - new: '1.12' old: '1.11' path: pkgsrc/devel/guile-www/PLIST pathrev: pkgsrc/devel/guile-www/PLIST@1.12 type: modified - new: '1.12' old: '1.11' path: pkgsrc/devel/guile-www/distinfo pathrev: pkgsrc/devel/guile-www/distinfo@1.12 type: modified id: 20121107T005006Z.83422aaf2e595e7cfeeb32f4241245494d2110cb log: "Update to 2.35.\n\n- 2.35 | 2012-11-06\n\n - support for \"make installcheck\"\n\n\tThis does \"make check\" at its core, so the same env vars apply.\n\tObviously, you should do \"make install\" first for sane results.\n\n - â\x80\x98receive-responseâ\x80\x99 spec takes thunk for â\x80\x98customâ\x80\x99 handling\n\n\tThis means that instead of using:\n\n\t(list 'custom (values P1 P2 P3 P4))\n\n\tto construct a â\x80\x98receive-responseâ\x80\x99 spec, you can use:\n\n\t(list 'custom (lambda () (values P1 P2 P3 P4)))\n\n\tIn fact, this is now the preferred way, as support for using\n\tâ\x80\x98valuesâ\x80\x99 directly is now deemed a misfeature and marked for\n\teventual removal (see below).\n\n \ - tested against Guile 2.0.6\n\n\tThere are some deprecation warnings, however.\n\n \ - MARKED FOR REMOVAL\n\n - 2013-05-15 support for values in â\x80\x98receive-responseâ\x80\x99 spec\n\n\tOut of an explicit â\x80\x98call-with-valuesâ\x80\x99 context, Guile 2.x will\n\tsilently discard all values but the first if the continuation\n\texpects one value only (the common case).\n\n\t(cons 0 (values 1 2)) => (0 . 1)\n\n\tThus, the â\x80\x98receive-responseâ\x80\x99 spec constructed by:\n\n\t(list 'custom (values P1 P2 P3 P4))\n\n\tfails to preserve P2, P3, P4 (w/ Guile 2.x). See above for\n\tthunk support, the future-proof way to specify â\x80\x98customâ\x80\x99 procs.\n\n - bootstrap tools upgrade\n - GNU Autoconf 2.69\n - GNU Automake 1.12.4\n - Guile-BAUX 20121026.1208.c21074a\n\n- 2.34 | 2012-03-29\n\n - bug fix: unbotch â\x80\x98upath->filename-procâ\x80\x99\n - bug fix: unbotch â\x80\x98log-http-response-procâ\x80\x99\n\n\tIn Guile-WWW 2.33 (2012-02-29), these procs were changed to use\n\t(ice-9 optargs), but the change was botched.\n\n - bug fix: â\x80\x98(www http) http:post-formâ\x80\x99 reserves â\x80\x98#\\&â\x80\x99 and â\x80\x98#\\=â\x80\x99\n\n\tThat is, in the absense of upload specs, the urlencoding for the\n\tkey names and values now also includes ampersand and equal sign,\n\tas those used to delimit pairs and separate key and value within\n\ta pair, respectively.\n\n - bug fix: â\x80\x98url-coding:encodeâ\x80\x99 outputs two hex digits for 0-15\n\n\tPreviously, octets 0 through 15 were output as \"%0\" through\n\t\"%f\", respectively. Now, they are \"%00\" through \"%0f\".\n\n \ - portability fix: big dishing loop handles #f â\x80\x98sockaddr:pathâ\x80\x99\n\n\tFor Unix domain sockets, some Guile versions have â\x80\x98sockaddr:pathâ\x80\x99\n\tsometimes return the empty string, others â\x80\x98#fâ\x80\x99.\n\n - â\x80\x98(www url) url:parseâ\x80\x99 can handle IPv6 addresses\n\n\tHandling doesn't imply validation, however. For example:\n\t(url:host (url:parse \"http://fool@[::ffff:bogus.0.2.1]:42/foo\"))\n\t=> \"[::ffff:bogus.0.2.1]\"\n\n - â\x80\x98(... answer) hqf<-upathâ\x80\x99 more efficient\n\n\tThe char-set computation has been moved to load-time.