- Dillo-dev - lists.mailman3.com

Valgrind errors from downloads.cc
by corvid＠lavabit.com Feb. 10, 2011

Feb. 10, 2011

Jeremy wrote: > Valgrind is reporting uninitialised data in DLItem::update() in > downloads.cc[1]. Here's the code with the offending lines (705, 706) > marked. > > /* Rate */ > if (csec >= 2) { > tsec = (float) (curr_time - twosec_time); > ! rate = ((float)(curr_bytesize-twosec_bytesize) / 1024) / tsec; > ! snprintf(str, 64, (rate < 100) ? "%.1fK/s" : "%.0fK/s", rate); > prRate->copy_label(str); > prRate->redraw_label(); > } > > I'm wondering what is happening here. DLItem::update() is scheduled > to be run every second. I think that twosec_time is only set to an > initialised value when the function has been run at least twice > before. Also, csec is calculated from the difference between the > current time and the initial time. Is it possible that the "csec >= > 2" could be true even when the function has been run only once before? > After all, the timing of callbacks is never exact. Perhaps the > callback should explicitly count how many times it has been called > rather than attempt to work it out from the timing? (The list software never did send me a copy of your msg...) Would something like the following work? --- a/dpi/downloads.cc Tue Jan 11 01:18:28 2011 +0000 +++ b/dpi/downloads.cc Sun Feb 06 20:40:53 2011 +0000 @@ -318,6 +318,8 @@ // Init value. Reset later, upon the first data bytes arrival init_time = time(NULL); + twosec_time = onesec_time = init_time; + // BUG:? test a URL with ' inside. /* escape "'" character for the shell. Is it necessary? */ esc_url = Escape_uri_str(url, "'"); > [1] > http://starurchin.org/dillo/valgrind.html > http://starurchin.org/dillo/valgrind/6444c6ff40e01e517ce6e802f23670c5422c6e… > http://starurchin.org/dillo/valgrind/68bd8732669bbac8e11f1ea5a692db022fc6a6… > http://starurchin.org/dillo/valgrind/6f1d6a4feadf8136cbed37497a1d6dbbfea4a3… > http://starurchin.org/dillo/valgrind/68026e750cc92b08699c26392ecfd92cf370a9… >

3 5

Valgrind errors from downloads.cc
by onepoint＠starurchin.org Feb. 7, 2011

Feb. 7, 2011

Valgrind is reporting uninitialised data in DLItem::update() in downloads.cc[1]. Here's the code with the offending lines (705, 706) marked. /* Rate */ if (csec >= 2) { tsec = (float) (curr_time - twosec_time); ! rate = ((float)(curr_bytesize-twosec_bytesize) / 1024) / tsec; ! snprintf(str, 64, (rate < 100) ? "%.1fK/s" : "%.0fK/s", rate); prRate->copy_label(str); prRate->redraw_label(); } I'm wondering what is happening here. DLItem::update() is scheduled to be run every second. I think that twosec_time is only set to an initialised value when the function has been run at least twice before. Also, csec is calculated from the difference between the current time and the initial time. Is it possible that the "csec >= 2" could be true even when the function has been run only once before? After all, the timing of callbacks is never exact. Perhaps the callback should explicitly count how many times it has been called rather than attempt to work it out from the timing? Regards, Jeremy Henty [1] http://starurchin.org/dillo/valgrind.html http://starurchin.org/dillo/valgrind/6444c6ff40e01e517ce6e802f23670c5422c6e… http://starurchin.org/dillo/valgrind/68bd8732669bbac8e11f1ea5a692db022fc6a6… http://starurchin.org/dillo/valgrind/6f1d6a4feadf8136cbed37497a1d6dbbfea4a3… http://starurchin.org/dillo/valgrind/68026e750cc92b08699c26392ecfd92cf370a9…

