- Dillo-dev - lists.mailman3.com

[Dillo-dev]Re: race condition patch

by Matthias Franz

On Fri, Oct 15, 2004 at 10:16:54AM -0300, Jorge Arellano Cid wrote: > BTW, did you had the time to test the race condition patch in > the CVS? Yes, but not very much. I've tested several critical pages, and there were no problems at all. I want to test it more thoroughly during the weekend. All the best, -- Matthias Franz Section de Mathématiques, Université de Genève, Suisse

Oct. 15, 2004

Re: [Dillo-dev]anchors_table not cleared

by Jorge Arellano Cid

On Fri, Oct 15, 2004 at 02:51:21PM +0200, Matthias Franz wrote: > Hi everybody, > > I'm writing a patch for a better treatment of anchor names in HTML pages. > During this, I've come across the following bug in Dillo. > > Jorge: I think this is what produced the bug in the patch I sent you. Thanks for the tip. (more on this later). But keep on reading... ;) > > > Entries are never removed from the hash table anchors_table, which stores > the anchors of a page (see dw_gtk_viewport.c). This means that whenever a > new page is displayed in the same viewport, the old anchors remain in > the hash table, but their keys are overwritten. Like in this example: > (The "###" lines are debug information I've inserted.) > > ~$ dillo http://www.google.de > dillo_dns_init: Here we go! > Disabling cookies. > ### Dw_gtk_viewport_init: anchors_table init > Nav_open_url: Url=>about:splash< > Nav_open_url: Url=>http://www.google.de< > Dns_server [0]: www.google.de is 0x8100ea8 > Connecting to 66.102.9.99 > ### size(anchors_table) = 0 > ### entries: > ### Html_add_anchor: adding anchor 1a > ### size(anchors_table) = 1 > ### entries: <1a> > ### Html_add_anchor: adding anchor 2a > ### size(anchors_table) = 2 > ### entries: <2a> <1a> > ### Html_add_anchor: adding anchor 3a > ### size(anchors_table) = 3 > ### entries: <3a> <2a> <1a> > ### Html_add_anchor: adding anchor 4a > ### size(anchors_table) = 4 > ### entries: <4a> <3a> <2a> <1a> > ### Html_add_anchor: adding anchor all > ### size(anchors_table) = 5 > ### entries: <4a> <all> <3a> <2a> <1a> > ### Html_add_anchor: adding anchor lgr > ### size(anchors_table) = 6 > ### entries: <4a> <all> <3a> <2a> <1a> <lgr> > ### Html_add_anchor: adding anchor cty > Nav_open_url: Url=>http://www.google.de/imghp?hl=de&tab=wi&ie=UTF-8< > Connecting to 66.102.9.99 > ### size(anchors_table) = 7 > ### entries: <> <> <h3> <> <¸1> <À(> <> > ### Html_add_anchor: adding anchor 0a > ### size(anchors_table) = 8 > ### entries: <0a> <> <> <h3> <> <¸1> <> <> > ### Html_add_anchor: adding anchor 2a > ### size(anchors_table) = 9 > ### entries: <0a> <> <> <h3> <> <2a> <¸1> <> <> > ### Html_add_anchor: adding anchor 3a > ### size(anchors_table) = 10 > ### entries: <0a> <> <> <h3> <3a> <> <2a> <¸1> <> <> > ### Html_add_anchor: adding anchor 4a > Dillo: normal exit! > > > Obviously, the anchors_table must be cleared whenever a new page is > displayed. I don't know the dw* part of Dillo very well, and the > connection between the html structure and the viewport seems to be quite > indirect. Can anyone suggest a fix? Please wait for Sebastian to answer this one. I remember being through the handling code for anchors in the past, and it has its subtleties for sure. I'll review the patch you sent me in detail now, and send you my comments on the approach. I'm busy with the rc2 work now. This patch shouldn't go in because this stage is a bugfix only (and the dillo-0.8.3 release is very delayed now :( ). BTW, did you had the time to test the race condition patch in the CVS? -- Cheers Jorge.-

