Warnings when pressing undo keys in tutorials

[R-mosaic]

Warnings when pressing undo keys in the Shapes tutorial

Inkscape 0.45.1, Mar. 21, 2007

Steps to reproduce:
1. Run Inkscape.
2. Open the Shapes tutorial (exactly this one).
3. Press Ctrl+Z.

Result:
Inkscape displays warnings about incomplete undo history on commandline

Expected result:
no warnings

[R-mosaic (r-mosaic)]

Originator: YES

gdb bt output:

Program received signal SIGSEGV, Segmentation fault.
[Switching to thread 3276.0x8a8]
0x004f5868 in ?? ()
(gdb) bt
#0 0x004f5868 in ?? ()
#1 0x0045eb2f in ?? ()
#2 0x004d133a in ?? ()
#3 0x005c91a0 in ?? ()
#4 0x005cd6cb in ?? ()
#5 0x005c89f9 in ?? ()
#6 0x605783d2 in _libmsvcrt_a_iname ()
#7 0x62743935 in _libmsvcrt_a_iname ()
#8 0x62757325 in _libmsvcrt_a_iname ()
#9 0x62757a3c in _libmsvcrt_a_iname ()
#10 0x62757f56 in _libmsvcrt_a_iname ()
#11 0x6069b8a4 in _libmsvcrt_a_iname ()
#12 0x605756b0 in _libmsvcrt_a_iname ()
#13 0x6057686d in _libmsvcrt_a_iname ()
#14 0x6b070e9e in _libmsvcrt_a_iname ()
#15 0x672dda27 in _libmsvcrt_a_iname ()
#16 0x672deefb in _libmsvcrt_a_iname ()
#17 0x672df0da in _libmsvcrt_a_iname ()
#18 0x60575dde in _libmsvcrt_a_iname ()
#19 0x014d39e3 in ?? ()
#20 0x00402ae7 in ?? ()
#21 0x00408503 in ?? ()
#22 0x00402242 in ?? ()
#23 0x00401237 in ?? ()
#24 0x004012a8 in ?? ()
#25 0x7c82f23b in _libmsvcrt_a_iname ()

[Buliabyak-users (buliabyak-users)]

Originator: NO

no crash in svn build

[R-mosaic (r-mosaic)]

Originator: YES

I downloaded Inkscape0711021622.7z, and tried it. It still crashes. The
platform is Windows XP.

[Johan Engelen (johanengelen)]

Just tested (svn 17700), it no longer crashes, but it does gives warnings about incomplete undo history. Removing crash tag.

[Johan Engelen (johanengelen)]

I have removed the win32 tag, are we definitely sure this only happens on windows? (seems strange for this bug)

[ScislaC (scislac)]

Johan,

Good call... not win32 only. In opening via command line on Ubuntu Hardy I get the following output (also, no crash here).

** (inkscape:18278): WARNING **: Incomplete undo transaction:

** (inkscape:18278): WARNING **: Event: Set attribute inkscape:original to "M 96.125 5217.4375 C 90.627188 5218.3264 88.703454 5238.2845 82.65625 5239.9688 C 75.920688 5241.8446 62.131885 5223.9675 57.03125 5228.75 C 52.146599 5233.3298 70.825719 5246.1407 69.15625 5252.625 C 67.412966 5259.3961 45.033761 5262.3791 46.625 5269.1875 C 48.148858 5275.7076 68.594118 5265.9372 73.375 5270.625 C 78.367278 5275.5203 69.74563 5296.4115 76.4375 5298.4375 C 82.846009 5300.3779 84.612149 5277.7965 91.0625 5276 C 97.798062 5274.1242 111.58687 5292.0012 116.6875 5287.2188 C 121.57215 5282.639 102.89303 5269.8281 104.5625 5263.3438 C 106.30578 5256.5727 128.68499 5253.5897 127.09375 5246.7812 C 125.56989 5240.2611 105.12463 5250.0315 100.34375 5245.3438 C 95.351472 5240.4485 103.97312 5219.5572 97.28125 5217.5312 C 96.880718 5217.41 96.491521 5217.3782 96.125 5217.4375 z " on #<Element:0x0xac3d9f0>

