aboutsummaryrefslogtreecommitdiffstats
path: root/README.xwidget
diff options
context:
space:
mode:
authorJoakim Verona2011-10-28 16:26:43 +0200
committerJoakim Verona2011-10-28 16:26:43 +0200
commitfc7ee8f78653ddcd6a46bf1ac1b50b80d939c52e (patch)
treecb39f16cf2f391669cb6341f5b3d444a53b13dc2 /README.xwidget
parentb70516db12801e015c538c4e7eb44cb87316afc4 (diff)
downloademacs-fc7ee8f78653ddcd6a46bf1ac1b50b80d939c52e.tar.gz
emacs-fc7ee8f78653ddcd6a46bf1ac1b50b80d939c52e.zip
some fruitless webkit size experiments
Diffstat (limited to 'README.xwidget')
-rw-r--r--README.xwidget176
1 files changed, 93 insertions, 83 deletions
diff --git a/README.xwidget b/README.xwidget
index 6b735cae13a..e9b8817b17b 100644
--- a/README.xwidget
+++ b/README.xwidget
@@ -264,7 +264,8 @@ Currently I just allocate as large an area as the offscreen widget
264desires. This works well most of the time. But a HTML page might in 264desires. This works well most of the time. But a HTML page might in
265principle be of infinite height so there will probably be cases where 265principle be of infinite height so there will probably be cases where
266this doesn't work too well. 266this doesn't work too well.
267** TODO again a trace 267** DONE again a trace
268 CLOSED: [2011-10-28 Fri 13:48]
268[2011-08-23 Tue] 269[2011-08-23 Tue]
269the hunch is that since I still hand-wave the view storage the array 270the hunch is that since I still hand-wave the view storage the array
270can get out of synchronous. so maybe switching to a lisp structure 271can get out of synchronous. so maybe switching to a lisp structure
@@ -533,6 +534,9 @@ xdotool
533xte xautomation 534xte xautomation
534crikey 535crikey
535libxdo 536libxdo
537
538*** TODO webkit raw keyboard event escape
539c-c tab could send a raw tab to the webkit instance.
536** DONE remove the special-case for when the minibuffer is 540** DONE remove the special-case for when the minibuffer is
537 active. I added some code to reduce the annoying problem display artefacts 541 active. I added some code to reduce the annoying problem display artefacts
538 when making the minibuffer the selected window. This made xwidgets in the 542 when making the minibuffer the selected window. This made xwidgets in the
@@ -1203,6 +1207,10 @@ to see the problem surface to http://www.slashdot.org
1203 1207
1204and then compare by resizing in Epiphany, which is also webkit based. 1208and then compare by resizing in Epiphany, which is also webkit based.
1205 1209
1210**** TODO try putting webkit osr inside a scrolling window
1211it seems webkit is supposed to behave differently while embedded in a
1212scrolling window. This is a bit cumbersome because the container stack
1213is already deep.
1206*** TODO xwidget webkit allow loading from string from emacs 1214*** TODO xwidget webkit allow loading from string from emacs
1207*** DONE xwidget-webkit-last-session 1215*** DONE xwidget-webkit-last-session
1208 CLOSED: [2011-08-01 Mon 22:38] 1216 CLOSED: [2011-08-01 Mon 22:38]
@@ -1230,6 +1238,90 @@ that image-mode has special code to handle scrolling. the browser mode
1230and image mode has some similarities. 1238and image mode has some similarities.
1231 1239
1232I made some delegation code frrom webkit mode to image mode. 1240I made some delegation code frrom webkit mode to image mode.
1241*** TODO internal links (page.html#section) do not work
1242see xwidget-webkit-show-named-element
1243
1244*** TODO sindicat notes
1245Here are some comments from user "sindikat" and my replies
1246
1247- site.com and http://site.com should be equivalent (simple site.com
1248 throws error)
1249
1250Yes, but its unclear at what level in Emacs to do this
1251properly. I added a url-tidy function as a start.
1252
1253- http://ya.ru renders inadequatly (compare with any other browser) -
1254 the search text-input is way below
1255
1256The problem is the size communication between Emacs and Webkit.
1257
1258- doing PageDown is endless; so if you do 100 PageDowns, you have to
1259 do 100 PageUps to retun to the header of the page
1260
1261True, I hadn't noticed. Thanks.
1262
1263- http://linux.org.ru (just an example) renders incorrectly too - it
1264 should stretch horizontally
1265
1266Size communication.
1267
1268- obviously, pointing of mouse over some link should change it to
1269 pointing hand cursor
1270
1271Need to verify with some other webkit browser.
1272
1273- when you are somewhere on the middle of a long page, than go to some
1274 other page, you are still in the middle, instead of being again on
1275 the top
1276
1277This is because I inherit from Image view mode. I kind of like it so
1278we can add an option for it.
