- Dillo-dev - lists.mailman3.com

Dillo working directory

by p37sitdu＠lavabit.com

On Sun, May 27, 2012 at 04:57:24PM +0400, 123 wrote: > In Paths::init Dillo changes working directory to /tmp. This way every > time I try to open some page stored on the hard disk (Ctrl-O) I have > to start in /tmp. What is the reason for changing WD to /tmp? BTW > oldWorkingDir is not used anywhere. oldWorkingDir is used in fact, it was a mistake. I have searched mailing list and found thread [1] [2] [3] discussing the same issue, but there is no rationale. I also found line in ChangeLog - Fixed Dillo and its dpis to work from "/tmp" (for easy device unmount). Most of the time Dillo is running from $HOME and unmounting it is not common. It may be that someone changes directory to, for example, /mnt, runs dillo there and exits shell (or changes its WD), but it is not common scenario too. The problem with changing directory to /tmp is that open dialog starts in /tmp, it is a bug. As a workaround I am typing "file:" in location, it uses oldWorkingDir then. I have attached patch, it fixes the problem. It is still not finished without removing obsolete oldWorkingDir and using real current directory. Are there any other hidden problems? If not, I will make finished patch. [1] http://lists.auriga.wearlab.de/pipermail/dillo-dev/2011-December/009221.html [2] http://lists.auriga.wearlab.de/pipermail/dillo-dev/2011-December/009223.html [3] http://lists.auriga.wearlab.de/pipermail/dillo-dev/2011-December/009225.html

June 14, 2012

dillo.cc cleanup and bugfix

by p37sitdu＠lavabit.com

On Thu, Jun 14, 2012 at 07:00:17PM +0000, corvid wrote: > Johannes wrote: > > On Sun, Jun 10, 2012 at 10:39:21PM +0400, 123 wrote: > > > aboutblank.diff removes unnecessary check. > > > > Removed check for about:blank as start_page, it is not a special case. > > > > Committed. > > about:blank no longer gives focus to Location. Right, it is because a_UIcmd_open_url works differently for NULL and about:blank. And it is commented in *identical* static function UIcmd_open_url_nbw. I overlooked this difference. If you change it back, add comment please.

June 14, 2012

dillo.cc cleanup and bugfix

by Johannes.Hofmann＠gmx.de

On Thu, Jun 14, 2012 at 07:00:17PM +0000, corvid wrote: > Johannes wrote: > > On Sun, Jun 10, 2012 at 10:39:21PM +0400, 123 wrote: > > > aboutblank.diff removes unnecessary check. > > > > Removed check for about:blank as start_page, it is not a special case. > > > > Committed. > > about:blank no longer gives focus to Location. > ah, the details... I reverted it.

June 14, 2012

dillo.cc cleanup and bugfix

by corvid＠lavabit.com

Johannes wrote: > On Sun, Jun 10, 2012 at 10:39:21PM +0400, 123 wrote: > > aboutblank.diff removes unnecessary check. > > > Removed check for about:blank as start_page, it is not a special case. > > Committed. about:blank no longer gives focus to Location.

June 14, 2012

[patch] VOIDP2INT, INT2VOIDP and pointer size

by joerg.sonnenberger＠web.de

On Wed, Jun 13, 2012 at 12:38:33PM -0400, Jorge Arellano Cid wrote: > On Tue, Jun 12, 2012 at 10:46:38PM +0200, Johannes Hofmann wrote: > > On Tue, Jun 12, 2012 at 11:25:33PM +0400, 123 wrote: > > > Now (long) is used for converting (void *) to integer. This patch > > > replaces (long) with intptr_t. > > > > > > I don't think this fixes the real problem. When VOIDP2INT and > > > INT2VOIDP macros are used, it is possible that converted type won't > > > fit into intptr_t. It is better to rewrite code that uses these macros > > > in portable way and remove these macros. > > > > Agreed, so let's go for the real fix - step by step. > > Even better would be if we could come up with a case where the > > current scheme causes a real problem - not sure if this is possible, > > some compiler warnings maybe? > > There's an old adage: if it ain't broken don't fix it. > > I'd much prefer if you could focus your patches on real bugs > instead of cleanups or changes to code form. Microsoft uses IL32LLP64 as memory model for Win64. So it certainly "ain't broken". Joerg

June 13, 2012

[patch] VOIDP2INT, INT2VOIDP and pointer size

by jcid＠dillo.org

On Tue, Jun 12, 2012 at 10:46:38PM +0200, Johannes Hofmann wrote: > On Tue, Jun 12, 2012 at 11:25:33PM +0400, 123 wrote: > > Now (long) is used for converting (void *) to integer. This patch > > replaces (long) with intptr_t. > > > > I don't think this fixes the real problem. When VOIDP2INT and > > INT2VOIDP macros are used, it is possible that converted type won't > > fit into intptr_t. It is better to rewrite code that uses these macros > > in portable way and remove these macros. > > Agreed, so let's go for the real fix - step by step. > Even better would be if we could come up with a case where the > current scheme causes a real problem - not sure if this is possible, > some compiler warnings maybe? There's an old adage: if it ain't broken don't fix it. I'd much prefer if you could focus your patches on real bugs instead of cleanups or changes to code form. -- Cheers Jorge.-

