- Inkscape Devel - Inkscape Mailman

Re: [Inkscape-devel] Waf build system, second try
by Johan Engelen 31 Jan '14

31 Jan '14

Hi Isobel, I'll be on vacation for a week, can't give a detailed answer now. But I am not the right person to ask anyway :) I think _building_ is the easy part, but getting all the include flags correct for all platforms is not. (which is "configuration" that you mean?) An example of what would be nice to have is the ability to specify additional compiler flags for each folder. Current architecture's biggest problem is that it is not crossplatform. Note that on Windows, cygwin is cumbersome and not something we want to do. ciao, Johan ----- Reply message ----- From: "Isobel Knowles" <ik5@...2950...> To: "Johan Engelen" <jbc.engelen@...2592...> Cc: <inkscape-devel(a)lists.sourceforge.net> Subject: [Inkscape-devel] Waf build system, second try Date: Fri, Jan 31, 2014 00:14 On Thu, 30 Jan 2014 20:11:14 +0100 Johan Engelen <jbc.engelen@...2592...> wrote: > > On 30-1-2014 20:02, Isobel Knowles wrote: > > PS. My offer to write some 'nice' makefiles is still open. > > No-one is stopping you :-) > Make us dazzle... > > :) > Johan Dear Johan, OK. Before I do, do you think you could tell me what it is you want from the build? What are the essentials and what are the nice-to-haves? Where do you see the problems in the current architecture and what are its strengths? I'm only after some pointers so I don't set off in the wrong direction. I'll start the ball rolling with these: - Documentation. People who are new to Inkscape should find it easy to edit the makefiles. - Separate configuration and build cleanly so when the configuration phase doesn't do what you want for one reason or another, it is straight-forward to manually configure the build options in one or two files. - Separate compilation and installation/packaging. Installation is generally performed as root, and all too often in GNU software libtool and the like get their teeth into a 'make install', turning it into compilation job. This is wrong. - An ordinary build job should do the minimum work in the minimum time. For this I propose a flat makefile structure, perhaps composed of included makefiles. Parallel makes like this perform correctly and have no scalability issues whereas recursive makefiles are beset with problems and complication. - It should be simple to modify build flags and make alternative targets for debugging purposes. I wouldn't describe it as dazzling, but it's a start. Yours, Is.

1 0

Copying text with pattern stroke and fill + text decoration issue
by mathog 30 Jan '14

