Hi Silvio,

Not a big surprise to me, the current G'MIC filter which splits colors into layers works in a quite straightforward way, and if you give it an anti-aliased image, it will definitely find a lot of different colors in it (tens or hundreds, even if there is only a single 'color' for you ). This filter also outputs one float-valued image (layer) by color, and 32 layers with the resolution you used already requires a lot of memory. What happens is probably you ran out of memory, and it started swapping.
I think this filter should be improved in the future, with a possible 'tolerance' parameter that would tell that a color can appear several time with slight variations while being still considered as a single color.
I'll put this on my TODO list

Hi David,

> I think this filter should be improved in the future, with a possible 'tolerance' parameter that would tell that a color can appear several time with slight variations while being still considered as a single color.
I'll put this on my TODO list

Thanks a lot indeed ! :-)

When applying a filter on a selection and with 'new layers' selected, the canvas in gimp is resized to the section size. Nothing is lost, but you need to use "fit canvas to layers" to get back to where you were.

Iain, ha thanks for noticing this, I'll fix that as soon as possible!

Hello David,

I can break -gmic_survey by populating the text widgets sufficiently so that the URL string being handed off to -command is around 200 bytes or more. Small URL strings 200 bytes or less, seem to work fine.

See the attached gmicdump.pdf, from the console in verbose(debug) mode. As you can see, I wrote a pretty verbose response (typical tech writer running off the mouth). That created an raw image 1305 bytes long. I make it through -str2hex but failed in -command. (at line 234 - Page 6 - in gmicdump.pdf) It looks like a sub-shell was started and its parser was having trouble seeing a closing double quote. Running wireshark, I never even saw the process open a connection and attempt to write out a packet to gmic.eu. Even small strings in the text boxes caused a break at the invocation of -command. Made about three attempts with increasingly smaller responses. The dump is a rerun of the first attempt with long responses in the text boxes.

Finally leaving all text fields empty got me success. An outbound 188 byte packet went to gmic.eu and the response came right back.

The .update1610.gmic file was pulled down about an hour before this post. gimp_gmic was built about four hours ago against this commit:

commit c3e5b0178d65dbce8c6cdafd890f383594decae4
Author: ronounours <David.Tschumperle@ensicaen.fr>
Date: Sat Mar 7 18:26:50 2015 +0100

Didn't chase the rabbit into CImg.h to see what was going on there. At least, in this odd, roundabout way you got my survey response. So much for anonymity, but you'd probably guess who it was anyway.

Cheers!

Garry

Ah, very good Garry, thanks for discovering this one !
I've been able to fix it, was a problem in the CImg library source code.

Hello David,

Short answer: Success!

Long answer:
rebuilt everything against this:
commit f45e7858d527ae5c4b4dd4f8756b91b1c9fc6fe2
Author: ronounours <David.Tschumperle@ensicaen.fr>
Date: Mon Mar 9 09:06:57 2015 +0100
Updated .update1610.gmic
Removed the old lock file
Sent the same long-winded survey that failed Sunday.

A 2,740 byte packet left here for gmic.eu at 06:29:02.42 EDT (11:29:02.42 CET) [Frame 4]
A 60 byte packet came back from gmic.eu at 06:29:02.61 (11:29:02.61)[Frame 9]
The preview pane thanked me very much.
Everything seems good to go for verbose, long-winded respondents.
Garry