** (inkscape:18278): WARNING **: Event: Set attribute inkscape:original to "M 96.125 5217.4375 C 90.627188 5218.3264 88.703454 5238.2845 82.65625 5239.9688 C 75.920688 5241.8446 62.131885 5223.9675 57.03125 5228.75 C 52.146599 5233.3298 70.825719 5246.1407 69.15625 5252.625 C 67.412966 5259.3961 45.033761 5262.3791 46.625 5269.1875 C 48.148858 5275.7076 68.594118 5265.9372 73.375 5270.625 C 78.367278 5275.5203 69.74563 5296.4115 76.4375 5298.4375 C 82.846009 5300.3779 84.612149 5277.7965 91.0625 5276 C 97.798062 5274.1242 111.58687 5292.0012 116.6875 5287.2188 C 121.57215 5282.639 102.89303 5269.8281 104.5625 5263.3438 C 106.30578 5256.5727 128.68499 5253.5897 127.09375 5246.7812 C 125.56989 5240.2611 105.12463 5250.0315 100.34375 5245.3438 C 95.351472 5240.4485 103.97312 5219.5572 97.28125 5217.5312 C 96.880718 5217.41 96.491521 5217.3782 96.125 5217.4375 z " on #<Element:0x0xac3d970>

** (inkscape:18278): CRITICAL **: SPObject* sp_object_unref(SPObject*, SPObject*): assertion `object != NULL' failed

[ScislaC (scislac)]

given that it's just a warning, lowering priority

[prkos (prkos)]

I can't reproduce on WinXP from March 3rd, and fedora 7 from Jan 10th.

I don't get any warnings; the first time I hit Ctrl+z the redo icon lights up and after the second hit I get "Nothing to redo" in the statusbar.

[ScislaC (scislac)]

prkos,

Start inkscape from a command prompt, you will see the warnings output there.

[su_v (suv-lp)]

confirmed with Inkscape 0.46+devel r22271, OS X 10.5.8:
happens with every tutorial linked in the Help menu (English version)

[jazzynico (jazzynico)]

I confirm it happens with the other tutorials too (tested on Ubuntu 9.04, Inkscape rev. 22293).

[jazzynico (jazzynico)]

If I understand correctly, the d attribute (in path elements) coding style used in the tutorials is a bit old and Inkscape changes it at load time, but can't undo this operation.
I plan to fix the tutorial, even if it doesn't fix the compatibility issue with other old files. At least we'll get rid of the bug with the official tutorials.
Does someone remember (or just know) when Inkscape change the way it handles the paths?

[jazzynico (jazzynico)]

Attached is a reduced test file with three different elements showing the same (translated) path.

The first one is extracted from the basic tutorial in Inkscape 0.46 (d="M 606.48149 567.70831 A 179.39815 53.819443 0 1 1 247.68520,567.70831 A 179.39815 53.819443 0 1 1 606.48149 567.70831 z").

The second one shows how the d attribute is converted after saving to Inkscape SVG (d="m 606.48149,567.70831 c 0,29.72366 -80.31929,53.81945 -179.39815,53.81945 -99.07886,0 -179.39814,-24.09579 -179.39814,-53.81945 0,-29.72365 80.31928,-53.81944 179.39814,-53.81944 99.07886,0 179.39815,24.09579 179.39815,53.81944 z").

The last one shows how the d attribute is converted after saving to Optimized SVG, as tutorials are as of Inkscape 0.48 (d="m606.48149,567.70831c0,29.72366-80.31929,53.81945-179.39815,53.81945s-179.39814-24.09579-179.39814-53.81945c0-29.72365,80.31928-53.81944,179.39814-53.81944s179.39815,24.09579,179.39815,53.81944z").

When undoing, warnings appear for the first and last paths, but not for the one saved as Inkscape SVG.

[jazzynico (jazzynico)]

Planned workaround for the tutorials: replacement of arc paths (with sodipodi:type="arc") with real SVG ellipses.

