(U 'old->new
'new->old))
+(define-type Result
+ (∀ (α β) (U (cons 'ok α)
+ (cons 'error β))))
+
(struct Msg
([ts-epoch : Integer]
[ts-orig : String]
[mentions : (Listof Peer)]))
(struct Peer
- ([nick : (Option String)]
- [uri : Url])
+ ([nick : (Option String)]
+ [uri : Url]
+ [comment : (Option String)])
#:transparent)
(struct Resp
([status-line : String]
[headers : (Listof Bytes)]
- [body-input : Input-Port]))
+ [body-input : Input-Port])
+ #:transparent)
(: tt-home-dir Path-String)
(define tt-home-dir (build-path (expand-user-path "~") ".tt"))
(define (url->cache-lmod-path uri)
(build-path cache-dir "lmods" (uri-encode (url->string uri))))
-; TODO Return Option
-(: uri-read-cached (-> Url String))
+(: uri-read-cached (-> Url (Option String)))
(define (uri-read-cached uri)
(define path-v1 (url->cache-file-path-v1 uri))
(define path-v2 (url->cache-file-path-v2 uri))
(file->string path-v2)
(begin
(log-warning "Cache file not found for URI: ~a" (url->string uri))
- "")))
+ #f)))
-(: uri? (-> String Boolean))
-(define (uri? str)
- (regexp-match? #rx"^[a-z]+://.*" (string-downcase str)))
+(: str->url (-> String (Option String)))
+(define (str->url s)
+ (with-handlers*
+ ([exn:fail? (λ (e) #f)])
+ (string->url s)))
(: str->peer (String (Option Peer)))
(define (str->peer str)
(log-debug "Parsing peer string: ~v" str)
- (with-handlers*
- ([exn:fail?
- (λ (e)
- (log-error "Invalid URI in string: ~v, exn: ~v" str e)
- #f)])
- (match (string-split str)
- [(list u) #:when (uri? u) (Peer #f (string->url u))]
- [(list n u) #:when (uri? u) (Peer n (string->url u))]
- [_
- (log-error "Invalid peer string: ~v" str)
- #f])))
+ (match
+ (regexp-match
+ #px"(([^\\s\t]+)[\\s\t]+)?([a-zA-Z]+://[^\\s\t]*)[\\s\t]*(#\\s*(.*))?"
+ str)
+ [(list _wholething
+ _nick-with-space
+ nick
+ url
+ _comment-with-hash
+ comment)
+ (match (str->url url)
+ [#f
+ (log-error "Invalid URI in peer string: ~v" str)
+ #f]
+ [url (Peer nick url comment)])]
+ [_
+ (log-error "Invalid peer string: ~v" str)
+ #f]))
+(module+ test
+ (check-equal?
+ (str->peer "foo http://bar/file.txt # some rando")
+ (Peer "foo" (str->url "http://bar/file.txt") "some rando"))
+ (check-equal?
+ (str->peer "http://bar/file.txt # some rando")
+ (Peer #f (str->url "http://bar/file.txt") "some rando"))
+ (check-equal?
+ (str->peer "http://bar/file.txt #")
+ (Peer #f (str->url "http://bar/file.txt") ""))
+ (check-equal?
+ (str->peer "http://bar/file.txt#") ; XXX URLs can have #s
+ (Peer #f (str->url "http://bar/file.txt#") #f))
+ (check-equal?
+ (str->peer "http://bar/file.txt")
+ (Peer #f (str->url "http://bar/file.txt") #f))
+ (check-equal?
+ (str->peer "foo http://bar/file.txt")
+ (Peer "foo" (str->url "http://bar/file.txt") #f))
+ (check-equal?
+ (str->peer "foo bar # baz")
+ #f)
+ (check-equal?
+ (str->peer "foo bar://baz # quux")
+ (Peer "foo" (str->url "bar://baz") "quux"))
+ (check-equal?