1 0

dillo fltk-1.3 scroll bar delay
by rogerx.oss＠gmail.com Feb. 4, 2011

Feb. 4, 2011

I messed around with ltrace to debug the scrollbar delay within dillo fltk-1.3 port. Here's what I get so far. roger@localhost1 ~ $ ltrace -c -p 11442 % time seconds usecs/call calls function ------ ----------- ----------- --------- -------------------- 18.32 0.291790 6785 43 _ZN8Fl_Group4drawEv 11.69 0.186144 1861 100 _ZN7Fl_Tabs6handleEi 9.70 0.154394 8126 19 _ZN9Fl_Window5flushEv 9.41 0.149853 7887 19 _ZN9Fl_Window4drawEv 9.05 0.144073 7582 19 _ZN7Fl_Tabs4drawEv 8.98 0.142919 793 180 _ZN8Fl_Group6handleEi 6.62 0.105367 115 911 _Znwj 5.82 0.092602 101 911 _ZdlPv 4.82 0.076755 102 746 _ZN2Fl12event_insideEPK9Fl_Widget 2.34 0.037301 153 243 strlen 2.11 0.033615 107 314 malloc 2.06 0.032840 104 314 free 1.81 0.028844 1696 17 _Z9fl_scrolliiiiiiPFvPviiiiES_ 1.49 0.023788 141 168 _ZdaPv 1.11 0.017654 105 168 _Znaj 1.04 0.016517 105 157 __vsnprintf_chk 1.01 0.016079 102 157 _ZN9Fl_Input_5valueEPKc 0.89 0.014229 459 31 _ZN8Fl_Input6handleEi 0.39 0.006271 169 37 _ZNK8Fl_Group10draw_childER9Fl_Widget 0.34 0.005480 103 53 _ZNK8Fl_Group5arrayEv 0.31 0.004997 138 36 _ZNK8Fl_Group12update_childER9Fl_Widget 0.23 0.003682 108 34 _ZN9Fl_Slider11scrollvalueEiiii 0.16 0.002508 192 13 _ZN9Fl_Button4drawEv 0.12 0.001868 116 16 _ZN9Fl_Button6handleEi 0.11 0.001818 106 17 _ZN9Fl_Widget6damageEh 0.02 0.000380 190 2 _ZN6Fl_Box4drawEv 0.02 0.000361 180 2 _ZN8Fl_Input4drawEv 0.01 0.000227 113 2 _ZN9Fl_Window6resizeEiiii ------ ----------- ----------- --------- -------------------- 100.00 1.592356 4729 total roger@localhost1 ~ $ ltrace -c -p 11442 % time seconds usecs/call calls function ------ ----------- ----------- --------- -------------------- 29.26 29.729943 120364 247 _ZN8Fl_Group4drawEv 14.75 14.984334 149843 100 _ZN9Fl_Window5flushEv 14.72 14.952232 149522 100 _ZN9Fl_Window4drawEv 14.69 14.922999 149229 100 _ZN7Fl_Tabs4drawEv 13.48 13.693784 228229 60 _Z9fl_scrolliiiiiiPFvPviiiiES_ 4.09 4.157141 329 12601 strlen 2.28 2.320390 860 2695 _Znwj 2.28 2.317818 860 2695 _ZdlPv 2.15 2.186969 1436 1522 free 0.41 0.421299 141 2982 _ZNK8Fl_Group5arrayEv 0.23 0.238538 126 1884 _ZN2Fl12event_insideEPK9Fl_Widget 0.22 0.220701 147 1500 _ZNK8Fl_Group12update_childER9Fl_Widget 0.19 0.188841 2121 89 _ZN7Fl_Tabs6handleEi 0.18 0.179153 248 720 _ZN8Fl_Group6resizeEiiii 0.18 0.179127 117 1522 malloc 0.16 0.159278 556 286 _ZNK8Fl_Group10draw_childER9Fl_Widget 0.13 0.128988 503 256 _ZN8Fl_Group6handleEi 0.09 0.094990 124 761 __vsnprintf_chk 0.09 0.089326 117 761 _ZN9Fl_Input_5valueEPKc 0.08 0.082873 2762 30 _Z9fl_cursor9Fl_Cursorjj 0.08 0.078284 130 600 _ZN9Fl_Input_6resizeEiiii 0.05 0.048376 115 420 _ZN9Fl_Widget6resizeEiiii 0.04 0.042491 118 360 _Z8fl_widthPKci 0.04 0.036692 136 269 _Znaj 0.03 0.030249 112 269 _ZdaPv 0.02 0.024592 105 233 fl_utf8fwd 0.01 0.013969 194 72 _ZN8Fl_Input4drawEv 0.01 0.013501 112 120 _ZN9Fl_Slider11scrollvalueEiiii 0.01 0.012305 146 84 _ZN8Fl_Input6handleEi 0.01 0.012155 289 42 _ZNK9Fl_Widget10draw_labelEv 0.01 0.007515 121 62 fl_utf8back 0.01 0.007343 174 42 _ZNK9Fl_Widget13draw_backdropEv 0.01 0.006733 112 60 _ZN9Fl_Widget6damageEh 0.00 0.004839 115 42 _ZNK9Fl_Widget8draw_boxE10Fl_Boxtypej 0.00 0.002366 182 13 _ZN9Fl_Button4drawEv 0.00 0.001462 104 14 _ZN6Fl_Box6handleEi 0.00 0.001444 103 14 _ZN9Fl_Button6handleEi 0.00 0.000528 264 2 _ZN6Fl_Box4drawEv 0.00 0.000226 113 2 _ZN9Fl_Window6resizeEiiii ------ ----------- ----------- --------- -------------------- 100.00 101.593794 33631 total

