December 2012 - Dillo-dev - lists.mailman3.com

config.h.in and configure.in

by onepoint＠starurchin.org

I thought that we agreed[1] to remove config.h.in from the repository and to rename configure.in to configure.ac . Can I make these changes or is more discussion needed? Regards, Jeremy Henty [1]: http://lists.auriga.wearlab.de/pipermail/dillo-dev/2012-December/009618.html

Dec. 19, 2012

Patch for xembed (updated for resize)

by johannes.hofmann＠gmx.de

Hi, On Sat, Dec 15, 2012 at 12:04:35PM +0100, podarcis wrote: > Hi, > > here is an updated version making resize work. > It still works with tabbed too. > Note that I have had no knowledge of X or xembed at all, so all > changes come from experimenting... > I removed large parts of the existing code, because they did nothing > in my environment. > > Please have a look at it. hm, the patch basically removes support for the xembed protocol. Is it meant as a diagnostic patch to investigate a problem, or meant for inclusion into the dillo repository? It won't work for the latter case. You can't just remove code that is not needed for your specific case. If you just want to get embedding of dillo into an fltk based window working, you would need to send the proper xembed X11 event from your embedding fltk program. Look at what tabbed is doing. I can also dig into my old stuff, I think I once had a minimal fltk based embedder working. Cheers, Johannes

Dec. 19, 2012

dpid patches

by p37sitdu＠lavabit.com

First patch is for removing expired TODO item, just resending it. Second patch is new, for dpid.c. Function fill_services_list has a bug: on line 463 there is a check *services_list != NULL. If it is true, then -1 is returned, but dpidrc_stream is not closed. Error message on line 456 says "popen failed", should be "fopen failed". fill_services_list checks for availability of user_dpidir and sys_dpidir, but after line 449 these variables are not used. They are just dFreed in the end. Looks like code was just copypasted from register_all. fill_services_list is only called from main.c:248 and dpid.c:807 of dpid.c. register_all is called from main.c:235 and dpid.c:806. So fill_services_list is always called after register_all and doing the same check is useless. On line 503 there is a conversion of services_alpha_comp to type (dCompareFunc). I changed the function services_alpha_comp to the usual style of writing comparison functions so no conversion is required. In the current state it is undefined behaviour. [1] [1] http://stackoverflow.com/questions/559581/casting-a-function-pointer-to-ano…

Dec. 18, 2012

[patch] Gopher DPI

by p37sitdu＠lavabit.com

Fixed memory leak related to dGetline.

Dec. 17, 2012

[patch] Gopher DPI

by p37sitdu＠lavabit.com

Small changes - Added 'p' filetype for images (used by some gopher servers) - Use dGetline instead of getline

Dec. 17, 2012

Tooltip bugs revisited

by jcid＠dillo.org

On Tue, Dec 11, 2012 at 04:30:59AM +0000, corvid wrote: > Jorge wrote: > > So far (and despite the warning inside FLTK), I've experimented > > with a custom tooltip handling class, and a couple custom UI > > widgets, that seem to work well together (i.e. solve the problem > > without the workaround :) ). > > I see that this code is derived from the code at > http://seriss.com/people/erco/fltk/#SliderTooltip > I think we should mention that at the tops of the tipwin files and > in AUTHORS. Yes, I wrote to Greg and got his answer stating that all the code there is public domain, and that he'll soon add some lines declaring it so in his "Cheat Page". -- Cheers Jorge.-