Oct. 15, 2004

[Dillo-dev]anchors_table not cleared

by Matthias Franz

Hi everybody, I'm writing a patch for a better treatment of anchor names in HTML pages. During this, I've come across the following bug in Dillo. Jorge: I think this is what produced the bug in the patch I sent you. Entries are never removed from the hash table anchors_table, which stores the anchors of a page (see dw_gtk_viewport.c). This means that whenever a new page is displayed in the same viewport, the old anchors remain in the hash table, but their keys are overwritten. Like in this example: (The "###" lines are debug information I've inserted.) ~$ dillo http://www.google.de dillo_dns_init: Here we go! Disabling cookies. ### Dw_gtk_viewport_init: anchors_table init Nav_open_url: Url=>about:splash< Nav_open_url: Url=>http://www.google.de< Dns_server [0]: www.google.de is 0x8100ea8 Connecting to 66.102.9.99 ### size(anchors_table) = 0 ### entries: ### Html_add_anchor: adding anchor 1a ### size(anchors_table) = 1 ### entries: <1a> ### Html_add_anchor: adding anchor 2a ### size(anchors_table) = 2 ### entries: <2a> <1a> ### Html_add_anchor: adding anchor 3a ### size(anchors_table) = 3 ### entries: <3a> <2a> <1a> ### Html_add_anchor: adding anchor 4a ### size(anchors_table) = 4 ### entries: <4a> <3a> <2a> <1a> ### Html_add_anchor: adding anchor all ### size(anchors_table) = 5 ### entries: <4a> <all> <3a> <2a> <1a> ### Html_add_anchor: adding anchor lgr ### size(anchors_table) = 6 ### entries: <4a> <all> <3a> <2a> <1a> <lgr> ### Html_add_anchor: adding anchor cty Nav_open_url: Url=>http://www.google.de/imghp?hl=de&tab=wi&ie=UTF-8< Connecting to 66.102.9.99 ### size(anchors_table) = 7 ### entries: <> <> <h3> <> <¸1> <À(> <> ### Html_add_anchor: adding anchor 0a ### size(anchors_table) = 8 ### entries: <0a> <> <> <h3> <> <¸1> <> <> ### Html_add_anchor: adding anchor 2a ### size(anchors_table) = 9 ### entries: <0a> <> <> <h3> <> <2a> <¸1> <> <> ### Html_add_anchor: adding anchor 3a ### size(anchors_table) = 10 ### entries: <0a> <> <> <h3> <3a> <> <2a> <¸1> <> <> ### Html_add_anchor: adding anchor 4a Dillo: normal exit! Obviously, the anchors_table must be cleared whenever a new page is displayed. I don't know the dw* part of Dillo very well, and the connection between the html structure and the viewport seems to be quite indirect. Can anyone suggest a fix? Thanks in advance, -- Matthias Franz Section de Mathématiques, Université de Genève, Suisse

Oct. 15, 2004

[Dillo-dev][marianne@mathematicians.de: Compatibility]

by Jorge Arellano Cid

-- Cheers Jorge.-

Oct. 15, 2004

Re: [Dillo-dev]testing dillo-0.8.3-rc1

by Jorge Arellano Cid

On Wed, Oct 06, 2004 at 03:47:32PM +0200, Bjoern Brill wrote: > Hello, > > [...] > - This one is not new, but I fixed it a while ago for myself and then > forgot about it: the "This page uses the NON-STANDARD meta refresh > tag..." warning usually occurs inside <head>, but is immediately > sent to the parser, which in turn switches its HTML processing state > from IN_HEAD to IN_BODY prematurely. After that, head-only tags like > <title> and <base> are ignored. A minimal fix is at the end of the > mail. Done. -- Cheers Jorge.-

Oct. 15, 2004

[Dillo-dev]Crash on fast reload for dpi pending...

