Inkscape build test system + current results
A couple weeks ago, Kees and I set up an automated test harness for Inkscape. It's been running smoothly and collecting test results.
Currently, the harness simply retrieves the nightly snapshots and builds them, logging compilation or installation problems.
Anyway, I notice it's found a few warnings:
### BUILD ### Making all in src
display/canvas-arena.cpp: In function `void sp_canvas_arena_class_init(SPCanvasArenaClass*)': display/canvas-arena.cpp:91: warning: invalid access to non-static data member `SPCanvasArenaClass::arena_event' of NULL object display/canvas-arena.cpp:91: warning: (perhaps the `offsetof' macro was used incorrectly)
display/sp-canvas.cpp: In function `void sp_canvas_item_class_init(SPCanvasItemClass*)': display/sp-canvas.cpp:112: warning: invalid access to non-static data member `SPCanvasItemClass::event' of NULL object display/sp-canvas.cpp:112: warning: (perhaps the `offsetof' macro was used incorrectly)
libnrtype/Layout-TNG-Output.cpp: In function `const char* Inkscape::Text::weight_to_text(PangoWeight)': libnrtype/Layout-TNG-Output.cpp:226: warning: enumeration value `PANGO_WEIGHT_SEMIBOLD' not handled in switch
livarot/float-line.cpp: In member function `void FloatLigne::Affiche()': livarot/float-line.cpp:518: warning: long unsigned int format, size_t arg (arg 2) livarot/float-line.cpp:524: warning: long unsigned int format, size_t arg (arg 2)
Making all in po am.po: parsing error for `<defs>': #: ../src/main.cpp:326 msgid "Remove unused definitions from the <defs> section(s) of the document" msgstr "" (repeated for each language)
WARNING: Bad markup found in translations. Please consider fixing the above problems. BUILD STATUS: 0
======================================================================== pkgpath=/testing/packages/inkscape/inkscape-20050703-1400.tar.bz2 id=86 ========================================================================
Bryce
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
...
Currently, the harness simply retrieves the nightly snapshots and builds them, logging compilation or installation problems.
Anyway, I notice it's found a few warnings: ... libnrtype/Layout-TNG-Output.cpp: In function `const char* Inkscape::Text::weight_to_text(PangoWeight)': ... livarot/float-line.cpp:518: warning: long unsigned int format, size_t arg (arg 2)
I have patches for these, do you want to see them?
Not sure that I have a complete grasp of this web thingy
---------- Forwarded message ---------- From: Ben Fowler <ben.the.mole@...400...> Date: Jul 4, 2005 9:26 AM Subject: Re: [Inkscape-devel] Inkscape build test system + current results To: inkscape-devel@...6...
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
...
Currently, the harness simply retrieves the nightly snapshots and builds them, logging compilation or installation problems.
Anyway, I notice it's found a few warnings: ... libnrtype/Layout-TNG-Output.cpp: In function `const char* Inkscape::Text::weight_to_text(PangoWeight)': ... livarot/float-line.cpp:518: warning: long unsigned int format, size_t arg (arg 2)
I have patches for these, do you want to see them?
Not sure that I have a complere grasp of this web thingy
---------- Forwarded message ---------- From: Ben Fowler <ben.the.mole@...400...> Date: Jul 4, 2005 9:26 AM Subject: Re: [Inkscape-devel] Inkscape build test system + current results To: inkscape-devel@...6...
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
...
Currently, the harness simply retrieves the nightly snapshots and builds them, logging compilation or installation problems.
Anyway, I notice it's found a few warnings: ... libnrtype/Layout-TNG-Output.cpp: In function `const char* Inkscape::Text::weight_to_text(PangoWeight)': ... livarot/float-line.cpp:518: warning: long unsigned int format, size_t arg (arg 2)
I have patches for these, do you want to see them?
On Mon, Jul 04, 2005 at 09:26:02AM +0100, Ben Fowler wrote:
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
...