+ (str->peer "foo bar//baz # quux")
+ #f))
(: filter-comments (-> (Listof String) (Listof String)))
(define (filter-comments lines)
(define (peers->file peers path)
(display-lines-to-file
(map (match-lambda
- [(Peer n u)
- (format "~a~a" (if n (format "~a " n) "") (url->string u))])
+ [(Peer n u c)
+ (format "~a~a~a"
+ (if n (format "~a " n) "")
+ (url->string u)
+ (if c (format " # ~a" c) ""))])
peers)
path
#:exists 'replace))
[user
(if (file-exists? user-peer-file)
(match (first (file->peers user-peer-file))
- [(Peer #f u) (format "+~a" (url->string u) )]
- [(Peer n u) (format "+~a; @~a" (url->string u) n)])
+ [(Peer #f u _) (format "+~a" (url->string u) )]
+ [(Peer n u _) (format "+~a; @~a" (url->string u) n)])
(format "+~a" prog-uri))])
(format "~a/~a (~a)" prog-name prog-version user)))
[(list val) val]
[_ #f]))
-(: uri-download (-> Positive-Float Url Void))
-(define (uri-download timeout u)
+(: uri-download-from-port
+ (-> Url (Listof (U Bytes String)) Input-Port
+ (U 'skipped-cached 'downloaded-new))) ; TODO 'ok|'error ?
+(define (uri-download-from-port u headers body-input)
+ (define u-str (url->string u))
+ (log-debug "uri-download-from-port ~v into ~v" u-str cached-object-path)
(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))
+ (define etag (header-get headers #"ETag"))
+ (define lmod (header-get headers #"Last-Modified"))
+ (define lmod-curr (if lmod (rfc2822->epoch lmod) #f))
+ (define 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)
+ (define cached?
+ (or (and etag
+ (file-exists? cached-etag-path)
+ (bytes=? etag (file->bytes cached-etag-path))
+ (begin
+ (log-debug "ETags match, skipping the rest of ~v" u-str)
+ #t))
+ (and lmod-curr
+ lmod-prev
+ (<= lmod-curr lmod-prev)
+ (begin
+ (log-debug "Last-Modified <= current skipping the rest of ~v" u-str)
+ #t))))
+ (if (not cached?)
+ (begin
+ (log-debug
+ "Downloading the rest of ~v. ETag: ~a, Last-Modified: ~v"
+ u-str 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))
+ 'downloaded-new)
+ 'skipped-cached))
+
+(: uri-download
+ (-> Positive-Float Url
+ (Result (U 'skipped-cached 'downloaded-new)
+ Any))) ; TODO Maybe more-precise error type?
+(define (uri-download timeout u)
(define u-str (url->string u))
- (log-debug "uri-download ~v into ~v" u-str cached-object-path)
(define timeout-chan (make-channel))
(define result-chan (make-channel))
(define timeout-thread
; Doing this instead of sync/timeout to distinguish error values,
; rather than just have #f to work with.
(sleep timeout)
- (channel-put timeout-chan (cons 'error 'timeout)))))
+ (channel-put timeout-chan '(error . timeout)))))
(define result-thread
(thread (λ ()
; XXX We timeout getting a response, but body download could
; and read-from disk as a whole.
(define result
(with-handlers
- ([exn:fail? (λ (e) (cons 'error (cons 'net-error e)))])
+ ; TODO Maybe name each known errno? (exn:fail:network:errno-errno e)
+ ([exn:fail:network?
+ (λ (e) `(error . (net-error . ,e)))]
+ [exn?
+ (λ (e) `(error . (other . ,e)))])
(define-values (status-line headers body-input)
(http-sendrecv/url
u
#:headers (list (format "User-Agent: ~a" user-agent))))
- (cons 'ok (Resp status-line headers body-input))))
+ `(ok . ,(Resp status-line headers body-input))))
(channel-put result-chan result))))
(define result
(sync timeout-chan
(kill-thread result-thread)
(kill-thread timeout-thread)
(match result
- [(cons 'error 'timeout)
- (log-error "Download failed: timeout. URL:~v" u-str)]
- [(cons 'error (cons 'net-error e))
- (log-error "Download failed. Network error. URL:~v EXN:~v" u-str e)]
+ [(cons 'error _)
+ result]
[(cons 'ok (Resp status-line headers body-input))
(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" u-str)
- #t))
- (and lmod-curr
- lmod-prev
- (<= lmod-curr lmod-prev)
- (begin
- (log-info "Last-Modified <= current skipping the rest of ~v" u-str)
- #t)))
- (begin
- (log-info
- "Downloading the rest of ~v. ETag: ~a, Last-Modified: ~v"
- u-str 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))]
- [_
- (log-error "HTTP error URL:~a status:~a" u-str status)])]))
+ ; TODO Handle redirects. Should be within same timeout as req and body.
