[Bioc-devel] rtracklayer bug?

[Marc Carlson]

Hi Arne,

So this time when I look at the bioc-devel email list, I don't see a 
record for this last name (or this email).  In fact the only way I could 
be sure it was you was that your post was the same...  ;) If you want to 
post from gmail, then you will need to subscribe the gmail address to 
the list here:

https://stat.ethz.ch/mailman/listinfo/bioc-devel



  Marc



On 06/30/2015 02:26 AM, Arne Müller wrote:
> Hello,
>
>
> I think there’s a problem in UCSCSession initializer in rtracklayer:
>
> setMethod("initialize", "UCSCSession",
>
>            function(.Object, url =http://genome.ucsc.edu/cgi-bin/",
>
>                     user =ULL, session = NULL, force = FALSE, ...)
>
>            {
>
>              .Object at url <- url
>
>              .Object at views <- new.env()
>
>              gwURL <- ucscURL(.Object, "gateway")
>
>              if (force) {
>
>                  gwURL <- paste0(gwURL, '?redirect=anual"')
>
>              }
>
>              gw <- httpGet(gwURL, cookiefile =empfile(), header = TRUE,
>
>                            .parseúLSE)
>
>              if (grepl("redirectTd", gw)) {
>
>                  url <- sub(".*?a href=h([^[:space:]]+cgi-bin/).*",
> "h\\1", gw)
>
>                  return(initialize(.Object, url, user=er, session=session,
>
>                                    force=UE, ...))
>
>              }
>
>              cookie <- grep("Set-[Cc]ookie: hguid[^==", gw)
>
>              if (!length(cookie))
>
>                stop("Failed to obtain 'hguid' cookie")
>
>              hguid <- sub(".*Set-Cookie: (hguid[^==[^;]*);.*", "\\1", gw)
>
>              .Object at hguid <- hguid
>
>              if (!is.null(user) && !is.null(session)) { ## bring in other
> session
>
>                ucscGet(.Object, "tracks",
>
>                        list(hgS_doOtherUser =submit", hgS_otherUserName user,
>
>                             hgS_otherUserSessionName =ession))
>
>              }
>
>              .Object
>
>            })
>
>
>
> Shouldn’t ‘…’ be passed to httpGet that in turn is passed to RCURL, I.e.
>
>
> gw <- httpGet(gwURL, cookiefile =empfile(), header = TRUE,
>
>                            .parseúLSE, …) ?
>
> We run an internal instance of the UCSC genome browser and need to pass a
> cookie to all http-requests. The problem is that
>
> session =ew ('UCSCSession', url=myInternalURL, cookie=myAuthCookie)
>
>
> Does not pass the ‘cookie’ argument to httpGet.
>
>
> Regards,
>
>
> Arne
>
> 	[[alternative HTML version deleted]]
>