Currently, the harness simply retrieves the nightly snapshots and builds them, logging compilation or installation problems.
Anyway, I notice it's found a few warnings: ... libnrtype/Layout-TNG-Output.cpp: In function `const char* Inkscape::Text::weight_to_text(PangoWeight)': ... livarot/float-line.cpp:518: warning: long unsigned int format, size_t arg (arg 2)
I have patches for these, do you want to see them?
Yes, please do!
Bryce
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
On Mon, Jul 04, 2005 at 09:26:02AM +0100, Ben Fowler wrote:
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
... livarot/float-line.cpp:518: warning: ...
I have patches for these, do you want to see them?
Yes, please do!
This could be a case of needing the modern format specifier for printf, http://unixhelp.ed.ac.uk/CGI/man-cgi?printf+3 videlicet:
Index: livarot/float-line.cpp =================================================================== RCS file: /cvsroot/inkscape/inkscape/src/livarot/float-line.cpp,v retrieving revision 1.8 diff -r1.8 float-line.cpp 518c518 < printf("%lu : \n", bords.size()); ---
printf("%zu : \n", bords.size());
524c524 < printf("%lu : \n", runs.size()); ---
printf("%zu : \n", runs.size());
If you O.K. this, I'll check it it in.
Ben
I am having difficulty seeing my posts ...
---------- Forwarded message ---------- From: Ben Fowler <ben.the.mole@...400...> Date: Jul 5, 2005 8:10 AM Subject: long unsigned int format, size_t arg (was Re: Inkscape build test system + current results) To: inkscape-devel@...6...
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
On Mon, Jul 04, 2005 at 09:26:02AM +0100, Ben Fowler wrote:
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
... livarot/float-line.cpp:518: warning: ...
I have patches for these, do you want to see them?
Yes, please do!
This could be a case of needing the modern format specifier for printf, http://unixhelp.ed.ac.uk/CGI/man-cgi?printf+3 videlicet:
Index: livarot/float-line.cpp =================================================================== RCS file: /cvsroot/inkscape/inkscape/src/livarot/float-line.cpp,v retrieving revision 1.8 diff -r1.8 float-line.cpp 518c518 < printf("%lu : \n", bords.size()); ---
printf("%zu : \n", bords.size());
524c524 < printf("%lu : \n", runs.size()); ---
printf("%zu : \n", runs.size());
If you O.K. this, I'll check it it in.
Ben
On Tue, Jul 05, 2005 at 08:10:35AM +0100, Ben Fowler wrote:
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
On Mon, Jul 04, 2005 at 09:26:02AM +0100, Ben Fowler wrote:
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
... livarot/float-line.cpp:518: warning: ...
I have patches for these, do you want to see them?
Yes, please do!
This could be a case of needing the modern format specifier for printf, http://unixhelp.ed.ac.uk/CGI/man-cgi?printf+3 videlicet:
Index: livarot/float-line.cpp
RCS file: /cvsroot/inkscape/inkscape/src/livarot/float-line.cpp,v retrieving revision 1.8 diff -r1.8 float-line.cpp 518c518
< printf("%lu : \n", bords.size());
printf("%zu : \n", bords.size());524c524
< printf("%lu : \n", runs.size());
printf("%zu : \n", runs.size());If you O.K. this, I'll check it it in.
Looks good, please commit it if you haven't already. (Sorry I didn't reply sooner, this msg got lost in my inbox)
Bryce
Index: livarot/float-line.cpp
RCS file: /cvsroot/inkscape/inkscape/src/livarot/float-line.cpp,v retrieving revision 1.8 diff -r1.8 float-line.cpp 518c518
< printf("%lu : \n", bords.size());
printf("%zu : \n", bords.size());524c524
< printf("%lu : \n", runs.size());
printf("%zu : \n", runs.size());If you O.K. this, I'll check it it in.
Looks good, please commit it if you haven't already.
