- Dillo-dev - lists.mailman3.com

Nav_repush() is hurting valgrind

by jcid＠dillo.org

On Sat, May 09, 2009 at 08:16:05PM +0100, Jeremy Henty wrote: > On Sat, May 09, 2009 at 12:19:17PM -0400, Jorge Arellano Cid wrote: > > > I've found this: > > > > ==9830== 1,420 (352 direct, 1,068 indirect) bytes in 11 blocks are definitely > > lost in loss record 153 of 192 > > [...] > > in www.lacuarta.cl > > I can reproduce this with a vanilla Dillo, but the leak disappears if > I apply the imgref patch that was suggested to me a while back. > (Patch attached.) Good! Image reference counting looks fishy inside the code. I'll review the patch. > > Please check whether the reported errors also happen in the > > vanilla repo. > > OK, I'll back out the image patches and see what the logs show then. > (It may take a few days before I can be sure if it's made a > difference.) I'd still suggest applying the imgref patch to the > mainline as it fixes your leak above. It looks as though DilloImages > aren't getting unrefed enough, and the patch fixes that. I still don't know what kind of valgrind problems are happening in vanilla dillo. This is useful information, but getting logs of a patched dillo is certainly confusing. -- Cheers Jorge.-

May 9, 2009

Nav_repush() is hurting valgrind

by onepoint＠starurchin.org

On Sat, May 09, 2009 at 12:19:17PM -0400, Jorge Arellano Cid wrote: > I've found this: > > ==9830== 1,420 (352 direct, 1,068 indirect) bytes in 11 blocks are definitely > lost in loss record 153 of 192 > [...] > in www.lacuarta.cl I can reproduce this with a vanilla Dillo, but the leak disappears if I apply the imgref patch that was suggested to me a while back. (Patch attached.) > Please check whether the reported errors also happen in the > vanilla repo. OK, I'll back out the image patches and see what the logs show then. (It may take a few days before I can be sure if it's made a difference.) I'd still suggest applying the imgref patch to the mainline as it fixes your leak above. It looks as though DilloImages aren't getting unrefed enough, and the patch fixes that. Regards, Jeremy Henty

May 9, 2009

Nav_repush() is hurting valgrind

by jcid＠dillo.org

On Sat, May 09, 2009 at 04:13:08PM +0100, Jeremy Henty wrote: > On Sat, May 09, 2009 at 10:00:55AM -0400, Jorge Arellano Cid wrote: > > On Sat, May 09, 2009 at 10:27:34AM +0100, Jeremy Henty wrote: > > > > > > The valgrind logs[1] show a recent burst of errors from calls to > > > Nav_repush(). > > > > Oh, here I can't reproduce that. I've tried with some sites and > > the valgrind log here is clean (a bit better than it used to be). > > Do you have an example URL or test case? > > Not yet. I'll see what I can dig out of the logs. I've found this: ==9830== 1,420 (352 direct, 1,068 indirect) bytes in 11 blocks are definitely lost in loss record 153 of 192 ==9830== at 0x4025D2E: malloc (vg_replace_malloc.c:207) ==9830== by 0x8081270: dMalloc (dlib.c:36) ==9830== by 0x807E46A: a_Image_new (image.cc:40) ==9830== by 0x80703C5: a_Html_image_new(DilloHtml*, char const*, int, _DilloUrl*) (html.cc:2067) ==9830== by 0x8074F49: _ZL16Html_input_imageP9DilloHtmlPKci (form.cc:1905) ==9830== by 0x8078C5A: Html_tag_open_input(DilloHtml*, char const*, int) (form.cc:484) ==9830== by 0x8071F71: _ZL16Html_process_tagP9DilloHtmlPci (html.cc:3455) ==9830== by 0x80725E8: _ZL14Html_write_rawP9DilloHtmlPcii (html.cc:3756) ==9830== by 0x807276F: DilloHtml::write(char*, int, int) (html.cc:573) ==9830== by 0x8072802: _ZL13Html_callbackiP12_CacheClient (html.cc:3647) ==9830== by 0x8060BC7: Cache_process_queue (cache.c:1154) ==9830== by 0x80603A0: a_Cache_process_dbuf (cache.c:876) in www.lacuarta.cl > > BTW, I'm talking about the vanilla repository. > > Are you testing with a patched repo? > > Yes. If it continues I'll back out any relevant patches and see if it > changes anything. I've added a list of my local patches to the > valgrind log page so that developers can see what I've done. > > http://starurchin.org/dillo/valgrind.html > > The only two that could (I think) possibly be relevant are: > > http://starurchin.org/dillo/patch/img_ref.diff.txt > http://starurchin.org/dillo/patch/make-a_Cache_stop_client-close-client.txt > > The first patch was suggested to me by the list when I first reported > the image access problems, the second is my attempt to fix it. I'd > really appreciate any comments on whether they are doing things the > right way. Now I'm working on this, comments should follow along the way. Please check whether the reported errors also happen in the vanilla repo. -- Cheers Jorge.-

