You are not logged in.
@WonderWoofy: Thanks, but that webkit bug.
Okay, thanks portix. Maybe when I have some free time I will look in to posting (or commenting on) a bug with webkit upstream.
Offline
A really nice browser this dwb. As, at least in my case, jumanji is having some windows resizing issues and other stuff in the latest version I have installed the community version and so far I am positively surprised.
There is one thing though that bothers me. When I add a new searchengine and try to make a search each time it searches the term 'undefined'. Has this been addressed before in this thread (to check the remaining posts in search for the fix if there is one) or is this something new?
In the same way, I use a couple of searchengines that I have stripped down to its 'essentials' (by default besides the result there is some additional content displayed in frames which I am not interested in) therefore, is it possible to add that personalized url in dwb?
I have tried so far editing the searchengines file directly but I am not sure if dwb did accept it or not. So far during my tests it seems to be affected by a similar or even the same problem of being unable to 'define' the search term.
Aside of this, and as long as I do not find something else to 'complain' about ;P, props to you portix for this nice browser.
\(o_X)/
'Cause I wanna be an Archy - An Archy in the UK // Sex Pistols
Offline
@root: That's a bug in the last release, i have posted a fix for it https://bbs.archlinux.org/viewtopic.php?id=95589&p=49. If you use dwb-hg from AUR this is already fixed.
Offline
Is it possible to assign Left and Right Arrow keys and the Shift key in the keys configuration n the "Custom commands" section?
Thanks.
Last edited by rix (2012-11-21 20:31:28)
Offline
Is it possible to assign Left and Right Arrow keys and the Shift key in the keys configuration n the "Custom commands" section?
Thanks.
Shift @Left@:...
Offline
Sorry but I can't get the Shift working.
[...]
Shift @Tab@:focus_prev
[...]
Offline
If a key doesn't work you should first check the output of xev for that key, dwb uses the same names, so Shift Tab is
@ISO_Left_Tab@:focus_prev
Offline
Doesn't work.
Xev checked. Thanks for the suggestion, I didn't know it.
Last edited by rix (2012-11-21 21:00:23)
Offline
Offline
Works. Thank so much.
But why Xev doesn't mention "Shift" but just "ISO_Left_Tab"?
Last edited by rix (2012-11-21 21:19:21)
Offline
The modifier is dwb-specific, only the name for the key isn't. If you use a modifier you will always have to specify it. Shift is an exception, since shortcuts are case-sensitive you only need to use shift for keys that don't have an output like Tab, Left, Right, ....
Offline
Got it. I think.
Thank you.
Offline
is there any way to specify which keys should be used for hints? sometimes i feel lazy and want to use just my left hand. it would be nice, e.g., to specify that only the numbers 1-5 should be used.
Last edited by anti-destin (2012-11-21 21:44:10)
Offline
@anti-destin: hint-letter-seq. It only works for hint-style letter, but you can use 12345 even if your hint-style is letter.
Offline
ahh, thanks, portix! that was simple.
Offline
Isola wrote:Thank you for a wonderful browser, portix!
I was wondering if you're interested in making it possible to have the tabs vertically, while keeping the text horizontally left-to-right
That way more letters and more tabs will be visible, while only sacrificing some width!
What do you think?
I had already implemented it but didn't push it to the repository because personally i wouldn't ever use it because it just takes too much space. Maybe i reimplement it but i'm not sure about it.
I'm pretty interested in this feature. Where would I poke around in the source to add something like this?
Offline
Hi portix
I'm registred today to archlinux only because of dwb. Thank you for this great little browser.
But...
I want to use dwb like a binary addon for a slim Linux multimedia distribution - OpenELEC
Compiled webkit and all the necesary libraries, compiled also uzbl and dwb.
One problem - kind of solved but is not very good - I applied this patch
{ "dwb:startpage", NULL, NULL, 0, html_startpage },
};
+void
+u_insert_str(GString* cont, char* fcontent, const char* inserted_str) {
+ g_string_append_printf(cont, "%s", fcontent);
+ char* strfirstocc = strstr(cont->str, "%s");
+ int pos = strfirstocc - cont->str;
+ g_string_erase(cont, pos, 2);
+ g_string_insert(cont, pos, inserted_str);
+}
+
static char current_uri[BUFFER_LENGTH];
DwbStatus
html_load_page(WebKitWebView *wv, HtmlTable *t, char *panel) {
@@ -71,12 +80,12 @@
if (path && headpath) {
/* load head */
g_file_get_contents(headpath, &filecontent, NULL, NULL);
- g_string_append_printf(content, filecontent, t->title);
+ u_insert_str(content, filecontent, t->title);
g_free(filecontent);
/* load content */
g_file_get_contents(path, &filecontent, NULL, NULL);
if (panel)
- g_string_append_printf(content, filecontent, panel);
+ u_insert_str(content, filecontent, panel);
webkit_web_frame_load_alternate_string(webkit_web_view_get_main_frame(wv), content->str, current_uri, current_uri);
g_string_free(content, true);
g_free(filecontent);
You can see that I replaced the original g_string_append_printf function. Whitout this I can not be able to go to settings, keys,... pages.