June 13, 2012

dillo.cc cleanup and bugfix

by p37sitdu＠lavabit.com

On Tue, Jun 12, 2012 at 10:26:45PM +0200, Johannes Hofmann wrote: > > geometry.diff fixes --geometry flag > > > Fixed --geometry flag. It is --geometry in man page but -geometry in > > > usage and only -geometry is recognized by program. > > I'm not sure here. -geometry seems to be somewhat standard [1] and I > don't want to break people's scripts etc without very good > reason. So for now I adjusted the man page. It is standard for traditional X applications (xclock, xterm etc.) to have only "long" options that start with one dash together with corresponding resource. Standard X tools mostly use XrmParseCommand [1] and you can see examples of options that start with one dash in Xlib documentation. I would even remove short options (like -v) and replace them with standard long options such as -version, but I guess you don't want to break compatibility for it. > > getcmdoption.diff simplifies getCmdOption > > > Simplified getCmdOption. Removed State enum, it was overkill. It is > > > also possible to remove support for mandatory arguments (opt_argc < 0) > > > as it is not used, but it does not add LOC so I left it there. > > Hm, if the patch would fix some bug - even some minor stuff - I would > commit it together with the cleanup, but not just for the reduced > LOC. > Any other takers? Now it looks like someone started writing a state machine for option parsing but then created a separate function for parsing just one option. Or maybe it is copied from legacy code. I will keep it until I find some related bug then, but IMO it is not good to mix cleanup with bugfixes (and it is not good to mix different patches in one in general). [1] http://www.tronche.com/gui/x/xlib/resource-manager/XrmParseCommand.html

June 12, 2012

[patch] VOIDP2INT, INT2VOIDP and pointer size

by Johannes.Hofmann＠gmx.de

On Tue, Jun 12, 2012 at 11:25:33PM +0400, 123 wrote: > Now (long) is used for converting (void *) to integer. This patch > replaces (long) with intptr_t. > > I don't think this fixes the real problem. When VOIDP2INT and > INT2VOIDP macros are used, it is possible that converted type won't > fit into intptr_t. It is better to rewrite code that uses these macros > in portable way and remove these macros. Agreed, so let's go for the real fix - step by step. Even better would be if we could come up with a case where the current scheme causes a real problem - not sure if this is possible, some compiler warnings maybe? Cheers, Johannes

June 12, 2012

[patch] HTML parser bugfix

by Johannes.Hofmann＠gmx.de

On Thu, Jun 07, 2012 at 04:45:42PM +0400, 123 wrote: > On Thu, Jun 07, 2012 at 12:14:06AM +0100, Jeremy Henty wrote: > > > > 123 wrote: > > > > > On Sun, Jun 03, 2012 at 09:00:55PM +0100, Jeremy Henty wrote: > > > > > > If you don't detect and ignore that extra double quote you will > > > > break many pages that every other browser renders perfectly well. > > > > > > Then there should be some logic for detecting double quotes. > > > > I agree. The hard question is: what logic? > > IMO the right way to do it is to implement HTML Standard [1]. When > Html_write_raw is called, parser is in the Data state. When it returns > to Data state again, it is the end of token. > > I have implemented standard comment/DOCTYPE parsing. It is incomplete, > EOF is not handled and DOCTYPE parsing is not changed. Patch is > attached. Next step is to rewrite tag parsing in standard way. > > [1] http://www.whatwg.org/specs/web-apps/current-work/multipage/tokenization.ht… With your patch text disappears e.g. on http://www.cnas.org/blogs/abumuqawama/2011/04/quote-day.html-0 which was mentioned by Jeremy while it renders ok with current dillo and firefox. I would rather put together a test page that includes all the cases Jeremy brought up plus the reddit one. Also looking into firefox or other browser sources might be a good start. Cheers, Johannes

June 12, 2012

dillo.cc cleanup and bugfix

by Johannes.Hofmann＠gmx.de

Hi, On Sun, Jun 10, 2012 at 10:39:21PM +0400, 123 wrote: > I have split patch into three patches. Good! > > geometry.diff fixes --geometry flag > > Fixed --geometry flag. It is --geometry in man page but -geometry in > > usage and only -geometry is recognized by program. I'm not sure here. -geometry seems to be somewhat standard [1] and I don't want to break people's scripts etc without very good reason. So for now I adjusted the man page. > > getcmdoption.diff simplifies getCmdOption > > Simplified getCmdOption. Removed State enum, it was overkill. It is > > also possible to remove support for mandatory arguments (opt_argc < 0) > > as it is not used, but it does not add LOC so I left it there. Hm, if the patch would fix some bug - even some minor stuff - I would commit it together with the cleanup, but not just for the reduced LOC. Any other takers? > > aboutblank.diff removes unnecessary check. > > Removed check for about:blank as start_page, it is not a special case. Committed. > > Minor fixes (#include order, indentation) are not included. Good. Cheers, Johannes [1] http://www.progsoc.org/tfm/tfm97/node32.html

June 12, 2012