1279
1280
1281- changing dropdown menus cause flickering
1282
1283
1284- string entering is incorrect - by default it enters the title of the
1285 page, while it should be empty
1286
1287The cause is the lack of return value in the webkit evaluation
1288API. Ive made some fixes.
1289
1290- internal links (page.html#section) do not work
1291
1292 ive added a rudimentary function "xwidget-webkit-show-named-element" for this
1293
1294- maybe it's a good idea to implement Conkeror or some other
1295 keybindings, where you press 'f' then select the exact <input
1296 type="text"> where you want to enter text, without using mouse,
1297 etc.;
1298
1299Indeed, this would require better DOM integration.
1300
1301- pressing 'home' and 'end' puts nonsense into minibuffer
1302
1303Probably because the Image mode derivative is mostly a hack.
1304fixed now I think.
1305
1306
1307
1308
1309
1310
1311- implement search (emacs internal isearch obviously doesn't work)
1312
1313Either use the webkit search but that doesn't feel right. It would be
1314better to expose the DOM and search that.
1315
1316- some sites intercept with keyboard; example -
1317 http://www.artlebedev.ru/kovodstvo/business-lynch/2011/10/03/ uses
1318 Ctrl+left/right/up/down to navigate between pages - this should be
1319 implemented too
1320
1321Keyboard integration is the unloved step-child of xwidgets, unfortunately.
1322
1323
1324
1233** TODO xwidget image display spec compatibility 1325** TODO xwidget image display spec compatibility
1234some history: the first version of the xwidget display spec was 1326some history: the first version of the xwidget display spec was
1235the same as an image spec. This turned out not to be fantastic because 1327the same as an image spec. This turned out not to be fantastic because
@@ -1369,85 +1461,3 @@ splint -Demacs -DHAVE_CONFIG_H -I. -I/home/joakim/build_myprojs/emacsnew/emac
1369user reports that xwidgets segfaults on the 32 bit Mint distribution 1461user reports that xwidgets segfaults on the 32 bit Mint distribution
1370but not the 64 bit. Mint is an Ubuntu derivative. I got some 1462but not the 64 bit. Mint is an Ubuntu derivative. I got some
1371VirtualBox images to test with. 1463VirtualBox images to test with.
1372** TODO sindicat notes
1373Here are some comments from user "sindikat" and my replies
1374
1375- site.com and http://site.com should be equivalent (simple site.com
1376 throws error)
1377
1378Yes, but its unclear at what level in Emacs to do this
1379properly. I added a url-tidy function as a start.
1380
1381- http://ya.ru renders inadequatly (compare with any other browser) -
1382 the search text-input is way below
1383
1384The problem is the size communication between Emacs and Webkit.
1385
1386- doing PageDown is endless; so if you do 100 PageDowns, you have to
1387 do 100 PageUps to retun to the header of the page
1388
1389True, I hadn't noticed. Thanks.
1390
1391- http://linux.org.ru (just an example) renders incorrectly too - it
1392 should stretch horizontally
1393
1394Size communication.
1395
1396- obviously, pointing of mouse over some link should change it to
1397 pointing hand cursor
1398
1399Need to verify with some other webkit browser.
1400
1401- when you are somewhere on the middle of a long page, than go to some
1402 other page, you are still in the middle, instead of being again on
1403 the top
1404
1405This is because I inherit from Image view mode. I kind of like it so
1406we can add an option for it.
1407
1408
1409- changing dropdown menus cause flickering
1410
1411
1412- string entering is incorrect - by default it enters the title of the
1413 page, while it should be empty
1414
1415The cause is the lack of return value in the webkit evaluation
1416API. Ive made some fixes.
1417
1418- internal links (page.html#section) do not work
1419
1420 ive added a rudimentary function "xwidget-webkit-show-named-element" for this
1421
1422- maybe it's a good idea to implement Conkeror or some other
1423 keybindings, where you press 'f' then select the exact <input
1424 type="text"> where you want to enter text, without using mouse,
1425 etc.;
1426
1427Indeed, this would require better DOM integration.
1428
1429- pressing 'home' and 'end' puts nonsense into minibuffer
1430
1431Probably because the Image mode derivative is mostly a hack.
1432fixed now I think.
1433
1434
1435
1436
1437
1438
1439- implement search (emacs internal isearch obviously doesn't work)
1440
1441Either use the webkit search but that doesn't feel right. It would be
1442better to expose the DOM and search that.
1443
1444- some sites intercept with keyboard; example -
1445 http://www.artlebedev.ru/kovodstvo/business-lynch/2011/10/03/ uses
1446 Ctrl+left/right/up/down to navigate between pages - this should be
1447 implemented too
1448
1449Keyboard integration is the unloved step-child of xwidgets, unfortunately.
1450
1451
1452* TODO - internal links (page.html#section) do not work
1453see xwidget-webkit-show-named-element