Now I have problem with one page (until now) - subtitrari.regielive.ro - this is a romanian subtitles site
when I want to search for a subtitle and push the seach (cauta) button dwb crashes.
In terminal I can see this errors
** (dwb:2507): CRITICAL **: dwb_update_status_text: assertion `gl == dwb.state.fview' failed
** (dwb:2507): CRITICAL **: void webkit_web_frame_load_alternate_string(WebKitWebFrame*, const gchar*, const gchar*, const gchar*): assertion `content' failed
Also when I close with D some tabs dwb often crashes showing before crashing repeteadly same errors** (dwb:2507): CRITICAL **: void
** (dwb:2507): CRITICAL **: void webkit_web_frame_load_alternate_string(WebKitWebFrame*, const gchar*, const gchar*, const gchar*): assertion `content' failed
** (dwb:2507): CRITICAL **: void webkit_web_frame_load_alternate_string(WebKitWebFrame*, const gchar*, const gchar*, const gchar*): assertion `content' failed
** (dwb:2507): CRITICAL **: void webkit_web_frame_load_alternate_string(WebKitWebFrame*, const gchar*, const gchar*, const gchar*): assertion `content' failed
webkitgtk 1.8.3, icu 4.8.1.1, libsoup, 2.38.1, pango 1.28.4, cairo 1.12.0, gtk 2.24.10, glib 2.34.1
In uzbl that page is working ok so I think it is not a webkit bug.
Regards
PS sorry for my not to good english
Last edited by ancalex (2012-11-22 13:21:24)
Offline
Thanks for that patch, i will decide if i accept it in about two weeks, currently i don't have time to test it.
The webkit_web_frame_load_alternate_string assertions should only appear if you don't install dwb. dwb uses some paths that are defined at compile time to avoid searching for that file, so if you want to test dwb without installing it you have to either copy lib and scripts to ~/.local/share/dwb/lib and scripts to ~/.local/share/dwb/scripts or you can 'install' it without installing it into the system with
make -B PREFIX=/home/whatever/dwb install
Offline
Hi portix
and thank you for your response.
I must clarify a little bit.
I have compiled and installed dwb under ubuntu 12.04 and is working ok, the site I mentioned is working ok
That patch is working for me, for OpeELEC, but it is not necesarry for others, the problem is with how to compile, what libraries to use so etc..
Maybe it is related to the fact that I make a package to install dwb in a specific location, and dwb is not finding his files?
I must make other tests to see what is happening, already thank you for your tip.
Will report back soon.
Regards
Offline
@portix, I keep forgetting to tell you that the PKGBUILD for dwb-hg is kinda wonky. It download and unpacks the source to $startdir/src rather than $srcdir. I just put a one liner in there to move it where it was supposed to be before continuing the script. I added this as line 20:
mv $startdir/src/* $srcdir
Now it works fine. I'm not very familiar with mercurial, so I am not sure if this is right or what... but it works.
Offline
@WonderWoofy: That is makepkg's fault, the PKGBUILD uses makepkg's builtin mercurial support:
if [[ -d ./src/$_hgrepo ]] ; then cd ./src/$_hgrepo local ret=0 hg pull || ret=$? if (( ! ret )); then hg update elif (( ret != 1 )); then return 1 fi else [[ ! -d ./src/ ]] && mkdir ./src/ hg clone $_hgroot/$_hgrepo ./src/$_hgrepo cd ./src/$_hgrepo fi
Offline
Ah, that explains a lot, thanks portix
Offline
@portix
Patch applied and all working fine as far as I can tell. I even could set up my custom search engines in the searchengines file once I realized to simply use a modified dwb submit string for my setup.
I have noticed though that memory use stays the same even if I delete tabs. This, in my case, leads in the end to a heavy swap use
Is this dwb's standard behaviour? Are there recommended tweaks to undergo 'memory clearing'?
I wasn't sure if it may be webkit related thus, in order to 'compare' (if that is actually possible), I installed my old version of jumanji (which for me performs better than the git ones) and the behaviour was the expected one, tab deleted = memory use reduced.
Therefore I guess this is a dwb related behaviour. Any 'fix' for this is most welcome.
\(o_X)/
'Cause I wanna be an Archy - An Archy in the UK // Sex Pistols
Offline
I have noticed though that memory use stays the same even if I delete tabs. This, in my case, leads in the end to a heavy swap use
Is this dwb's standard behaviour? Are there recommended tweaks to undergo 'memory clearing'?
IIRC this is webkit's standard behaviour to make reopening closed tabs faster than other browser engines by keeping the last 10 closed ones in memory. See chromium for example.
Offline
I know about 2 leaks, i don't know if i will fix it, at least i won't fix it this year, the only solution i see to fix this leak is to remove the adblocker and the plugin blocker. I don't see a possibility to implement an adblock plus compatible adblocker without the leak. The problem is, the adblocker uses webkit's dom-api to create elements on the site and elements created with the dom-api can not be unreferenced. I tried to implement it with javascript but it would make the adblocker unusably slow.
I hope that i will have time to fix it in a few months, currently i don't have time to work on dwb.
Offline