30 Jan '14

Hi all, After adding pattern fill and stroke support for text decoration I found that if both text decoration and a pattern are active, then trying to copy a text object with that combination crashes inkscape. That is, in rough terms, if the input file contains: <text > <tspan style=" pattern fill " > A <tspan style="text-decoration:underline">B</tspan></tspan></text> it loads and renders properly. But if "AB" is selected as an object (not as the text) and "^c" to copy, it crashes in style.cpp in sp_style_paint_server_ref_modified because "server" is not the same as style->getStrokePaintServer() (or style->getTextDecorationStrokePaintServer(), but these last two are identical in the test cases.) That is (sorry about the wrap): (gdb) print *server $1 = {<SPObject> = {_vptr.SPObject = 0x8a13e48, cloned = 0, uflags = 0, mflags = 0, xml_space = {set = 0, value = 0}, hrefcount = 1, _total_hrefcount = 1, document = 0x9232a80, parent = 0xa082ad8, children = 0xa086210, _last_child = 0xa086210, next = 0xa0c2ea8, id = 0xbadc000 "Packedcircles", repr = 0xbbb26f0, refCount = 4, style = 0xa096200, _release_signal = {<sigc::signal1<void, SPObject*, sigc::nil>> = {<sigc::signal_base> = {<sigc::trackable> = {callback_list_ = 0x0}, impl_ = 0xa0f6d48}, <No data fields>}, <No data fields>}, _delete_signal = {<sigc::signal1<void, SPObject*, sigc::nil>> = {<sigc::signal_base> = {<sigc::trackable> = {callback_list_ = 0x0}, impl_ = 0x0}, <No data fields>}, <No data fields>}, _position_changed_signal = {<sigc::signal1<void, SPObject*, sigc::nil>> = {<sigc::signal_base> = {<sigc::trackable> = {callback_list_ = 0x0}, impl_ = 0x0}, <No data fields>}, <No data fields>}, _modified_signal = {<sigc::signal2<void, SPObject*, unsigned int, sigc::nil>> = {<sigc::signal_base> = {<sigc::trackable> = { callback_list_ = 0x0}, impl_ = 0xa0765e8}, <No data fields>}, <No data fields>}, _successor = 0x0, _collection_policy = SPObject::ALWAYS_COLLECT, _label = 0x0, _default_label = 0x0}, swatch = false} (gdb) print *style->getTextDecorationStrokePaintServer() $2 = {<SPObject> = {_vptr.SPObject = 0x8a13e48, cloned = 0, uflags = 0, mflags = 0, xml_space = {set = 0, value = 0}, hrefcount = 7, _total_hrefcount = 7, document = 0x9232a80, parent = 0xa082ad8, children = 0x0, _last_child = 0x0, next = 0xa099530, id = 0xa059ab0 "pattern5506", repr = 0xbbb5ff0, refCount = 1, style = 0xa07be48, _release_signal = {<sigc::signal1<void, SPObject*, sigc::nil>> = {<sigc::signal_base> = {<sigc::trackable> = {callback_list_ = 0x0}, impl_ = 0xa0ccfc0}, <No data fields>}, <No data fields>}, _delete_signal = {<sigc::signal1<void, SPObject*, sigc::nil>> = {<sigc::signal_base> = {<sigc::trackable> = {callback_list_ = 0x0}, impl_ = 0x0}, <No data fields>}, <No data fields>}, _position_changed_signal = {<sigc::signal1<void, SPObject*, sigc::nil>> = {<sigc::signal_base> = {<sigc::trackable> = {callback_list_ = 0x0}, impl_ = 0x0}, <No data fields>}, <No data fields>}, _modified_signal = {<sigc::signal2<void, SPObject*, unsigned int, sigc::nil>> = {<sigc::signal_base> = {<sigc::trackable> = { callback_list_ = 0x0}, impl_ = 0xa0e8fd0}, <No data fields>}, <No data fields>}, _successor = 0x0, _collection_policy = SPObject::ALWAYS_COLLECT, _label = 0x0, _default_label = 0x0}, swatch = false} It seems like the objects are put together differently when they are copied than when they are just read in. The clue comes from the "pattern5506" above. That ID isn't in the original file, it is somehow created during the copy operation. Can somebody tell me where that is happening? The object construction when the file is opened or reverted is in style.cpp, if the copy feeds through there too, I guess the question becomes "where is the SVG created" that is fed into it. If the text-decoration just uses solid colors and fills there are no issues copying a complex formatted text object with text decorations. It is only when patterns and text decorations are both present that things go south. Thanks, David Mathog mathog@...1176... Manager, Sequence Analysis Facility, Biology Division, Caltech

1 0

Trunk is broken enough to merit complaining
by Josh Andler 29 Jan '14

29 Jan '14

Hey All, Thankfully it's been a while since this has been an issue, but it seems that a reminder is needed. The only real rule we tend to have for committing to trunk is "don't break trunk". Unfortunately, trunk is currently pretty broken when trying to work with imported bitmaps. If whomever broke it can get back to fixing it, it would be greatly appreciated. In the future, just remember, if you won't have the free time within 1 day to fix something you might have accidentally broken... please wait to commit it until you do. If you can not figure out an appropriate fix, please revert the changes and work via patches in the tracker or in a branch until you can figure it out (so others can help test). Cheers, Josh

3 2

Re: [Inkscape-devel] Trunk is broken enough to merit complaining
by mathog 29 Jan '14

29 Jan '14

> Thankfully it's been a while since this has been an issue, but it > seems that a reminder is needed. The only real rule we tend to have > for committing to trunk is "don't break trunk". Unfortunately, trunk > is currently pretty broken when trying to work with imported bitmaps. > If whomever broke it can get back to fixing it, it would be greatly > appreciated. It's more broken than that, unfortunately, since on my Ubuntu test machine the inkscape binaries from 12978 on no longer even get to the point where they put any graphics on the screen. The reason it is crashing is complex enough that it really is above and beyond the call of duty for other developers to have to figure it out. (For instance, the failure mode is different in the debugger than outside of it.) So, Martin, please pull out all of the image changes from around 12939 through 12978 and test them for a while in a separate branch. Before resubmitting them to trunk, please have people test your branch on other platforms. Or at the very least run valgrind on it and make sure there are no (new) memory access issues in the code you are adding. (Admittedly this is a bit of a pain since there are already so many memory issues in Inkscape.) These image changes must have worked on Martin's machine, presumably because the memory problems didn't happen to blow up anything there, but they are show stoppers for me. Thanks, David Mathog mathog@...1176... Manager, Sequence Analysis Facility, Biology Division, Caltech