Dec. 17, 2012

Dpi_connect_socket retry

by jcid＠dillo.org

Hi, Finally I had the time to review this issue in more detail... On Sun, Dec 09, 2012 at 02:58:43PM +0400, p37sitdu at lavabit.com wrote: > Dpi_connect_socket function from src/IO/dpi.c (also duplicated in > test/cookies.c) has retry argument. It is set to TRUE on first call, > and Dpi_connect_socket calls itself recursively with retry == FALSE in > case of error. > > This retry function is not working. If connection fails, at line 645 > sock_fd is set, but ret is still set to -1. Even if the second attempt > to connect is successful, -1 is returned and conected sock_fd is lost. > > It is possible to fix it by adding "ret = sock_fd", but I think it is > better to just remove retries completely. It is unlikely that > connecting using loopback interface will fail on first try and succeed > on retry. Yes, this code is obsolete now. It was meant for Unix domain sockets (when dpis/dpid used them). Sometimes the filename lingered in the filesystem after a dpi or server crash. With unlink&retry it was possible to recover then. Details in patchset: 7faa2c7a544f (Nov, 2009 ;) Removed in the repo (see below). > Also, comment says "the server closes sock_fd on error". However, > sock_fd should still be closed by client to free allocated file > descriptor. OS don't know if Dillo is going to read from it (in which > case error such as ECONNRESET should happen) so file descriptor can't > be reused. > > File descriptor should be closed in case of error in a_Dpip_build_cmd > too. The file descriptor is properly closed by its CCC afterwards (Visible by defining VERBOSE in chain.c). The CCC manages the error propagation/handling, in both the requesting and answer branches. > Patch to fix this function and its calls in src/IO/dpi.c and > test/cookies.c is attached. No need to fix, it works properly. -- Cheers Jorge.-

Dec. 16, 2012

Extensions for hyphenation

by corvid＠lavabit.com

Sebastian wrote: > On Sat, Dec 08, corvid wrote: > > Sebastian wrote: > > > On Fri, Dec 07, corvid wrote: > > > > For instance, > > > > > > > > dillo -geometry 850x700 http://fltk.org > > > > > > > > makes the names of the snapshots very narrow, and if I try > > > > > > > > penalty_hyphen=inf > > > > > > > > part of the fltk-2 snapshot link is obscured by the image. > > > > I still see this, or at least the penalty_hyphen=inf case. I didn't try > > the default penalty_hyphen case. > > I was finally able to reproduce this, by using "trickle". > > It is fixed now, at least for me; please test it again. It's working :)

Dec. 16, 2012

Extensions for hyphenation

by sgeerken＠dillo.org

On Sat, Dec 08, corvid wrote: > Sebastian wrote: > > On Fri, Dec 07, corvid wrote: > > > For instance, > > > > > > dillo -geometry 850x700 http://fltk.org > > > > > > makes the names of the snapshots very narrow, and if I try > > > > > > penalty_hyphen=inf > > > > > > part of the fltk-2 snapshot link is obscured by the image. > > I still see this, or at least the penalty_hyphen=inf case. I didn't try > the default penalty_hyphen case. I was finally able to reproduce this, by using "trickle". It is fixed now, at least for me; please test it again. As far as I see, this is the last bug related to hypenation, at least the last one relevant for the next release. Sebastian

Dec. 16, 2012

[patch] Gopher DPI

by jcid＠dillo.org

On Sun, Dec 16, 2012 at 12:31:19AM +0400, p37sitdu at lavabit.com wrote: > Now search works properly. URL looks like > gopher://gopher.floodgap.com:70/7/v2/vs?query=dillo, but ?query is > replaced with %09 internally. I have not found the way to make > redirect. The patch is ready to be commited and should not break > anything as it is a separate DPI that is only called for gopher URLs. > > I will try to implement search without forms by adding input dialogs > as a separate patch in my patch queue on top of it. User experience > will be similar to lynx: you open the page and then browser asks you > for search query. Are you sure a dialog is better than a simple form (as with bookmarks)? I'll review the patch as soon as I find the necessary time. So far from what I've read in the list, it looks OK. Just some friendly advice: don't rush; you'll find plenty of options along the way, and independent as a dpi is, it looks much like it will make it into the next release. > Also patch for src/IO/dpi.c is attached to make Dpi_parse_token more > robust as I accedently made it segfault during development of gopher > DPI. Yes, there're subtle interactions and assumptions in dpip. That's why I'd like to take the necessary time to review the FD leak patch and then the gopher dpi. -- Cheers Jorge.-

Dec. 15, 2012