\n\n - â\x80\x98(... cookies) simple-parse-cookiesâ\x80\x99 takes optional arg SEP\n\n\tBy default, tokens are â\x80\x98#\\,â\x80\x99 (comma)-separated. You can use SEP\n\tto specify another character, e.g., â\x80\x98#\\;â\x80\x99 (semicolon).\n\n - new â\x80\x98(www cgi) cgi:initâ\x80\x99 option: cookies-split-on-semicolon\n\n\tIf the server provides the CGI script with a \"Netscape style\"\n\t(i.e., old and deprecated) set of cookies, you can use this\n\toption to split them.\n\n - â\x80\x98(www cgi) cgi:getenvâ\x80\x99 recognizes some more \"idiomatic\" names\n\n\tThe following keys now are recognized by â\x80\x98cgi:getenvâ\x80\x99 (that is,\n\tthey no longer signal \"unrecognized key\" error): â\x80\x98server-nameâ\x80\x99,\n\tâ\x80\x98server-softwareâ\x80\x99, â\x80\x98server-protocolâ\x80\x99, â\x80\x98auth-typeâ\x80\x99, â\x80\x98http-acceptâ\x80\x99.\n\tAlso, the manual now marks those keys that either are aliases,\n\tor return parts of the associated value, of these idiomatic names.\n\n - new (... parse-request) proc: receive-request\n\n\tThis uses the same underlying machinery as â\x80\x98receive-responseâ\x80\x99\n\tfrom (www http), and thus sports the same features.\n\n\tAdditionally, there is a new â\x80\x98requestâ\x80\x99 type with five accessor\n\tprocs: request-{method,upath,protocol-version,headers,body} and\n\tthe new predicate â\x80\x98request?â\x80\x99.\n\n - new (... log) proc: string<-sockaddr\n\n\tThis formats AF_UNIX or AF_INET addresses (including port).\n\n \ - new (... answer) proc: compose-response\n\n\tThis is a stripped down variant of â\x80\x98mouthpieceâ\x80\x99. It uses the\n\tsame underlying machinery as â\x80\x98send-requestâ\x80\x99 from (www http) and\n\tthus sports the same features.\n\n \ - MARKED FOR REMOVAL\n\n - 2012-12-31 module (www main)\n\n - 2012-12-31 (www url) procs\n - url:address\n - url:unknown\n\n - 2013-02-28 (... parse-request) procs\n - read-first-line\n - read-headers\n - skip-headers\n - read-body\n\n - 2013-02-28 (www http) procs\n - http:head\n - http:get\n\n - test suite expanded\n\n\tPreviously, there was just one test. Now, there are several,\n\tand they are in directory v/ (for \"validation\", \"verification\",\n\t\"vouchsafed\", and so on).\n\n - doc license upgraded to GNU FDL 1.3\n\n - maintenance uses Guile-BAUX 20120309.1509.1c4bb92\n\n- 2.33 | 2012-02-29\n\n - distribution now .tar.xz\n\n\tIf you have GNU tar, you can use \"tar xf\" and it will DTRT.\n\tIf not, you can use \"xz -dc TARBALL | tar xf -\" to unpack.\n\n - configure no longer checks for some widely-available modules\n\n\tThe modules â\x80\x98(ice-9 and-let-star)â\x80\x99, â\x80\x98(ice-9 rdelim)â\x80\x99 and\n\tâ\x80\x98(ice-9 regex)â\x80\x99 have been widely-available since Guile 1.4.x.\n\n - configure uses SNUGGLE for cross-Guile portability\n\n\tSNUGGLE is Free Software: .\n\tIt provides â\x80\x98build-aux/snuggle.m4â\x80\x99, which is more amicable than\n\tthe various â\x80\x98guile.m4â\x80\x99 in the wild.\n\n \ - dropped â\x80\x98(www server-utils answer) string<-header-componentsâ\x80\x99\n\n\tThis was pre-announced below in Guile-WWW 2.28 (2010-06-22).