Download Class Reference

Class containing the state of a download. More...

#include <download.hh>


Public Member Functions
	Download (const string &uri, Output *o)
	~Download ()
void	setResumeOffset (uint64 offset)
	Set offset to resume from - download must not yet have been started, call before run().
uint64	resumeOffset () const
	Value passed to setResumeOffset().
void	run ()
	Whether to send a "Pragma: no-cache" header.
const string &	uri () const
void	pause ()
	Pause the request.
void	cont ()
	Continue with the download.
bool	paused () const
	Is the download paused? (i.e.
bool	failed () const
	Is the download paused, or is it not yet paused, but will be soon?
bool	succeeded () const
	Is download finished? (Also returns true if FTP/HTTP1.0 connection dropped).
bool	interrupted () const
	The connection was interrupted during the download, or it timed out.
void	stop ()
	Forcibly stop this download.
Output *	output () const
	Return the Output object.
void	setOutput (Output *o)
	Set the Output object.
Static Public Member Functions
static void	init ()
	Initialize libwww - call this before starting any downloads.
static void	cleanup ()
	Clean up - call this after all requests are finished.
Classes
class	Output
	A derived class of Output must be supplied by the user of Download, to take care of the downloaded data and to output progress reports. More...

Detailed Description

Class containing the state of a download.

libwww must be properly initialised before usage (this happens in glibwww-init.cc).

The purpose of this class is a minimal API for downloading HTTP and FTP URLs. A few useful extra features are added by Job::SingleURL.

The Output object is e.g. a GtkSingleUrl for single-file downloads

Constructor & Destructor Documentation

Download::Download	(	const string &	uri,
		Output *	o
	)

Download::~Download ( )

References Assert, debug, and glibcurl_remove().

Member Function Documentation

void Download::init ( ) [static]

Initialize libwww - call this before starting any downloads.

References debug, glibcurl_init(), glibcurl_set_callback(), info, and JIGDO_VERSION.

Referenced by main().

void Download::cleanup ( ) [static]

Clean up - call this after all requests are finished.

References debug, and glibcurl_cleanup().

Referenced by main().

void Download::setResumeOffset ( uint64 offset ) [inline]

Set offset to resume from - download must not yet have been started, call before run().

Caution: The offset is not reset after the download has finished/failed and will be reused if you re-run(), so better always* call this immediately before run().

Referenced by Job::SingleUrl::setResumeOffset().

uint64 Download::resumeOffset ( ) const [inline]

Value passed to setResumeOffset().

Referenced by run().

void Download::run ( )

Whether to send a "Pragma: no-cache" header.

The header is sent iff pragmaNoCache==true. Caution: The setting is not reset after the download has finished/failed and will be reused if you re-run(), so better *always* call this immediately before run(). Start downloading. Returns (almost) immediately and runs the download in the background via the glib/libwww event loop. If resumeOffset > 0, continue downloading the data from a given byte position. If a resume is not possible (e.g. server does not support HTTP ranges), there's an error.

References Assert, curlDebug, curlWriter(), debug, glibcurl_add(), and resumeOffset().

Referenced by Job::SingleUrl::run().

const string & Download::uri ( ) const [inline]

Referenced by Job::SingleUrl::location().

void Download::pause ( )

Pause the request.

Actually, this only sets a flag, which causes the request to be paused at the next convenient opportunity. The socket of the Download is no longer passed to select().

References Assert, debug, and glibcurl_remove().

Referenced by Job::SingleUrl::pause().

void Download::cont ( )

Continue with the download.

References Assert, debug, glibcurl_add(), and paused().

Referenced by Job::SingleUrl::cont().

bool Download::paused ( ) const [inline]

Is the download paused? (i.e.

really paused now, not just pause() called.)

Referenced by cont(), and Job::SingleUrl::paused().

bool Download::failed ( ) const [inline]

Is the download paused, or is it not yet paused, but will be soon?

Did the download fail with an error?

Referenced by Job::SingleUrl::failed().

bool Download::succeeded ( ) const [inline]

Is download finished? (Also returns true if FTP/HTTP1.0 connection dropped).

Referenced by Job::SingleUrl::succeeded().

bool Download::interrupted ( ) const [inline]

The connection was interrupted during the download, or it timed out.

Referenced by Job::SingleUrl::resumePossible().

void Download::stop ( )

Forcibly stop this download.

Careful, in case we are pipelining it is unavoidable that other (pending and not pending) downloads in the same pipeline also fail; they'll have to be resumed. [outofdate:stop() must not be called when libwww is delivering data, i.e. you must not call it from your download_data() method.] stop() is special in that it neither calls job_failed() nor job_succeeded(). Usually, you will have to call one or the other (depending on context) to notify everyone that the request was stopped.

References Assert, debug, glibcurl_remove(), and NULL.

Referenced by Job::SingleUrl::stop().

Download::Output * Download::output ( ) const [inline]

Return the Output object.

void Download::setOutput ( Download::Output * o ) [inline]

Set the Output object.

The documentation for this class was generated from the following files: