- Dillo-dev - lists.mailman3.com

Startup time of the latest dev versions.

by onepoint＠starurchin.org

John Found wrote: > "First" means first after restart of the computer. I am not sure if > later, after some other programs started the startup time is not > again slow. You know, such things are pretty hard to track. But even > after the "first" startup, the subjective impression is that the dev > versions starts much slower than 3.0.5; But once started, they > render (and generally browse) faster than 3.0.5. This is a *wild* guess, but the dev versions have a different ssl library now. Perhaps it is trying to collect entropy when the programs starts? You would expect that to take longer immediately after a reboot. Crypto libraries have been known to do surprising things to gather entropy, such as reading /etc/passwd ! Regards Jeremy Henty

Sept. 4, 2016

Startup time of the latest dev versions.

by johnfound＠asm32.info

On Sat, 3 Sep 2016 23:43:15 -0700 Ben Woolley <tautolog at gmail.com> wrote: > > > On Sep 3, 2016, at 11:10 PM, John Found <johnfound at asm32.info> wrote: > > > > On Sat, 3 Sep 2016 21:30:07 +0100 > > Nick Warne <nick at linicks.net> wrote: > > > >> Hi John, > >> > >> On Sat, 3 Sep 2016 21:12:46 +0300 > >> John Found <johnfound at asm32.info> wrote: > >> > >>> I am using Dillo as a help viewer of my IDE, because it has really > >>> short startup times and pressing F1 I get instant access to the > >>> documentation. > >>> > >>> Recently I tried to compile and use the recent development versions > >>> from the hg repository, because of the recent changes in the > >>> rendering and the general improvements in the browser. > >>> > >>> But unfortunately, the startup time seems to be increased and now I > >>> have to wait 2..3 seconds for browser to start and to show my help > >>> files. > >> > >> Is that not fast? You blink and miss it. > > > > Well, no, I don't blink for 2..3 seconds. More like 300ms; :) > > The official Dillo 3.0.5 always starts instantly. > > > > I made some more tests and found that the "big wait" is actually the first time > > you start Dillo - there is an intensive hard disk activity (swap?) without any messages in the console. > > > > The next times, Dillo starts faster, but still noticeable slower than v 3.0.5 > > > > v3.0.5 always starts instantly, at least I can't see any delay or difference between the first start and the next. > > > > Can you go into more detail on what constitutes "first" here? Maybe disk cache explains most of that. Are the two binaries on the same disk? Maybe one on an SSD system disk, and another in some home folder on a spinning disk? There could be lots of different things at play here. Are you rebuilding 3.0.5 in the same way, in the same parent folder? "First" means first after restart of the computer. I am not sure if later, after some other programs started the startup time is not again slow. You know, such things are pretty hard to track. But even after the "first" startup, the subjective impression is that the dev versions starts much slower than 3.0.5; But once started, they render (and generally browse) faster than 3.0.5. There is only one hard disk (not SSD) on the computer. The different Dillo versions are on the same hard disk. The v3.0.5 has been installed from the applications repository (Manjaro/Arc Linux). -- http://fresh.flatassembler.net http://asm32.info John Found <johnfound at asm32.info>

Sept. 4, 2016

Startup time of the latest dev versions.

by tautolog＠gmail.com

> On Sep 3, 2016, at 11:10 PM, John Found <johnfound at asm32.info> wrote: > > On Sat, 3 Sep 2016 21:30:07 +0100 > Nick Warne <nick at linicks.net> wrote: > >> Hi John, >> >> On Sat, 3 Sep 2016 21:12:46 +0300 >> John Found <johnfound at asm32.info> wrote: >> >>> I am using Dillo as a help viewer of my IDE, because it has really >>> short startup times and pressing F1 I get instant access to the >>> documentation. >>> >>> Recently I tried to compile and use the recent development versions >>> from the hg repository, because of the recent changes in the >>> rendering and the general improvements in the browser. >>> >>> But unfortunately, the startup time seems to be increased and now I >>> have to wait 2..3 seconds for browser to start and to show my help >>> files. >> >> Is that not fast? You blink and miss it. > > Well, no, I don't blink for 2..3 seconds. More like 300ms; :) > The official Dillo 3.0.5 always starts instantly. > > I made some more tests and found that the "big wait" is actually the first time > you start Dillo - there is an intensive hard disk activity (swap?) without any messages in the console. > > The next times, Dillo starts faster, but still noticeable slower than v 3.0.5 > > v3.0.5 always starts instantly, at least I can't see any delay or difference between the first start and the next. > Can you go into more detail on what constitutes "first" here? Maybe disk cache explains most of that. Are the two binaries on the same disk? Maybe one on an SSD system disk, and another in some home folder on a spinning disk? There could be lots of different things at play here. Are you rebuilding 3.0.5 in the same way, in the same parent folder? >> >>> Is such slow startup time side effect of some changes that will be >>> fixed some day, or it is kind of compromise, because of other >>> improvements and this way will persists in the project from now on? >>> >>> Or (what is pretty possible) I messed something with the compilation? >> >> What ./configure line do you use, and what install options? (I use >> make install-strip) > > I am using "./configure --enable-ssl", then "make", "strip ./dillo" and then copy the browser manually in "/usr/local/bin" > I will attach the "config.log" file to this message for details. > > -- > http://fresh.flatassembler.net > http://asm32.info > John Found <johnfound at asm32.info> > <config.log> > _______________________________________________ > Dillo-dev mailing list > Dillo-dev at dillo.org > http://lists.dillo.org/cgi-bin/mailman/listinfo/dillo-dev

Sept. 4, 2016

Startup time of the latest dev versions.

by johnfound＠asm32.info

On Sat, 3 Sep 2016 21:30:07 +0100 Nick Warne <nick at linicks.net> wrote: > Hi John, > > On Sat, 3 Sep 2016 21:12:46 +0300 > John Found <johnfound at asm32.info> wrote: > > > I am using Dillo as a help viewer of my IDE, because it has really > > short startup times and pressing F1 I get instant access to the > > documentation. > > > > Recently I tried to compile and use the recent development versions > > from the hg repository, because of the recent changes in the > > rendering and the general improvements in the browser. > > > > But unfortunately, the startup time seems to be increased and now I > > have to wait 2..3 seconds for browser to start and to show my help > > files. > > Is that not fast? You blink and miss it. Well, no, I don't blink for 2..3 seconds. More like 300ms; :) The official Dillo 3.0.5 always starts instantly. I made some more tests and found that the "big wait" is actually the first time you start Dillo - there is an intensive hard disk activity (swap?) without any messages in the console. The next times, Dillo starts faster, but still noticeable slower than v 3.0.5 v3.0.5 always starts instantly, at least I can't see any delay or difference between the first start and the next. > > > Is such slow startup time side effect of some changes that will be > > fixed some day, or it is kind of compromise, because of other > > improvements and this way will persists in the project from now on? > > > > Or (what is pretty possible) I messed something with the compilation? > > What ./configure line do you use, and what install options? (I use > make install-strip) > I am using "./configure --enable-ssl", then "make", "strip ./dillo" and then copy the browser manually in "/usr/local/bin" I will attach the "config.log" file to this message for details. -- http://fresh.flatassembler.net http://asm32.info John Found <johnfound at asm32.info>

Sept. 4, 2016

Startup time of the latest dev versions.

by waltdnes＠waltdnes.org

On Sat, Sep 03, 2016 at 09:12:46PM +0300, John Found wrote > Or (what is pretty possible) I messed something with the compilation? I use... CFLAGS="-O2 -march=native -fomit-frame-pointer -pipe -fno-unwind-tables -fno-asynchronous-unwind-tables" CXXFLAGS=${CFLAGS} ...for a binary that will be running on the same machine. -- Walter Dnes <waltdnes at waltdnes.org>