by Jorge Arellano Cid

Hi there, Please note that hitting reload as a mad man over a dpi-handled URL can crash Dillo. This happens because currently there's no code to account for this case (as there is for the HTTP branch). Thinking about it, there's a possibility of merging the CCC branches into a couple of generic and very powerful ones (much like what the dpi is now). If this design is possible, and succeeds, it would mean much simpler handling of IO, the elimination of a lot of complexity, more stability (all the data would use it and thus be more tested). I'll give it some thought. Not for the next release though... :-) -- Cheers Jorge.-

Oct. 14, 2004

Re: [Dillo-dev]dillo-0.8.3-rc1.tar.bz2: bug

by Jorge Arellano Cid

On Wed, Oct 13, 2004 at 02:01:01PM +0200, Matthias Franz wrote: > On Tue, Oct 12, 2004 at 12:49:21PM -0300, Jorge Arellano Cid wrote: > > Well, I've devoted several days to this BUG. > > Dear Jorge, > > the time and energy you have been spending on this bug shows me > how much you are dedicated to high quality software. > > Thanks a lot! > > PS: Once you have a bugfix, I'd be happy to test it. A patch is in the CVS! Finally I figured it out: there were two race conditions, one on closing the write end of the socket (the easy one) and another on the thin time-slice between server FD close and the IO watch for it being scheduled inside Dillo. The problem was very hard to spot as it wasn't on a variable shared explicitly among threads, but on a file descriptor. If somewhere inside the above mentioned time-slice, Dillo requested for a new FD and the kernel noticed, and decided to reuse the FD, then there were problems. I say there were problems because the data transfer had a good chance to succeed on a fascinatingly wicked way! (covering up the BUG). The patch makes changes in several places, the most interesting of them being the threads list being sorted by a primary key other than the FD (this allows handling the CCCs for both, the just closed and the just openned one). I left the thread list's debug messages for checking purposes. If you see two identical FDs in the list, that was the bug (handled now). It worked OK on all my tests for the file dpi server and also for the file dpi filter. I left the dpi server option as default because it spawns less processes. If you want to test the file filter dpi, just compile dillo and run from the command line. To test the file dpi server, do a: make install-strip and remove, or 'chmod -x', the file.filter.dpi. (usually /usr/local/lib/dillo/dpi/file/file.filter.dpi). ... and tell us how it works. -- Cheers Jorge.-

Oct. 14, 2004

Re: [Dillo-dev]dillo-0.8.3-rc1.tar.bz2: bug

by Matthias Franz

On Tue, Oct 12, 2004 at 12:49:21PM -0300, Jorge Arellano Cid wrote: > Well, I've devoted several days to this BUG. Dear Jorge, the time and energy you have been spending on this bug shows me how much you are dedicated to high quality software. Thanks a lot! PS: Once you have a bugfix, I'd be happy to test it. -- Matthias Franz Section de Mathématiques, Université de Genève, Suisse

Oct. 13, 2004

[Dillo-dev]Webmails compatibility with Dillo 0.8.3rc1

by Florent BERANGER

Hi, Here a little test result about webmails compatibility : Read Compose Horde : ok nok Squirrel : ok ok -> you just have to select a frame and it's ok:) Tiscali : nok nok Ms hotmail: nok nok Google gmail: nok nok Yahoo mail: nok nok Best regards, Florent ************************ ADSL ILLIMITE TISCALI + TELEPHONE GRATUIT ************************ Surfez 40 fois plus vite pour 30EUR/mois seulement ! Et téléphonez partout en France gratuitement, vers les postes fixes (hors numéros spéciaux). Tarifs très avantageux vers les mobiles et l'international ! Pour profiter de cette offre exceptionnelle, cliquez ici : http://register.tiscali.fr/adsl (voir conditions sur le site)

Oct. 12, 2004

[jcid@dillo.org: Re: [Dillo-dev]Caching]

by Jorge Arellano Cid

Hi, Here's an email that I mistakenly replied personally instead of to dillo-dev. -- Cheers Jorge.-

Oct. 12, 2004