Emacs25.1之后UrlHttpError
1 引言
当我们更新到Emacs25.1或者之后的版本后,原来可以使用的org2blog,或者其他类似使用metaweblog包的功能都无法使用了,会报一个"error: url-http-create-request: Multibyte text in HTTP request" 的错误信息。
这个问题出现之后,由于我用的是cnblogs插件,这个插件作者很久没有更新了,我首先想到的是这个插件是不是出问题了。然后,我把这个插件查找了,发现不是这个里面的问题。于是在google上搜索了一下,发现其他有好多人也在问这个问题(主要是日本的,说明emacs在日本使用普及度应该比中国要高很多)。
2 正文
在网上并没有找到解决方案,于是,我就决定自己去查看一下Emacs的url-http的源码:
c:/emacs/share/emacs/26.0.50/lisp/url/url-http.el.gz
然后,发现在第395行的位置,有这么一段代码:
;; Bug#23750 (unless (= (string-bytes request) (length request)) (error "Multibyte text in HTTP request: %s" request))
无疑就是这个地方报的错,为什么会报错呢?
去看一下之前版本的代码就知道了,原来是没有这几行代码的,作者也标明了,加入这几行代码是为了解决 Bug#23750,而这个bug我也去查了一下,应该是json.el中的一个bug,也就是说,如果代码不在这个地方检查报错,后面如果要使用到json中的函数的话,同样也会报错,并且据介绍,会更加隐晦,更难查找(去搜索json中的中文问题)。所以,作者在这里提前报错。
那这个错误到底是什么呢?
就是说 (string-bytes request) 和 (length request) 的返回值不一样的时候,就会报错。而中文字符的string-bytes,如果用utf-8来进行编码的话,字符长度是变化的,中文的编码平均大概是每个字会占用3个bytes,而 length 函数返回的会是1。那结果当然是不相等,当然会报错了。
怎么解决呢?一个简单的方法是:
;; Bug#23750 (setq request (url-http--encode-string request)) (unless (= (string-bytes request) (length request)) (error "Multibyte text in HTTP request: %s" request))
其中的 url-http--encode-string
也是作者新增的:
(defun url-http--encode-string (s) (if (multibyte-string-p s) (encode-coding-string s ‘us-ascii) s))
把传进来的中文先转换为‘us-ascii编码,这样,两种函数返回值就是相等的了。 当然,为什么可以这样,把中文转换为ascii,照理压根儿就不可行吧?这里没有想明白。但是可以工作了,是不是作用只是把这个判断跳过去了,后面使用json还是会报错?这个目前还没有去测试。 但是,确实是可以工作了,我目前的这个文章就是使用这个来发布的。
完整的函数,把它添加到配置文件中去就可以了:
(defun url-http-create-request (&optional ref-url) "Create an HTTP request for `url-http-target-url‘, referred to by REF-URL." (let* ((extra-headers) (request nil) (no-cache (cdr-safe (assoc "Pragma" url-http-extra-headers))) (using-proxy url-http-proxy) (proxy-auth (if (or (cdr-safe (assoc "Proxy-Authorization" url-http-extra-headers)) (not using-proxy)) nil (let ((url-basic-auth-storage ‘url-http-proxy-basic-auth-storage)) (url-get-authentication url-http-proxy nil ‘any nil)))) (real-fname (url-filename url-http-target-url)) (host (url-http--encode-string (url-host url-http-target-url))) (auth (if (cdr-safe (assoc "Authorization" url-http-extra-headers)) nil (url-get-authentication (or (and (boundp ‘proxy-info) proxy-info) url-http-target-url) nil ‘any nil)))) (if (equal "" real-fname) (setq real-fname "/")) (setq no-cache (and no-cache (string-match "no-cache" no-cache))) (if auth (setq auth (concat "Authorization: " auth "\r\n"))) (if proxy-auth (setq proxy-auth (concat "Proxy-Authorization: " proxy-auth "\r\n"))) ;; Protection against stupid values in the referrer (if (and ref-url (stringp ref-url) (or (string= ref-url "file:nil") (string= ref-url ""))) (setq ref-url nil)) ;; We do not want to expose the referrer if the user is paranoid. (if (or (memq url-privacy-level ‘(low high paranoid)) (and (listp url-privacy-level) (memq ‘lastloc url-privacy-level))) (setq ref-url nil)) ;; url-http-extra-headers contains an assoc-list of ;; header/value pairs that we need to put into the request. (setq extra-headers (mapconcat (lambda (x) (concat (car x) ": " (cdr x))) url-http-extra-headers "\r\n")) (if (not (equal extra-headers "")) (setq extra-headers (concat extra-headers "\r\n"))) ;; This was done with a call to `format‘. Concatenating parts has ;; the advantage of keeping the parts of each header together and ;; allows us to elide null lines directly, at the cost of making ;; the layout less clear. (setq request (concat ;; The request (or url-http-method "GET") " " (url-http--encode-string (if using-proxy (url-recreate-url url-http-target-url) real-fname)) " HTTP/" url-http-version "\r\n" ;; Version of MIME we speak "MIME-Version: 1.0\r\n" ;; (maybe) Try to keep the connection open "Connection: " (if (or using-proxy (not url-http-attempt-keepalives)) "close" "keep-alive") "\r\n" ;; HTTP extensions we support (if url-extensions-header (format "Extension: %s\r\n" url-extensions-header)) ;; Who we want to talk to (if (/= (url-port url-http-target-url) (url-scheme-get-property (url-type url-http-target-url) ‘default-port)) (format "Host: %s:%d\r\n" (puny-encode-domain host) (url-port url-http-target-url)) (format "Host: %s\r\n" (puny-encode-domain host))) ;; Who its from (if url-personal-mail-address (concat "From: " url-personal-mail-address "\r\n")) ;; Encodings we understand (if (or url-mime-encoding-string ;; MS-Windows loads zlib dynamically, so recheck ;; in case they made it available since ;; initialization in url-vars.el. (and (eq ‘system-type ‘windows-nt) (fboundp ‘zlib-available-p) (zlib-available-p) (setq url-mime-encoding-string "gzip"))) (concat "Accept-encoding: " url-mime-encoding-string "\r\n")) (if url-mime-charset-string (concat "Accept-charset: " (url-http--encode-string url-mime-charset-string) "\r\n")) ;; Languages we understand (if url-mime-language-string (concat "Accept-language: " url-mime-language-string "\r\n")) ;; Types we understand "Accept: " (or url-mime-accept-string "*/*") "\r\n" ;; User agent (url-http-user-agent-string) ;; Proxy Authorization proxy-auth ;; Authorization auth ;; Cookies (when (url-use-cookies url-http-target-url) (url-http--encode-string (url-cookie-generate-header-lines host real-fname (equal "https" (url-type url-http-target-url))))) ;; If-modified-since (if (and (not no-cache) (member url-http-method ‘("GET" nil))) (let ((tm (url-is-cached url-http-target-url))) (if tm (concat "If-modified-since: " (url-get-normalized-date tm) "\r\n")))) ;; Whence we came (if ref-url (concat "Referer: " ref-url "\r\n")) extra-headers ;; Length of data (if url-http-data (concat "Content-length: " (number-to-string (length url-http-data)) "\r\n")) ;; End request "\r\n" ;; Any data url-http-data)) ;; Bug#23750 (setq request (url-http--encode-string request)) (unless (= (string-bytes request) (length request)) (error "Multibyte text in HTTP request: %s" request)) (url-http-debug "Request is: \n%s" request) request))