Sept. 4, 2016

Startup time of the latest dev versions.

by nick＠linicks.net

Hi John, On Sat, 3 Sep 2016 21:12:46 +0300 John Found <johnfound at asm32.info> wrote: > I am using Dillo as a help viewer of my IDE, because it has really > short startup times and pressing F1 I get instant access to the > documentation. > > Recently I tried to compile and use the recent development versions > from the hg repository, because of the recent changes in the > rendering and the general improvements in the browser. > > But unfortunately, the startup time seems to be increased and now I > have to wait 2..3 seconds for browser to start and to show my help > files. Is that not fast? You blink and miss it. > Is such slow startup time side effect of some changes that will be > fixed some day, or it is kind of compromise, because of other > improvements and this way will persists in the project from now on? > > Or (what is pretty possible) I messed something with the compilation? What ./configure line do you use, and what install options? (I use make install-strip) Nick -- Gosh that takes me back... or is it forward? That's the trouble with time travel, you never can tell." -- Doctor Who "Androids of Tara"

Sept. 3, 2016

Startup time of the latest dev versions.

by johnfound＠asm32.info

I am using Dillo as a help viewer of my IDE, because it has really short startup times and pressing F1 I get instant access to the documentation. Recently I tried to compile and use the recent development versions from the hg repository, because of the recent changes in the rendering and the general improvements in the browser. But unfortunately, the startup time seems to be increased and now I have to wait 2..3 seconds for browser to start and to show my help files. Is such slow startup time side effect of some changes that will be fixed some day, or it is kind of compromise, because of other improvements and this way will persists in the project from now on? Or (what is pretty possible) I messed something with the compilation? -- http://fresh.flatassembler.net http://asm32.info John Found <johnfound at asm32.info>

Sept. 3, 2016

Get Dillo to honour target=_blank

by jcid＠dillo.org

On Mon, Aug 29, 2016 at 11:53:30AM +0100, Nick Warne wrote: > Hi Jorge, > > On Sun, 28 Aug 2016 21:57:21 -0300 > Jorge Arellano Cid <jcid at dillo.org> wrote: > > > The trick is to submit pressing the middle button, that way the data > > remains in the old tab! > > Duh! I must have read the help file 20 times and still missed that. > Thanks, works a treat! I knew it was somewhere and didn't find it either... ;) It's under the Help|Tabs section, as you found. A minor refactoring of the help text looks like a good idea. -- Cheers Jorge.-

Aug. 29, 2016

Get Dillo to honour target=_blank

by jcid＠dillo.org

Hi, On Mon, Aug 29, 2016 at 01:29:39PM +0300, Martin Atukunda wrote: > Hi dillo-dev, > > While reading, incidentally, news.ycombinator.com, I came across this: > > https://dev.to/ben/the-targetblank-vulnerability-by-example I remember long ago we decided to avoid it for resons similar to popup windows, and some others probably, but the details left me... Thanks for the link. -- Cheers Jorge.-

Aug. 29, 2016

Get Dillo to honour target=_blank

by nick＠linicks.net

Hi Martin, On Mon, 29 Aug 2016 13:29:39 +0300 Martin Atukunda <matlads at gmail.com> wrote: > Hi dillo-dev, > > While reading, incidentally, news.ycombinator.com, I came across this: > > https://dev.to/ben/the-targetblank-vulnerability-by-example > Interesting! I use it a lot, but having read that page and then this one: https://css-tricks.com/use-target_blank/ Basically, if a user _wants_ to open a link in a new tab, it is up to them! I have removed all my target="_blank" stuff now, and will not use it again. Nick -- Gosh that takes me back... or is it forward? That's the trouble with time travel, you never can tell." -- Doctor Who "Androids of Tara"

Aug. 29, 2016