Frame 4:
2740 bytes on wire (21920 bits), 2740 bytes captured (21920 bits) on interface 0
Interface id: 0 (eth0)
Encapsulation type: Ethernet (1)
Arrival Time: Mar 9, 2015 06:29:02.415444000 EDT
[Time shift for this packet: 0.000000000 seconds]
Epoch Time: 1425896942.415444000 seconds
[Time delta from previous captured frame: 0.000154000 seconds]
[Time delta from previous displayed frame: 0.000154000 seconds]
[Time since reference or first frame: 0.110428000 seconds]
Frame Number: 4
Frame Length: 2740 bytes (21920 bits)
Capture Length: 2740 bytes (21920 bits)
[Frame is marked: False]
[Frame is ignored: False]
[Protocols in frame: eth:ethertype:ip:tcp:http]
[Number of per-protocol-data: 1]
[Hypertext Transfer Protocol, key 0]
[Coloring Rule Name: HTTP]
[Coloring Rule String: http || tcp.port == 80 || http2]
Ethernet II, Src: DellInc_a1:f5:b0 (00:13:72:a1:f5:b0), Dst: Cisco-Li_4e:6d:1f (00:0f:66:4e:6d:1f)
Destination: Cisco-Li_4e:6d:1f (00:0f:66:4e:6d:1f)
Address: Cisco-Li_4e:6d:1f (00:0f:66:4e:6d:1f)
.... ..0. .... .... .... .... = LG bit: Globally unique address (factory default)
.... ...0 .... .... .... .... = IG bit: Individual address (unicast)
Source: DellInc_a1:f5:b0 (00:13:72:a1:f5:b0)
Address: DellInc_a1:f5:b0 (00:13:72:a1:f5:b0)
.... ..0. .... .... .... .... = LG bit: Globally unique address (factory default)
.... ...0 .... .... .... .... = IG bit: Individual address (unicast)
Type: IP (0x0800)
Internet Protocol Version 4, Src: 192.168.2.211 (192.168.2.211), Dst: 213.186.33.40 (213.186.33.40)
Version: 4
Header Length: 20 bytes
Differentiated Services Field: 0x00 (DSCP 0x00: Default; ECN: 0x00: Not-ECT (Not ECN-Capable Transport))
0000 00.. = Differentiated Services Codepoint: Default (0x00)
.... ..00 = Explicit Congestion Notification: Not-ECT (Not ECN-Capable Transport) (0x00)
Total Length: 2726
Identification: 0xc837 (51255)
Flags: 0x02 (Don't Fragment)
0... .... = Reserved bit: Not set
.1.. .... = Don't fragment: Set
..0. .... = More fragments: Not set
Fragment offset: 0
Time to live: 64
Protocol: TCP (6)
Header checksum: 0xadbc [validation disabled]
[Good: False]
[Bad: False]
Source: 192.168.2.211 (192.168.2.211)
Destination: 213.186.33.40 (213.186.33.40)
Transmission Control Protocol, Src Port: 45867 (45867), Dst Port: http (80), Seq: 1, Ack: 1, Len: 2686
Source Port: 45867 (45867)
Destination Port: http (80)
[Stream index: 0]
[TCP Segment Len: 2686]
Sequence number: 1 (relative sequence number)
[Next sequence number: 2687 (relative sequence number)]
Acknowledgment number: 1 (relative ack number)
Header Length: 20 bytes
.... 0000 0001 1000 = Flags: 0x018 (PSH, ACK)
000. .... .... = Reserved: Not set
...0 .... .... = Nonce: Not set
.... 0... .... = Congestion Window Reduced (CWR): Not set
.... .0.. .... = ECN-Echo: Not set
.... ..0. .... = Urgent: Not set
.... ...1 .... = Acknowledgment: Set
.... .... 1... = Push: Set
.... .... .0.. = Reset: Not set
.... .... ..0. = Syn: Not set
.... .... ...0 = Fin: Not set
Window size value: 29200
[Calculated window size: 29200]
[Window size scaling factor: -2 (no window scaling used)]
Checksum: 0xc4f6 [validation disabled]
[Good Checksum: False]
[Bad Checksum: False]
Urgent pointer: 0
[SEQ/ACK analysis]
[iRTT: 0.110274000 seconds]
[Bytes in flight: 2686]
Hypertext Transfer Protocol
[truncated]GET /gmic_survey.cgi?362c312c312c312c312c312c312c312c312c312c302c322c302c322c302c302c54686520766572792073616e65206f7267616e697a6174696f6e206f66207468652067276d696320434c4920636f6d6d616e64206c696e652c54686520434d594b20636f6c6f72
[ [truncated]Expert Info (Chat/Sequence): GET /gmic_survey.cgi?362c312c312c312c312c312c312c312c312c312c302c322c302c322c302c302c54686520766572792073616e65206f7267616e697a6174696f6e206f66207468652067276d696320434c4920636f6d6d616e64206c696e652]
[GET /gmic_survey.cgi?362c312c312c312c312c312c312c312c312c312c302c322c302c322c302c302c54686520766572792073616e65206f7267616e697a6174696f6e206f66207468652067276d696320434c4920636f6d6d616e64206c696e652c54686520434d594b20636f6c6f72207370616365]
[Severity level: Chat]
[Group: Sequence]
Request Method: GET
Request URI [truncated]: /gmic_survey.cgi?362c312c312c312c312c312c312c312c312c312c302c322c302c322c302c302c54686520766572792073616e65206f7267616e697a6174696f6e206f66207468652067276d696320434c4920636f6d6d616e64206c696e652c54686520434d594b206
Request Version: HTTP/1.1
Host: gmic.eu\r\n
Accept: */*\r\n
\r\n
[Full request URI [truncated]: http://gmic.eu/gmic_survey.cgi?362c312c ... 6c696e652c]
[HTTP request 1/1]
[Response in frame: 9]