Yeah, and please revert back what I sent for it. Yours (Ben's) is the correct way.
ralf
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
On Mon, Jul 04, 2005 at 09:26:02AM +0100, Ben Fowler wrote:
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
... libnrtype/Layout-TNG-Output.cpp: In function ...
I have patches for these, do you want to see them?
Yes, please do!
The enumeration value `PANGO_WEIGHT_SEMIBOLD' is guarded by a highly questionable #ifdef, videlicet:
00216 switch (w) { 00217 case PANGO_WEIGHT_ULTRALIGHT: return "ultralight"; 00218 case PANGO_WEIGHT_LIGHT : return "light"; 00219 #ifdef PANGO_WEIGHT_SEMIBOLD // not available on pango before 1.8 00220 case PANGO_WEIGHT_SEMIBOLD : return "semibold"; 00221 #endif 00222 case PANGO_WEIGHT_NORMAL : return "normalweight"; 00223 case PANGO_WEIGHT_BOLD : return "bold"; 00224 case PANGO_WEIGHT_ULTRABOLD : return "ultrabold"; 00225 case PANGO_WEIGHT_HEAVY : return "heavy"; 00226 }
On line 219 it would be irrational for PANGO_WEIGHT_SEMIBOLD to be defined as a pre-processor symbol, it is needed as an enumeration value.
Also, I haven't seen PANGO_WEIGHT_SEMIBOLD not being available, so my first suggestion would be to remove the conditional compilation, and if this causes problems, choose a non-conflicting symbol for the macro or use the pango version directly.
See http://inkscape.gristle.org/2005-04-17.txt http://inkscape.gristle.org/2005-03-14.txt
Ben
On Tue, Jul 05, 2005 at 08:46:43AM +0100, Ben Fowler wrote:
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
On Mon, Jul 04, 2005 at 09:26:02AM +0100, Ben Fowler wrote:
On 7/4/05, Bryce Harrington <bryce@...260...> wrote:
... libnrtype/Layout-TNG-Output.cpp: In function ...
I have patches for these, do you want to see them?
Yes, please do!
The enumeration value `PANGO_WEIGHT_SEMIBOLD' is guarded by a highly questionable #ifdef, videlicet:
00216 switch (w) { 00217 case PANGO_WEIGHT_ULTRALIGHT: return "ultralight"; 00218 case PANGO_WEIGHT_LIGHT : return "light"; 00219 #ifdef PANGO_WEIGHT_SEMIBOLD // not available on pango before 1.8 00220 case PANGO_WEIGHT_SEMIBOLD : return "semibold"; 00221 #endif 00222 case PANGO_WEIGHT_NORMAL : return "normalweight"; 00223 case PANGO_WEIGHT_BOLD : return "bold"; 00224 case PANGO_WEIGHT_ULTRABOLD : return "ultrabold"; 00225 case PANGO_WEIGHT_HEAVY : return "heavy"; 00226 }
On line 219 it would be irrational for PANGO_WEIGHT_SEMIBOLD to be defined as a pre-processor symbol, it is needed as an enumeration value.
Looks like you're right. This shows it to be an enum:
http://developer.gimp.org/api/2.0/pango/pango-Fonts.html
Also, I haven't seen PANGO_WEIGHT_SEMIBOLD not being available, so my first suggestion would be to remove the conditional compilation, and if this causes problems, choose a non-conflicting symbol for the macro or use the pango version directly.
Agreed, can you remove the conditional, assuming no one disagrees?
Bryce
On Sun, Jul 03, 2005 at 11:40:34PM -0700, Bryce Harrington wrote:
Making all in po am.po: parsing error for `<defs>': #: ../src/main.cpp:326 msgid "Remove unused definitions from the <defs> section(s) of the document" msgstr "" (repeated for each language)
I've fixed this as of last night. I added an exception to the "check-markup" script.
participants (4)
-
Ben Fowler -
Bryce Harrington -
Kees Cook -
Ralf Stephan