X-Git-Url: https://git.xandkar.net/?a=blobdiff_plain;f=tt.rkt;h=19b61ec179af87900653f80ad63fb6474f32e29b;hb=9c34c974c9c5d324cd432499ae55b3e2b3f1b059;hp=2410f53ddead3bcd3cf421b2ee70f52028ef242b;hpb=d0a0e07370312fdb0ea04cfd33685f55d9e509ed;p=tt.git diff --git a/tt.rkt b/tt.rkt index 2410f53..19b61ec 100644 --- a/tt.rkt +++ b/tt.rkt @@ -3,10 +3,9 @@ (require openssl/sha1) (require racket/date) (require - net/http-client + net/head net/uri-codec - net/url-string - net/url-structs) + net/url) (require (prefix-in info: "info.rkt")) @@ -244,6 +243,8 @@ (: cache-dir Path-String) (define cache-dir (build-path tt-home-dir "cache")) +(define cache-object-dir (build-path cache-dir "objects")) + (: url->cache-file-path-v1 (-> Url Path-String)) (define (url->cache-file-path-v1 uri) (define (hash-sha1 str) : (-> String String) @@ -251,13 +252,19 @@ (define digest (sha1 in)) (close-input-port in) digest) - (build-path cache-dir (hash-sha1 (url->string uri)))) + (build-path cache-object-dir (hash-sha1 (url->string uri)))) (: url->cache-file-path-v2 (-> Url Path-String)) (define (url->cache-file-path-v2 uri) - (build-path cache-dir (uri-encode (url->string uri)))) + (build-path cache-object-dir (uri-encode (url->string uri)))) + +(define url->cache-object-path url->cache-file-path-v2) + +(define (url->cache-etag-path uri) + (build-path cache-dir "etags" (uri-encode (url->string uri)))) -(define url->cache-file-path url->cache-file-path-v2) +(define (url->cache-lmod-path uri) + (build-path cache-dir "lmods" (uri-encode (url->string uri)))) ; TODO Return Option (: uri-read-cached (-> Url String)) @@ -304,6 +311,43 @@ (log-error "File does not exist: ~v" (path->string file-path)) '()))) +(define re-rfc2822 + #px"^(Mon|Tue|Wed|Thu|Fri|Sat|Sun), ([0-9]{2}) (Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec) ([0-9]{4}) ([0-2][0-9]):([0-6][0-9]):([0-6][0-9]) GMT") + +(: b->n (-> Bytes (Option Number))) +(define (b->n b) + (string->number (bytes->string/utf-8 b))) + +(: mon->num (-> Bytes Natural)) +(define/match (mon->num mon) + [(#"Jan") 1] + [(#"Feb") 2] + [(#"Mar") 3] + [(#"Apr") 4] + [(#"May") 5] + [(#"Jun") 6] + [(#"Jul") 7] + [(#"Aug") 8] + [(#"Sep") 9] + [(#"Oct") 10] + [(#"Nov") 11] + [(#"Dec") 12]) + +(: rfc2822->epoch (-> Bytes (Option Nonnegative-Integer))) +(define (rfc2822->epoch timestamp) + (match (regexp-match re-rfc2822 timestamp) + [(list _ _ dd mo yyyy HH MM SS) + #:when (and dd mo yyyy HH MM SS) + (find-seconds (b->n SS) + (b->n MM) + (b->n HH) + (b->n dd) + (mon->num mo) + (b->n yyyy) + #f)] + [_ + #f])) + (: user-agent String) (define user-agent (let* @@ -319,38 +363,65 @@ (format "+~a" prog-uri))]) (format "~a/~a (~a)" prog-name prog-version user))) +(: header-get (-> (Listof Bytes) Bytes (Option Bytes))) +(define (header-get headers name) + (match (filter-map (curry extract-field name) headers) + [(list val) val] + [_ #f])) + (: uri-download (-> Url Void)) (define (uri-download u) - (define cache-file-path (url->cache-file-path u)) - (log-debug "uri-download ~v into ~v" u cache-file-path) - (match* ((url-scheme u) (url-host u) (url-port u)) - [(s h p) - #:when (and s h) - (define ssl? (string=? s "https")) - (define-values (status-line headers body-input) - ; TODO Timeout. Currently hangs on slow connections. - (http-sendrecv - h - (url->string (struct-copy url u [scheme #f] [host #f])) - #:ssl? ssl? - #:port (cond [p p] [ssl? 443] [else 80]) - #:headers (list (format "User-Agent: ~a" user-agent)) - )) - (log-debug "headers: ~v" headers) - (log-debug "status-line: ~v" status-line) - (define status - (string->number (second (string-split (bytes->string/utf-8 status-line))))) - (log-debug "status: ~v" status) - ; TODO Handle redirects - (if (= 200 status) - (begin - (make-parent-directory* cache-file-path) - (call-with-output-file cache-file-path - (curry copy-port body-input) - #:exists 'replace)) - (raise status))] - [(_ _ _) - (log-error "Invalid URI: ~v" u)])) + (define cached-object-path (url->cache-object-path u)) + (define cached-etag-path (url->cache-etag-path u)) + (define cached-lmod-path (url->cache-lmod-path u)) + (log-debug "uri-download ~v into ~v" u cached-object-path) + (define-values (status-line headers body-input) + ; TODO Timeout. Currently hangs on slow connections. + (http-sendrecv/url u #:headers (list (format "User-Agent: ~a" user-agent)))) + (log-debug "headers: ~v" headers) + (log-debug "status-line: ~v" status-line) + (define status + (string->number (second (string-split (bytes->string/utf-8 status-line))))) + (log-debug "status: ~v" status) + ; TODO Handle redirects + (match status + [200 + (let* ([etag (header-get headers #"ETag")] + [lmod (header-get headers #"Last-Modified")] + [lmod-curr (if lmod (rfc2822->epoch lmod) #f)] + [lmod-prev (if (file-exists? cached-lmod-path) + (rfc2822->epoch (file->bytes cached-lmod-path)) + #f)]) + (log-debug "lmod-curr:~v lmod-prev:~v" lmod-curr lmod-prev) + (unless (or (and etag + (file-exists? cached-etag-path) + (bytes=? etag (file->bytes cached-etag-path)) + (begin + (log-info "ETags match, skipping the rest of ~v" (url->string u)) + #t)) + (and lmod-curr + lmod-prev + (<= lmod-curr lmod-prev) + (begin + (log-info "Last-Modified <= current skipping the rest of ~v" (url->string u)) + #t))) + (begin + (log-info + "Downloading the rest of ~v. ETag: ~a, Last-Modified: ~v" + (url->string u) etag lmod) + (make-parent-directory* cached-object-path) + (make-parent-directory* cached-etag-path) + (make-parent-directory* cached-lmod-path) + (call-with-output-file cached-object-path + (curry copy-port body-input) + #:exists 'replace) + (when etag + (display-to-file etag cached-etag-path #:exists 'replace)) + (when lmod + (display-to-file lmod cached-lmod-path #:exists 'replace)))) + (close-input-port body-input))] + [_ + (raise status)])) (: timeline-print (-> Out-Format (Listof Msg) Void)) (define (timeline-print out-format timeline)