Frame 9:
Frame 9: 60 bytes on wire (480 bits), 60 bytes captured (480 bits) on interface 0
Interface id: 0 (eth0)
Encapsulation type: Ethernet (1)
Arrival Time: Mar 9, 2015 06:29:02.615850000 EDT
[Time shift for this packet: 0.000000000 seconds]
Epoch Time: 1425896942.615850000 seconds
[Time delta from previous captured frame: 0.000430000 seconds]
[Time delta from previous displayed frame: 0.000430000 seconds]
[Time since reference or first frame: 0.310834000 seconds]
Frame Number: 9
Frame Length: 60 bytes (480 bits)
Capture Length: 60 bytes (480 bits)
[Frame is marked: False]
[Frame is ignored: False]
[Protocols in frame: eth:ethertype:ip:tcp:http:data:data:media]
[Number of per-protocol-data: 1]
[Hypertext Transfer Protocol, key 0]
[Coloring Rule Name: HTTP]
[Coloring Rule String: http || tcp.port == 80 || http2]
Ethernet II, Src: Cisco-Li_4e:6d:1f (00:0f:66:4e:6d:1f), Dst: DellInc_a1:f5:b0 (00:13:72:a1:f5:b0)
Destination: DellInc_a1:f5:b0 (00:13:72:a1:f5:b0)
Address: DellInc_a1:f5:b0 (00:13:72:a1:f5:b0)
.... ..0. .... .... .... .... = LG bit: Globally unique address (factory default)
.... ...0 .... .... .... .... = IG bit: Individual address (unicast)
Source: Cisco-Li_4e:6d:1f (00:0f:66:4e:6d:1f)
Address: Cisco-Li_4e:6d:1f (00:0f:66:4e:6d:1f)
.... ..0. .... .... .... .... = LG bit: Globally unique address (factory default)
.... ...0 .... .... .... .... = IG bit: Individual address (unicast)
Type: IP (0x0800)
Padding: 00
Internet Protocol Version 4, Src: 213.186.33.40 (213.186.33.40), Dst: 192.168.2.211 (192.168.2.211)
Version: 4
Header Length: 20 bytes
Differentiated Services Field: 0x00 (DSCP 0x00: Default; ECN: 0x00: Not-ECT (Not ECN-Capable Transport))
0000 00.. = Differentiated Services Codepoint: Default (0x00)
.... ..00 = Explicit Congestion Notification: Not-ECT (Not ECN-Capable Transport) (0x00)
Total Length: 45
Identification: 0x26af (9903)
Flags: 0x00
0... .... = Reserved bit: Not set
.0.. .... = Don't fragment: Not set
..0. .... = More fragments: Not set
Fragment offset: 0
Time to live: 243
Protocol: TCP (6)
Header checksum: 0xe6bd [validation disabled]
[Good: False]
[Bad: False]
Source: 213.186.33.40 (213.186.33.40)
Destination: 192.168.2.211 (192.168.2.211)
Transmission Control Protocol, Src Port: http (80), Dst Port: 45867 (45867), Seq: 435, Ack: 2687, Len: 5
Source Port: http (80)
Destination Port: 45867 (45867)
[Stream index: 0]
[TCP Segment Len: 5]
Sequence number: 435 (relative sequence number)
[Next sequence number: 440 (relative sequence number)]
Acknowledgment number: 2687 (relative ack number)
Header Length: 20 bytes
.... 0000 0001 1000 = Flags: 0x018 (PSH, ACK)
000. .... .... = Reserved: Not set
...0 .... .... = Nonce: Not set
.... 0... .... = Congestion Window Reduced (CWR): Not set
.... .0.. .... = ECN-Echo: Not set
.... ..0. .... = Urgent: Not set
.... ...1 .... = Acknowledgment: Set
.... .... 1... = Push: Set
.... .... .0.. = Reset: Not set
.... .... ..0. = Syn: Not set
.... .... ...0 = Fin: Not set
Window size value: 10164
[Calculated window size: 10164]
[Window size scaling factor: -2 (no window scaling used)]
Checksum: 0xaa52 [validation disabled]
[Good Checksum: False]
[Bad Checksum: False]
Urgent pointer: 0
[SEQ/ACK analysis]
[iRTT: 0.110274000 seconds]
[Bytes in flight: 5]
TCP segment data (5 bytes)
[2 Reassembled TCP Segments (439 bytes): #7(434), #9(5)]
[Frame: 7, payload: 0-433 (434 bytes)]
[Frame: 9, payload: 434-438 (5 bytes)]
[Segment count: 2]
[Reassembled TCP length: 439]
[Reassembled TCP Data: 485454502f312e3120323030204f4b0d0a5365742d436f6f...]
Hypertext Transfer Protocol
HTTP/1.1 200 OK\r\n
[Expert Info (Chat/Sequence): HTTP/1.1 200 OK\r\n]
[HTTP/1.1 200 OK\r\n]
[Severity level: Chat]
[Group: Sequence]
Request Version: HTTP/1.1
Status Code: 200
Response Phrase: OK
Set-Cookie: 300gpBAK=R4178755467; path=/; expires=Mon, 09-Mar-2015 11:50:05 GMT\r\n
Content-Type: text\r\n
Set-Cookie: 300gp=R3395845342; path=/; expires=Mon, 09-Mar-2015 11:43:08 GMT\r\n
Server: Apache\r\n
Transfer-Encoding: chunked\r\n
Date: Mon, 09 Mar 2015 10:29:02 GMT\r\n
Connection: keep-alive\r\n
X-Geo: varn07.rbx5\r\n
X-Geo-Port: 1011\r\n
X-Cacheable: Not cacheable: no cache headers from backend\r\n
\r\n
[HTTP response 1/1]
[Time since request: 0.200406000 seconds]
[Request in frame: 4]
HTTP chunked response
Data chunk (26 octets)
Chunk size: 26 octets
Data (26 bytes)