[jazzynico (jazzynico)]

Unfortunately, Inkscape automatically adds a "d" attribute to the ellipse elements, and thus we're back to the initial bug.
But since the only goal of the problematic paths (in the tutorials) is to generate a blurred drop shadow under the Inkscape logo in the header and footer, replacing them with a filter would be a more effective workaround (positive side effect, the size of each tutorial is reduces by about 8 kio after optimization –1 Mio overall).

[su_v (suv-lp)]

<off-topic>
OTOH do we really need to use Scour for the tutorials? As far as I understand, it is intended to optimize SVG files e.g. for publishing on the web, not for the use in Inkscape itself… The tutorials on the other hand are not general SVG tutorials but intended specifically to be viewed in Inkscape.
(Yes, I do remember that it was about the file size… and overall size of the distributed packages)
</off-topic>

[su_v (suv-lp)]

From the scoured path attribute 'd':
> s179.39815,24.09579,179.39815,53.81944

Just curious - does Inkscape ever write 's' shorthand/smooth curveto path commands or always convert it (internally, even for rendering), to a path 'd' string limited to the same path commands as used when saving as Inskcape SVG (apparently m, c, z and the corresponding absolute commands for curved path segments)?

[jazzynico (jazzynico)]

@~suv.Optimizing the tutorials is not critical, but since scour provides a very easy and efficient way to do it, I think it's worth using it, even if it generate warnings. Size matters here, but providing the cleanest files we can is important too.
Ok, I must admit I'm a bit fussy sometimes...

As for the smooth curveto, I've also noticed that, but have no idea how it works internally.

[jazzynico (jazzynico)]

Workaround partially applied in inkscape-docs, revision 426.
Partially only because in some cases the tutorials (shapes and elements) describe the shape as an arc and must thus the arc specific features must be available from the tutorial.
BTW, the bug also affect stars.

I'm going to open a separate report for the generic part of this issue. The tutorials can be considered fixed for most of them, and won't fix for the others (until the generic bug is fixed).

[jazzynico (jazzynico)]

Updated tutorials in the trunk as of revision 10007.

[OpenContrail Admin (ci-admin-f)]

Review in progress for https://review.opencontrail.org/31426
Submitter: Nitish Krishna Kaveri (<email address hidden>)

[OpenContrail Admin (ci-admin-f)]

Review in progress for https://review.opencontrail.org/31427
Submitter: Nitish Krishna Kaveri (<email address hidden>)

[OpenContrail Admin (ci-admin-f)]

Reviewed: https://review.opencontrail.org/31427
Committed: http://github.com/Juniper/contrail-server-manager/commit/ae749d34c937f417057f128c0e0874b6dc1c3a36
Submitter: Zuul (<email address hidden>)
Branch: R4.0

commit ae749d34c937f417057f128c0e0874b6dc1c3a36
Author: nitishkrishna <email address hidden>
Date: Thu May 11 11:43:02 2017 -0700

Closes-Bug: #169020 - Trailing newline char in FQDN causes cmd to split

The newline char at the end of FQDN string caused new '-config' to be sent to next line
The caused error in cmd execution and prevented SM from coming up

Change-Id: I88ef395ea23a790dacaefeb57f7e2ac182e5cc85

[OpenContrail Admin (ci-admin-f)]

Reviewed: https://review.opencontrail.org/31426
Committed: http://github.com/Juniper/contrail-server-manager/commit/faa6ad90e106d4051024776b754e39776e3af859
Submitter: Zuul (<email address hidden>)
Branch: master

commit faa6ad90e106d4051024776b754e39776e3af859
Author: nitishkrishna <email address hidden>
Date: Thu May 11 11:43:02 2017 -0700

Closes-Bug: #169020 - Trailing newline char in FQDN causes cmd to split

The newline char at the end of FQDN string caused new '-config' to be sent to next line
The caused error in cmd execution and prevented SM from coming up

Change-Id: I88ef395ea23a790dacaefeb57f7e2ac182e5cc85