1 0

Writing Tests for URI
by Martin Owens 29 Jan '14

29 Jan '14

Hi devs, I want to write a test suite for uri.h/uri.cpp; I had a look to see if there are existing tests for the code and I admit I find myself a little confused. There's a directory called cxxtest which contains tests and then there are *-test.h files in the actual codebase. My feeling is to use extract-uri-test.h, copy it for a new uri-test.h file. But this might be wrong. New URI functionality includes: dealing with data uri, getting full path name (previous work) and further intergration between it and io/uristream.cpp to feed that lovely data in transparently when needed. Should conclude with the ability to quickly move any href from attribute to workable data. Martin,

2 1

Bug Hunt for 0.91
by Bryce Harrington 29 Jan '14

29 Jan '14

Welcome to the first bug hunt for the 0.91 release. This hunt focuses on regression bugs. Traditionally, we specify a point target for us to reach, and award us points based on the severity of the bug: 3, 6, 9, or 12 points for low, medium, high, and critical. This bug hunt's goal is 500 points worth of bugs from the following list: http://tinyurl.com/bughunt91a You'll note a few dozen regressions were already fixed, so we'll count that as a (207 point) headstart on the effort. Here's a couple more links to list only the open or closed bugs: http://tinyurl.com/bughunt91a-open http://tinyurl.com/bughunt91a-closed [I'll write a launchpadlib script to calculate the score from the latter list, unless someone's ambitious and beats me to it!] We only give points for true bug fixes - not bugs simply closed due to being dupes, invalid, works-for-me, or whatnot. Once the bug is fixed in the mainline branch, set its status to Fix Committed, and it'll tally as a fixed regression. Bug Hunt Goal: 500 points Current: 207 points If you're not a developer, you can still help by seeing if you can reproduce the bug yourself, and posting your findings. If you have an idea on how to fix the bug, but don't have the time to do it yourself, then outlining your thoughts on the bug may help junior developers pick it up and implement the changes. If you're an experienced developer, consider posting an invitation to an IRC bug party for the time period you'll be working, if you're willing to help pilot junior developers and sponsor their patches. I will be posting scoring updates a couple times a week until we're done. Bryce

4 9

Cairo and ISO C++
by Tavmjong Bah 28 Jan '14

28 Jan '14

Hi, There is a proposal before the ISO C++ standards group for a technical specification for 2D graphics based on the Cairo API.[1] It seems to have gained some traction. My question here is what is missing in the Cairo API that would be useful for rendering SVG? Tav [1] http://isocpp.org/files/papers/N3888.pdf

3 3

Strange Error
by Arshdeep Singh 26 Jan '14

26 Jan '14

Hi, I participated is GSoC 2013 and wrote code for Inkscape. I was working in WIndows last year and sometime around December I decided to port my branch to Linux (Ubuntu 12.10 , in my case) . Strangely, I am caught up with this error, which doesn't trouble me at all in my Windows branch: "ui/dialog/dialog-manager.o: In function `Inkscape::UI::Dialog::RecolorArtwork::getInstance()': /home/arshdeep/Desktop/InkS-Repo/recolor/src/./ui/dialog/recolor-artwork.h:38: undefined reference to `Inkscape::UI::Dialog::RecolorArtwork::RecolorArtwork()' collect2: error: ld returned 1 exit status " This happens only when I am associating a dialog box of the tool that I made last year thorugh the codeline: " registerFactory("RecolorArtwork", &create<RecolorArtwork, FloatingBehavior>); " inside of dialog-manager.cpp. The RecolorArtwork class is mirrored directly from FillAndStroke class and I can't pin point why this error crops up ? If someone can help , please do so. -- Arshdeep Singh Third Year, Computer Engineering Delhi Technological University Ph: +91-9654115614 https://sites.google.com/site/adsingh1729/home

1 1

Simple String Handling
by Martin Owens 25 Jan '14

25 Jan '14

Hi devs, I've tried out the following patch: http://paste.ubuntu.com/6809633/ And it's not working. The path which used to work well, it now being overwritten with garbage data before the file can be opened. As if the reference has been freed. I'm sure it's something simple to do with those consts or gchar types. Any ideas? Example output: ** (inkscape:4946): WARNING **: Can't open file: /home/doctormo/D\x89 \x99 (doesn't exist) ** (inkscape:4946): WARNING **: Can't get document for referenced URI: inner.svg#bar Martin,

3 3

Removal of popt
by Krzysztof Kosiński 24 Jan '14

24 Jan '14

Hello The bug which is keeping us with popt has been recently fixed. https://bugzilla.gnome.org/show_bug.cgi?id=722025 This means we can remove several kb of code and remove one of our dependencies by changing command line parsing from popt to GOption. Since the bug was Windows-specific, we can do this once the next stable version of glib is released. Regards, Krzysztof

4 5