0000 23 40 67 6d 69 63 0a 67 6d 69 63 5f 73 75 72 76 #@gmic.gmic_surv
0010 65 79 5f 63 68 65 63 6b 3a 0a ey_check:.
Data: 2340676d69630a676d69635f7375727665795f636865636b...
[Length: 26]
Chunk boundary
End of chunked encoding
Chunk size: 0 octets
Chunk boundary
Media Type
Media Type: text (26 bytes)

Good news Garry.
Anyway, as I'm using the GET procedure to provide the survey answers to the server, there is still a theoretical limit of 2048 characters for the URL.
That being said, we have then about 1000 chars available for the whole survey.
And to (almost) quote a (surely not) genius man : "1024 chars should be enough for everyone!"

@Ronounours

I was stressing the envelope. In real life, I doubt if anyone would approach my verbosity -- though slv grosso is in the same order of magnitude with this attempt (#9) on the survey design thread. I think the median response will be hardly more than the numeric parameters: hardly beyond 500 bytes. There are far more interesting (and pressing) things to work on now.

Take care,

Garry

OK, so I've done again a lot of changes and cleaning in the G'MIC and CImg source code.
I've just posted new binaries today, so if you want to help, it would be cool if you download them,
http://gmic.eu/
Feedback will be appreciated ! Thanks

Just downloaded it David. So far so good. I'll keep you posted.

Hello everyone,

Just installed the newest beta version
I have tested many filters (with the gmic_gimp plugin 64.exe on Windows 7): So far so good

I have also installed the CLI version (win64) always on Windows 7 (64 bit).
Quite often I get some runtime error by typing some command. This occurs both by working with the CMD Windows console and the MSYS shell.

Is it "normal" to get these kind of errors (they look harmless in the end...) ?

For instance:
EDIT:
gmic -h sharpen #works fine (the man page is displayed for this command)
gmic -h sharpen > log.txt # works fine (the man page is copied into the log.txt file)
gmic -h blur #does not work fine (in the end there is always a run-time error)
gmic -h blur> log.txt #does not work fine either (in the end there is always a run-time error)
gmic -h > log.txt #does not work fine (the man page is not copied into the text file > run-time error)

Here is a video with all my steps (Windows 7 - 64 bit):
https://dl.dropboxusercontent.com/u/309 ... _ERROR.mp4

P.s: these same Runtime errors occured with the previous versions of G'mic CLI as well

Ha Thanks Silvio, this is really interesting.
I'll check that.

Thanks Silvio, it should work now, after an update :

Hello David,

Thanks a lot indeed: just updated and now the CLI works fine (e.g. gmic -h blur etc)

not an error, simply a comment (and not specifically to this version I presume)
I was trying to experiment
Testing/Tom Keil/Dof manipulation
it seems interesting, but it's EXTREMELY SLOW.
Can someone (Tom if around) have a look to see whether performance can be improved?
thanks

this is more a request to David
because you are "cleaning and enhancing" the whole G'MIC treasury, may I ask whether you have in plan to enhance the "Faves" function?
many improvements can be done, to ease its use to the users:
- having a tree structure (i.e. subfolders, or at least more than just one Faves folder)
- having the possibility to rename the filter(s) stored there
- having the possibility to sort them in alpha order
...
maybe some of these are already available and simply I don't know them
thanks

Testing>Ronounours>Circlism
starts but stops after a while w/o errors and/or outcome

Testing>Ronounours>Shaded segmentation
creates an empty layer

another (hopefully small) request to David:
is it possible to have a "clean start" of the Gimp-GMIC interface from inside Gimp without restarting Gimp?
when a GMIC filter abends , if I re-enter again the G'MIC interface w/o re-opening Gimp, the interface starts with the abending filter, thus I'm obliged to restart Gimp to "clean" G'MIC memory