Calculating duration from a single uint64_t is simpler than from a struct
timeval. Change throughput measurement from gettimeofday() to
getnanotime().
Also calculate misec only if needed, and change integer division to integer
multiplication + shift, which should be slightly faster.
Signed-off-by: Karsten Blees <blees@dcon.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Humanization of downloaded size is done in the same function as text
formatting in 'process.c'. The code cannot be reused easily elsewhere.
Separate text formatting from size simplification and make the
function public in strbuf so that it can easily be used by other
callers.
We now can use strbuf_humanise_bytes() for both downloaded size and
download speed calculation. One of the drawbacks is that speed will
now look like this when download is stalled: "0 bytes/s" instead of
"0 KiB/s".
Signed-off-by: Antoine Pelisse <apelisse@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Switch to MiB/s when the connection is fast enough (i.e. on a LAN).
Signed-off-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Due to problems at cam.org, my nico@cam.org email address is no longer
valid. From now on, nico@fluxnic.net should be used instead.
Signed-off-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Often the throughput output is requested when the data read so far is
one smaller than multiple of 1024; because 1023/1024 is ~0.999, it often
shows up as 0.99 because the code currently truncates.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Dynamically sized arrays are gcc and C99 construct. Using them hurts
portability to older compilers, although using them is nice in this case
it is not desirable. This patch removes the only use of the construct
in stop_progress_msg(); the function is about writing out a single line
of a message, and the existing callers of this function feed messages
of only bounded size anyway, so use of dynamic array is simply overkill.
Signed-off-by: Boyd Lynn Gerber <gerberb@zenez.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This will make progress display from pack-objects (invoked via
upload-pack) more responsive on platforms with an implementation
of stdio whose stderr is line buffered.
The standard error stream is defined to be merely "not fully
buffered"; this is different from "unbuffered". If the
implementation of the stdio library chooses to make it line
buffered, progress reports that end with CR but do not contain
LF will accumulate in the stdio buffer before written out. A
fflush() after each progress display gives a nice continuous
progress.
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the same spirit of prettifying Git's output display for mere mortals,
here's a simple extension to the progress API allowing for a final
message to be provided when terminating a progress line, and use it for
the display of the number of objects needed to complete a thin pack,
saving yet one more line of screen display.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The minimum delay of 1/2 sec between successive throughput updates might
not have been elapsed when display_throughput() is called for the last
time, potentially making the display of total transferred bytes not
right when progress is said to be done.
Let's force an update of the throughput display as well when the
progress is complete. As a side effect, the total transferred will
always be displayed even if the actual transfer rate doesn't have time
to kickin.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The throughput display needs a delay period before accounting and
displaying anything. Yet it might be called after some amount of data
has already been transferred. The display of total data is therefore
accounted late and therefore smaller than the reality.
Let's call display_throughput() with an absolute amount of transferred
data instead of a relative number, and let the throughput code find the
relative amount of data by itself as needed. This way the displayed
total is always exact.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Right now it is infeasible to offer to the user a reasonable concept
of when a clone will be complete as we aren't able to come up with
the final pack size until after we have actually transferred the
entire thing to the client. However in many cases users can work
with a rough rule-of-thumb; for example it is somewhat well known
that git.git is about 16 MiB today and that linux-2.6.git is over
120 MiB.
We now show the total amount of data we have transferred over
the network as part of the throughput meter, organizing it in
"human friendly" terms like `ls -h` would do. Users can glance at
this, see that the total transferred size is about 3 MiB, see the
throughput of X KiB/sec, and determine a reasonable figure of about
when the clone will be complete, assuming they know the rough size
of the source repository or are able to obtain it.
This is also a helpful indicator that there is progress being made
even if we stall on a very large object. The thoughput meter may
remain relatively constant and the percentage complete and object
count won't be changing, but the total transferred will be increasing
as additional data is received for this object.
[from an initial proposal from Shawn O. Pearce]
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently the progress/throughput display update happens only through
display_progress(). If the progress based on object count remains
unchanged because a large object is being received, the latest throughput
won't be displayed. The display update should occur through
display_throughput() as well.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This adds the ability for the progress code to also display transfer
throughput when that makes sense.
The math was inspired by commit c548cf4ee0
from Linus.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This allows for better management of progress "object" existence,
as well as making the progress display implementation more independent
from its callers.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Each progress can be on a single line instead of two.
[sp: Changed "Checking files out" to "Checking out files" at
Johannes Sixt's suggestion as it better explains the
action that is taking place]
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
This allows for progress to be displayed only if the progress has not
reached a specified percentage treshold within a given delay in seconds.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
If the progress bar ends up in a box, better provide a title for it too.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
Instead of having this code duplicated in multiple places, let's have
a common interface for progress display. If someday someone wishes to
display a cheezy progress bar instead then only one file will have to
be changed.
Note: I left merge-recursive.c out since it has a strange notion of
progress as it apparently increase the expected total number as it goes.
Someone with more intimate knowledge of what that is supposed to mean
might look at converting it to the common progress interface.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