\n\n - changes to (www http)\n\n - bug fix: â\x80\x98http:requestâ\x80\x99 sends message body as-is\n\n\tThis procedure used to erroneously send â\x80\x98CRLFâ\x80\x99 after each string\n\tspecified in BODY. Now, it sends BODY only.\n\n - â\x80\x98http:status-ok?â\x80\x99 also accepts numeric arg\n\n\tE.g., â\x80\x98(http:status-ok? 200)â\x80\x99 and â\x80\x98(http:status-ok? \"200\")â\x80\x99\n\tboth return â\x80\x98#tâ\x80\x99.\n\n \ - â\x80\x98http:requestâ\x80\x99 also accepts a symbolic method\n\n\tOf course, a string is still acceptable. Additionally, it is\n\tnow an error if â\x80\x98methodâ\x80\x99 is neither string nor symbol.\n\n - new procs: â\x80\x98send-requestâ\x80\x99, â\x80\x98receive-responseâ\x80\x99\n\n\tThese give (www http) a more asynchronous feel; it is now\n\tpossible to implement pipelining, connection re-use, etc.\n\tSupport for \"chunked\" transfer encoding and other HTTP 1.1 bits\n\tis built-in, although you can explicitly downgrade to 1.0 if you\n\twish.\n\n\tOther features: (optional) caller-directed i/o to reduce data\n\tmotion; support for â\x80\x98u8vectorâ\x80\x99 body (see example in the docs);\n\tsupport for custom body type; caller-specifiable header-name\n\tnormalization; procedure names w/o \"http:\" prefix :-D.\n\n - message field types now documented\n\n\tThe message object returned by â\x80\x98receive-responseâ\x80\x99 has a numeric\n\tstatus-code and header names normalized by the caller.\n\n\tIn contrast, for backward compatability, the message object\n\treturned by â\x80\x98http:requestâ\x80\x99 has a string status-code and\n\tlower-case header names.\n\n\t(Header names are symbolic in any case.)\n\n - new fluid: protocol-version\n\n\tThis affects how â\x80\x98http:requestâ\x80\x99 calls â\x80\x98send-requestâ\x80\x99 internally.\n\tIt also impacts â\x80\x98http:getâ\x80\x99, â\x80\x98http:headâ\x80\x99, and â\x80\x98http:post-formâ\x80\x99,\n\twhich call â\x80\x98http:requestâ\x80\x99.\n\n\tThe default value is â\x80\x98(1 . 0)â\x80\x99, i.e., HTTP 1.0. The idea is to\n\tsupport backward compatability out of the box, but allow users\n\tto experimentally do:\n\n\t (set-fluid! protocol-version '(1 . 1))\n\n\twithout having to modify their current code. (New code\n\tshould use â\x80\x98send-requestâ\x80\x99 and â\x80\x98receive-responseâ\x80\x99 directly.)\n\tThis is a transitional feature of Guile-WWW 2.x; 3.x will\n\tdrop â\x80\x98http:requestâ\x80\x99 (et al) along with the need for this\n\tfluid, entirely.\n\n - (www cgi) recognizes â\x80\x98application/x-www-form-urlencodedâ\x80\x99 better\n\n\tBefore, Content-Type recognition was over-constrained, failing\n\tto handle parameters (usually of the form \"; charset=...\").\n\tThings are now more relaxed.\n\n - documentation now explicitly UTF-8\n\n\tThis has no immediate impact, but if you regenerate the .info,\n\tit may be possible that the makeinfo you use will generate nicer\n\tquotes, e.g., â\x80\x98fooâ\x80\x99 instead of `foo'. Time will tell.\n\n - maintenance: Autoconf 2.68, Automake 1.11.3, Guile-BAUX 20111227\n" module: pkgsrc subject: 'CVS commit: pkgsrc/devel/guile-www' unixtime: '1352249406' user: gdt