3 5

Yong Li added you as a business connection on Plaxo
by plaxo＠mx.plaxo.com Feb. 2, 2011

Feb. 2, 2011

Yong Li wants to add you as a business connection on Plaxo. To accept this connection request, go to: http://www.plaxo.com/invite?lang=en&i=120351393&k=1829487072&l=en&src=email… Thanks, The Plaxo team Plaxo - your address book for life. Don't want to receive emails from Plaxo any more? Go to: http://www.plaxo.com/stop?src=email&et=1&est=business&etv=jun10d&el=en&emai… Plaxo, Inc. - 203 Ravendale Drive - Mountain View - CA - 94043 - USA

2 1

Final release of FLTK 2.0 Dillo
by Hole.destructor＠gmx.de Jan. 30, 2011

Jan. 30, 2011

Dear Dillo developers, since you started porting Dillo to FLTK 1.3, I wonder if you plan to do a last release of the current FLTK 2.0 Dillo? (This would seem reasonable to me, since this version works pretty well and some of us can't wait so long for the next release. :) Many thanks in advance, Alexander Voigt

7 10

[patch]: clean up the build and make d_size.h static.
by onepoint＠starurchin.org Jan. 28, 2011

Jan. 28, 2011

These patches clean up the build system some more and then address the issue raised here[1], by using autoconf to define int32_t etc. and turning d_size.h into a static source file instead of building it. A nice simplification! Note that it depends on adding -I.$(top_srcdir) to AM_CPPFLAGS so that d_size.h can be #include-ed anywhere. 1: Finish the work begun in changeset 90603f572f75 by replacing paths that use ../ with full paths from $(top_builddir). 2: Add -I$(top_srcdir) to AM_CPPFLAGS everywhere. (This is necessary to allow the next change.) 3: Determine int32_i and friends with autoconf macros and make d_size.h a static file. Now "make distcheck" works at last! (It used to fail because d_size.h was left behind after a clean.) Comments are welcome. Regards, Jeremy Henty [1] http://lists.auriga.wearlab.de/pipermail/dillo-dev/2008-October/005135.html

3 9

1.3 port status
by obeythepenguin＠gmail.com Jan. 27, 2011

Jan. 27, 2011

Oops, I meant to send this to the list, not just Alexander -- apologies for that. On Wed, Jan 26, 2011 at 9:12 AM, Alexander Voigt <Hole.destructor(a)gmx.de> wrote: > Maybe this is a solution to the upcomming problem(s). ?What one needs > to do is to build an abstraction and an implementor and pack all > toolkit specific things into concrete implementors (i.e. the FLTK 2.0, > or FLTK 1.3 specific things). There's one obvious problem there: the toolkit is *already* an abstraction. ?(FLTK's not strictly an abstraction layer, because it implements its own widget set, but the purpose of any toolkit is to provide some sort of abstraction over the native APIs.) On top of that, this "solution" doesn't really solve anything -- it still takes time away from development, requires developers to learn the new abstraction layer, and won't make up for functional incompatibilities/missing features between different toolkits or versions. ?It sounds like a great idea in theory -- and IIRC, the Dillo renderer (but not the rest of the interface) is already built on an abstraction layer -- but in practice it would mostly amount to needlessly re-inventing the wheel. Disclaimer: I'm not one of the core developers, and this is only my personal view. ~Benjamin