+ (let ([result
+ (match status
+ [200
+ `(ok . ,(uri-download-from-port u headers body-input))]
+ [_
+ `(error . (http . ,status))])])
+ (close-input-port body-input)
+ result)]))
(: timeline-print (-> Out-Format (Listof Msg) Void))
(define (timeline-print out-format timeline)
timeline)))
(: peer->msgs (-> Peer (Listof Msg)))
-(define (peer->msgs f)
- (match-define (Peer nick uri) f)
+(define (peer->msgs peer)
+ (match-define (Peer nick uri _) peer)
(log-info "Reading peer nick:~v uri:~v" nick (url->string uri))
- (str->msgs nick uri (uri-read-cached uri)))
-
-(: peer-download (-> Positive-Float Peer Void))
+ (define msgs-data (uri-read-cached uri))
+ (if msgs-data
+ (str->msgs nick uri msgs-data)
+ '()))
+
+(: peer-download
+ (-> Positive-Float Peer
+ (Result (U 'skipped-cached 'downloaded-new)
+ Any)))
(define (peer-download timeout peer)
- (match-define (Peer nick uri) peer)
+ (match-define (Peer nick uri _) peer)
(define u (url->string uri))
- (log-info "Download BEGIN uri:~a" u)
- (define-values (_result _tm-cpu-ms tm-real-ms _tm-gc-ms)
+ (log-info "Download BEGIN URL:~a" u)
+ (define-values (results _tm-cpu-ms tm-real-ms _tm-gc-ms)
(time-apply uri-download (list timeout uri)))
- (log-info "Download END in ~a seconds, uri: ~a" (/ tm-real-ms 1000.0) u))
+ (define result (car results))
+ (log-info "Download END in ~a seconds, URL:~a, result:~s"
+ (/ tm-real-ms 1000.0)
+ u
+ result)
+ result)
(: timeline-download (-> Integer Positive-Float (Listof Peer) Void))
(define (timeline-download num-workers timeout peers)
- ; TODO No need for map - can just iter
- (void (concurrent-filter-map num-workers (λ (p) (peer-download timeout p)) peers)))
+ (define results
+ (concurrent-filter-map num-workers
+ (λ (p) (cons p (peer-download timeout p)))
+ peers))
+ (define peers-ok
+ (filter-map (match-lambda
+ [(cons p (cons 'ok _)) p]
+ [(cons _ (cons 'error e)) #f])
+ results))
+ (define peers-err
+ (filter-map (match-lambda
+ [(cons _ (cons 'ok _))
+ #f]
+ [(cons p (cons 'error e))
+ (struct-copy Peer p [comment (format "~s" e)])])
+ results))
+ (peers->file peers-ok (build-path tt-home-dir "peers-last-downloaded-ok"))
+ (peers->file peers-err (build-path tt-home-dir "peers-last-downloaded-err")))
(: uniq (∀ (α) (-> (Listof α) (Listof α))))
(define (uniq xs)
#:args ()
(let* ([peers-sort
(λ (peers) (sort peers (match-lambda**
- [((Peer n1 _) (Peer n2 _))
+ [((Peer n1 _ _) (Peer n2 _ _))
(string<? (if n1 n1 "")
(if n2 n2 ""))])))]
[peers-all-file
(build-path tt-home-dir "peers-all")]
[peers-mentioned-file
(build-path tt-home-dir "peers-mentioned")]
+ [peers-parsed-file
+ (build-path tt-home-dir "peers-parsed")]
[peers-mentioned-curr
(mentioned-peers-in-cache)]
[peers-mentioned-prev
[peers-discovered
(set-subtract peers-all (list->set peers-all-prev))]
[peers-all
- (peers-sort (set->list peers-all))])
+ (peers-sort (set->list peers-all))]
+ [peers-parsed
+ (filter
+ (λ (p) (< 0 (length (peer->msgs p))))
+ peers-all)])
(log-info "Known peers mentioned: ~a" (length peers-mentioned))
+ (log-info "Known peers parsed ~a" (length peers-parsed))
(log-info "Known peers total: ~a" (length peers-all))
(log-info "Discovered ~a new peers:~n~a"
(set-count peers-discovered)
(set->list peers-discovered))))
(peers->file peers-mentioned
peers-mentioned-file)
+ (peers->file peers-parsed
+ peers-parsed-file)
(peers->file peers-all
peers-all-file)))]
[command