- Dillo-dev - lists.mailman3.com

[Patch 3/4] Key bindings

by corvid＠lavabit.com

Jorge wrote: > The patch is not 100% complete. For instance there's > a need to provide a default keysrc file in a similar way > that cookiesrc or dillorc (with useful comments). Started looking into making a default keysrc. Some of the bindings in keys.cc are not the same as those in dillo2-help.html. Is that intentional? What is the purpose of KEYS_HIDE_PANELS when we have KEYS_FULLSCREEN?

May 4, 2009

[patch]: fix comment for a_Cache_open_url()

by jcid＠dillo.org

On Sun, May 03, 2009 at 12:12:41AM +0100, Jeremy Henty wrote: > > a_Cache_open_url() never returns 0, so its comment needs fixing. > (Bonus typo fix included too.) :) Committed. -- Cheers Jorge.-

May 3, 2009

[patch]: ensure Cache_client_make_key() returns > 0

by jcid＠dillo.org

On Sat, May 02, 2009 at 11:22:06PM +0100, Jeremy Henty wrote: > > Cache_client_make_key() carefully prevents integer overflow from > making it return a negative value, but it might return 0. This is a > problem because a_Capi_open_url() either calls a_Cache_open_url() and > returns the new client key or else returns 0 to signal that it did not > call a_Cache_open_url(). If Cache_client_make_key() returns 0 then > a_Capi_open_url() will return 0 too, and so its caller will think that > a_Cache_open_url() was not called even though it was. This will stop > the new cache client from being added to the browser window that > is responsible for creating it, which is wrong. Committed. Thanks for the detailed explanation. -- Cheers Jorge.-

May 3, 2009

Style question: cleaning up NULL

by jcid＠dillo.org

On Sun, May 03, 2009 at 05:16:31PM +0100, Jeremy Henty wrote: > > On Sun, May 03, 2009 at 09:35:34AM -0400, Jorge Arellano Cid wrote: > > On Sun, May 03, 2009 at 11:27:14AM +0100, Jeremy Henty wrote: > > > > > > ... a_Url_free() and a_Image_unref() both guard against being > > > passed a NULL argument, yet in several cases the code that calls > > > these functions also checks for a NULL, which is redundant. > > > > In general the API accepts NULL (e.g. dlib), so please remove the > > redundant checks outside it. > > OK, patches attached. Committed! -- Cheers Jorge.-

May 3, 2009

[Patch 3/4] Key bindings

by jcid＠dillo.org

On Sun, May 03, 2009 at 05:51:33PM +0000, corvid wrote: > Jorge wrote: > > A new, rewritten version of the patch was committed. > > Hurray! :) > PS Don't we need something like the following: Committed. -- Cheers Jorge.-

May 3, 2009

[Patch 3/4] Key bindings

by corvid＠lavabit.com

Jorge wrote: > A new, rewritten version of the patch was committed. Hurray! PS Don't we need something like the following:

May 3, 2009

Style question: cleaning up NULL

by onepoint＠starurchin.org

On Sun, May 03, 2009 at 09:35:34AM -0400, Jorge Arellano Cid wrote: > On Sun, May 03, 2009 at 11:27:14AM +0100, Jeremy Henty wrote: > > > > ... a_Url_free() and a_Image_unref() both guard against being > > passed a NULL argument, yet in several cases the code that calls > > these functions also checks for a NULL, which is redundant. > > In general the API accepts NULL (e.g. dlib), so please remove the > redundant checks outside it. OK, patches attached. Regards, Jeremy Henty

May 3, 2009

[Patch 3/4] Key bindings

by jcid＠dillo.org

Hi, A new, rewritten version of the patch was committed. It is O(log n) instead of O(n^3), supports named keys, and multiple keybindings. It supports the traditional key bindings by default, and they can be configured in ~/.dillo/keysrc. e.g. <Ctrl>a = bookmarks F6 = back . = forward <Ctrl>b = back The patch is not 100% complete. For instance there's a need to provide a default keysrc file in a similar way that cookiesrc or dillorc (with useful comments). Please test it and polish where needed. It has had very little testing. Note: O(log n) may be an overkill, but once it was O(n) it was easy to use dlib's sorted list instead of plain list. It can also be used as a template for other sources. -- Cheers Jorge.-

May 3, 2009

Style question: cleaning up NULL

by jcid＠dillo.org

On Sun, May 03, 2009 at 11:27:14AM +0100, Jeremy Henty wrote: > > I've noticed that a_Url_free() and a_Image_unref() both guard against > being passed a NULL argument, yet in several cases the code that calls > these functions also checks for a NULL, which is redundant. I think > that if the API officially accepts NULL then code that calls the API > shouldn't be checking for it. Conversely, if the API does not accept > NULL then its implementation shouldn't be checking for it. So I would > suggest we remove the redundancy by eliminating the NULL check from > either the callers or the API implementation. Thoughts? I'll provide > patches if people agree and decide whether or not these APIs should > support NULL. In general the API accepts NULL (e.g. dlib), so please remove the redundant checks outside it. -- Cheers Jorge.-

May 3, 2009

Style question: cleaning up NULL

by onepoint＠starurchin.org

I've noticed that a_Url_free() and a_Image_unref() both guard against being passed a NULL argument, yet in several cases the code that calls these functions also checks for a NULL, which is redundant. I think that if the API officially accepts NULL then code that calls the API shouldn't be checking for it. Conversely, if the API does not accept NULL then its implementation shouldn't be checking for it. So I would suggest we remove the redundancy by eliminating the NULL check from either the callers or the API implementation. Thoughts? I'll provide patches if people agree and decide whether or not these APIs should support NULL. Regards, Jeremy Henty

May 3, 2009