2 1

1.3 port status
by rogerx.oss＠gmail.com Jan. 26, 2011

Jan. 26, 2011

> On Mon, Jan 24, 2011 at 06:11:31PM +0000, corvid wrote: >Roger wrote: >> 'hg update' && patch from this thread. > >Your backtrace doesn't match up for dw/layout.cc, for instance. >Is there an 'hg pull' in there as well? >(Also, if it wasn't clear, a new version of the UI patch went up >yesterday.) > >> Going from memory, I believe they were window functions calling the old 2.0 >> branch names instead of 1.3. >> >> I believe I already sent you the details, but since they were only test >> programs and didn't hinder dillo from being built, they aren't a high priority. > >I fixed up the Browser test to compile a couple of days ago. Yup. I performed the 'hg tip' and compared with your times and mine still showed older. This is exactly what I was afraid of. svn/cvs update seems so much friendlier! So, I just wiped and re-cloned. Within the following compiler error stdout's, please ignore the "-I/usr/include/fltk-1.1" as the package maintainer for fltk-1.3 ebuilds on bugs.gentoo.org is working to get the library slotted correctly. I only have fltk-1.3 and fltk-2.0 libraries installed here. Here's that pthread issue again: g++ -I/usr/include/libpng14 -I/usr/include/fltk-1.1 -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -g -O2 -Wall -W -Wno-unused-parameter -fno-rtti -fno-exceptions -L/usr/local/lib -o dillo dillo.o paths.o ui.o uicmd.o bw.o cookies.o auth.o colors.o misc.o history.o prefs.o prefsparser.o keys.o url.o bitvec.o klist.o chain.o utf8.o timeout.o dialog.o web.o nav.o cache.o decode.o dicache.o capi.o css.o cssparser.o styleengine.o plain.o html.o form.o table.o bookmark.o dns.o gif.o jpeg.o png.o imgbuf.o image.o menu.o dpiapi.o findbar.o xembed.o ../dlib/libDlib.a ../dpip/libDpip.a IO/libDiof.a ../dw/libDw-widgets.a ../dw/libDw-fltk.a ../dw/libDw-core.a ../lout/liblout.a -ljpeg -L/usr/lib -lpng14 -L/usr/lib/fltk-1.1 -Wl,-rpath,/usr/lib/fltk-1.1 -lfltk -lXext -lXft -lfontconfig -ldl -lm -lX11 -lz dns.o: In function `Dns_server_req': /home/roger/src/dillo/dillo_port1.3/src/dns.c:360: undefined reference to `pthread_create' Too resolve, ./configure --disable-threaded-dns (or am I missing a dep?... I don't think so.) Here's the Flock() function call to the old fltk-2.0 lock function: g++ -I/usr/include/fltk-1.1 -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -g -O2 -Wall -W -Wno-unused-parameter -fno-rtti -fno-exceptions -L/usr/local/lib -o downloads.dpi downloads_dpi-downloads.o dpiutil.o -L/usr/lib/fltk-1.1 -Wl,-rpath,/usr/lib/fltk-1.1 -lfltk -lXext -lXft -lfontconfig -ldl -lm -lX11 ../dpip/libDpip.a ../dlib/libDlib.a downloads_dpi-downloads.o: In function `main': /home/roger/src/dillo/dillo_port1.3/dpi/downloads.cc:1105: undefined reference to `Fl::lock()' Test programs compiled without errors. On execute, I see the "Back, Forward, Home, ..." buttons with a slice of the bottom of the "File, View, ..." stuff. Everything else seems frozen. If I'm not mistaken, I can see the toolbar window extend a third of the way down the screen. Then the next 2/3's I see the light grey browser window with slider. Then 3/3 I can see the top slice of "File, View, ..." toolbar window. Oh. I found the hidden "File" button from "File, View, ..." toolbar. But I can't open a url. But did get a new blank tab open. No pages loaded yet. ctrl-s gets me search the web, and ctrl-r seems to be reloading an invisible page (likely google.com as I see it loads two images from the text info). ctrl-l (open file) and ctrl-n (new window) works. ctrl-l fails to open an applet (open url). Looks like pages are loading by looking at the console and the text info detailing "images loaded", but everything is invisible ... windows borking? Looks like rendering pages is being done invisible because the applets/windows are failing.

4 11

1.3 port status
by corvid＠lavabit.com Jan. 25, 2011

Jan. 25, 2011

Roger wrote: > Capi_filters_test: DENY from 'news.google.com' to 'nt3.ggpht.com' If you want all of those, in dillorc, change filter_auto_requests to allow_all.

1 0

1.3 port status
by corvid＠lavabit.com Jan. 24, 2011

Jan. 24, 2011

The fltk-1.3 port is decently usable now if you take http://hg.dillo.org/dillo_port1.3/ and apply my quick hacky UI patch http://www.dillo.org/test/1.3_ui_hardcoded.patch (Anyone want to write a proper one? :) I see that it's segfaulting when leaving pages with image inputs because of something to do with complexbutton, but...decently usable, certainly.

4 9