May 9, 2009

Suggestion: update the Doxyfile

by onepoint＠starurchin.org

Running doxygen is annoyingly noisy. One reason for that is that the Doxyfile is from an old version of doxygen and contains obsolete stuff that triggers warnings. I ran "doxygen -u" locally and submitted the patch to the list but it broke (just) the 40K limit and stuck in moderation. You can get it here: http://starurchin.org/dillo/patch/update-Doxyfile.txt Or a developer could run "doxygen -u" instead. Anyway, I think upgrading the Doxyfile is a good idea. Thoughts? Regards, Jeremy Henty

May 9, 2009

Nav_repush() is hurting valgrind

by onepoint＠starurchin.org

On Sat, May 09, 2009 at 10:00:55AM -0400, Jorge Arellano Cid wrote: > On Sat, May 09, 2009 at 10:27:34AM +0100, Jeremy Henty wrote: > > > > The valgrind logs[1] show a recent burst of errors from calls to > > Nav_repush(). > > Oh, here I can't reproduce that. I've tried with some sites and > the valgrind log here is clean (a bit better than it used to be). > Do you have an example URL or test case? Not yet. I'll see what I can dig out of the logs. > BTW, I'm talking about the vanilla repository. > Are you testing with a patched repo? Yes. If it continues I'll back out any relevant patches and see if it changes anything. I've added a list of my local patches to the valgrind log page so that developers can see what I've done. http://starurchin.org/dillo/valgrind.html The only two that could (I think) possibly be relevant are: http://starurchin.org/dillo/patch/img_ref.diff.txt http://starurchin.org/dillo/patch/make-a_Cache_stop_client-close-client.txt The first patch was suggested to me by the list when I first reported the image access problems, the second is my attempt to fix it. I'd really appreciate any comments on whether they are doing things the right way. > The DNS here can't resolve the hostname. D'oh! My bad, that's the copy on my local machine. The correct links are above. Regards, Jeremy Henty

May 9, 2009

Nav_repush() is hurting valgrind

by jcid＠dillo.org

On Sat, May 09, 2009 at 10:27:34AM +0100, Jeremy Henty wrote: > > The valgrind logs[1] show a recent burst of errors from calls to > Nav_repush(). Oh, here I can't reproduce that. I've tried with some sites and the valgrind log here is clean (a bit better than it used to be). Do you have an example URL or test case? BTW, I'm talking about the vanilla repository. Are you testing with a patched repo? > Also, I'm currently trialling a patch to fix the image access errors > from a_Cache_stop_client() [2]. I don't think it can be the cause of > the Nav_repush() problems but I'd like a second opinion (I don't yet > fully understand how this code should work). > > [1] http://starurchin.singularity/dillo/valgrind.html > [2] http://starurchin.singularity/dillo/patch/make-a_Cache_stop_client-close-cl… The DNS here can't resolve the hostname. -- Cheers Jorge.-

May 9, 2009

[patch]: (updated) replace DW2TB with HT2TB

by jcid＠dillo.org

On Sat, May 09, 2009 at 10:52:32AM +0100, Jeremy Henty wrote: > > <headdesk> *Here's* the patch. Committed! -- Cheers Jorge.-

May 9, 2009

[patch]: (updated) replace DW2TB with HT2TB

by onepoint＠starurchin.org

<headdesk> *Here's* the patch. Jeremy Henty

May 9, 2009

[patch]: (updated) replace DW2TB with HT2TB

by onepoint＠starurchin.org

This is an updated version of the cleanup patch that replaces the convenience macro DW2TB with an even more convenient HT2TB. It doesn't change anything, but it reduces the noise in html.cc and form.cc . Please comment. Regards, Jeremy Henty

May 9, 2009

Nav_repush() is hurting valgrind

by onepoint＠starurchin.org

The valgrind logs[1] show a recent burst of errors from calls to Nav_repush(). Also, I'm currently trialling a patch to fix the image access errors from a_Cache_stop_client() [2]. I don't think it can be the cause of the Nav_repush() problems but I'd like a second opinion (I don't yet fully understand how this code should work). Regards, Jeremy Henty [1] http://starurchin.singularity/dillo/valgrind.html [2] http://starurchin.singularity/dillo/patch/make-a_Cache_stop_client-close-cl…

May 9, 2009