Re: [elinks-users] Cache problems?
On Fri, Jul 14, 2006 at 01:41:38PM -0600, Reid Rivenburgh wrote: On Fri, Jul 14, 2006 at 06:39:23AM +, Miciah Dashiel Butler Masters wrote: On Thu, Jul 13, 2006 at 03:54:28PM -0600, Reid Rivenburgh wrote: Hi. I've been using elinks for awhile now; it seems like the best of the text browsers. But I've noticed one problem, and I haven't seen any bug reports or emails relating to it. I seem to be having cache problems. When I visit www.dailykos.com (for example) and click to view an article, if I then click on www.dailykos.com on the article page to go to the home page again (i.e. NOT going back in the history), the page I get seems to be the cached version from before. It's very fast and doesn't seem to access the network. The content on the home page changes frequently, so it's often out of date, requiring me to reload manually. I've set the formatted documents number and memory cache size to zero to try to fix this, but it still happens. Is it possible that this only happens to me for some reason? I'm currently using the latest 0.12 from 20060712, but it happened with 0.11 and 0.10. Go to Setup - Options manager - Document - Cache - Ignore cache-control info from server and set that option to 0. I forgot about that option. That's already set to 0. When you say 'already', do you mean that you have the problem even with the option disabled? -- Miciah Masters [EMAIL PROTECTED] / [EMAIL PROTECTED] ___ elinks-users mailing list elinks-users@linuxfromscratch.org http://linuxfromscratch.org/mailman/listinfo/elinks-users
Re: [elinks-users] Cache problems?
On Wed, Jul 19, 2006 at 05:32:19PM +, Miciah Dashiel Butler Masters wrote: On Fri, Jul 14, 2006 at 01:41:38PM -0600, Reid Rivenburgh wrote: On Fri, Jul 14, 2006 at 06:39:23AM +, Miciah Dashiel Butler Masters wrote: Go to Setup - Options manager - Document - Cache - Ignore cache-control info from server and set that option to 0. I forgot about that option. That's already set to 0. When you say 'already', do you mean that you have the problem even with the option disabled? Yes, that's what I meant. Sorry about the confusion. Reid ___ elinks-users mailing list elinks-users@linuxfromscratch.org http://linuxfromscratch.org/mailman/listinfo/elinks-users
Re: [elinks-users] Cache problems?
On Wed, Jul 19, 2006 at 06:05:35PM +, Miciah Dashiel Butler Masters wrote: Looking at the document, I don't see any headers to indicate that ELinks should reload it: HTTP/1.1 200 OK Date: Wed, 19 Jul 2006 17:42:19 GMT Server: Apache/1.3.34 (Unix) mod_gzip/1.3.26.1a Vary: Accept-Encoding Last-Modified: Wed, 19 Jul 2006 17:42:01 GMT ETag: 8d4622-c8b2-44be6ee9 Accept-Ranges: bytes Connection: close Content-Type: text/html Content-Encoding: gzip Content-Length: 18374 Do the browsers with which you are familiar actually check every time when you view a document (presumably using the HTTP If-Modified-Since header) whether there is a newer copy on the server? ELinks could do that, but it would be a little complex, and far too slow. I can't even stand the behaviour with ignore_cache_control disabled, which only affects documents that explicitely signal that they should be reloaded from the server. Such behaviour might be acceptable if done in the background, but then it would be a bit confusing (you load the document, you start to read it, then it suddenly updates while you're in the middling of reading it). My main browser is Firefox; I use ELinks under special circumstances. Unfortunately, I'm really just a user and don't know how exactly Firefox knows to reload a page. It does seem like there's some network activity, so it probably is checking if there's a newer copy on the server. If that's the case, I'm not sure why you think it would be so slow. I find Firefox to be pretty fast, so ELinks should be at least as fast if it was doing the same server query, no? Overall, it seems to me like ELinks is a very fast browser, which should be no surprise since it's just text. When loading a page, I personally would definitely trade a split-second of time for the proper page. (I wouldn't want it to work as you describe, showing the old page and then updating it in the background. I'd prefer to wait.) Maybe this could be a setting for those that prefer speed and don't ever use ELinks to visit frequently-changing pages? (Notice I'm avoiding the whole issue of the complexity of implementing this...!) Reid ___ elinks-users mailing list elinks-users@linuxfromscratch.org http://linuxfromscratch.org/mailman/listinfo/elinks-users
Re: [elinks-users] Cache problems?
On Wed, Jul 19, 2006 at 12:25:16PM -0600, Reid Rivenburgh wrote: On Wed, Jul 19, 2006 at 06:05:35PM +, Miciah Dashiel Butler Masters wrote: [...] Do the browsers with which you are familiar actually check every time when you view a document (presumably using the HTTP If-Modified-Since header) whether there is a newer copy on the server? ELinks could do that, but it would be a little complex, and far too slow. I can't even stand the behaviour with ignore_cache_control disabled, which only affects documents that explicitely signal that they should be reloaded from the server. Such behaviour might be acceptable if done in the background, but then it would be a bit confusing (you load the document, you start to read it, then it suddenly updates while you're in the middling of reading it). My main browser is Firefox; I use ELinks under special circumstances. Unfortunately, I'm really just a user and don't know how exactly Firefox knows to reload a page. It does seem like there's some network activity, so it probably is checking if there's a newer copy on the server. If that's the case, I'm not sure why you think it would be so slow. I find Firefox to be pretty fast, so ELinks should be at least as fast if it was doing the same server query, no? Overall, it seems to me like ELinks is a very fast browser, which should be no surprise since it's just text. When loading a page, I personally would definitely trade a split-second of time for the proper page. (I wouldn't want it to work as you describe, showing the old page and then updating it in the background. I'd prefer to wait.) Maybe this could be a setting for those that prefer speed and don't ever use ELinks to visit frequently-changing pages? (Notice I'm avoiding the whole issue of the complexity of implementing this...!) The split-second delay would be mildly annoying, but that is the best case, when the server is responsive. How is the performance with slow servers? In any case, I do not consider Firefox to be anywere near 'pretty fast'. Doing what you describe might be feasible. It would be a little more code for the HTTP part and a little re-engineering of the cache code, but it doesn't sound too difficult. Whether it will actually get done also depends on whether the developers desire such behaviour, however. -- Miciah Masters [EMAIL PROTECTED] / [EMAIL PROTECTED] ___ elinks-users mailing list elinks-users@linuxfromscratch.org http://linuxfromscratch.org/mailman/listinfo/elinks-users
Re: [elinks-users] Cache problems?
On Wed, Jul 19, 2006 at 06:33:07PM +, Miciah Dashiel Butler Masters wrote: On Wed, Jul 19, 2006 at 12:25:16PM -0600, Reid Rivenburgh wrote: My main browser is Firefox; I use ELinks under special circumstances. Unfortunately, I'm really just a user and don't know how exactly Firefox knows to reload a page. It does seem like there's some network activity, so it probably is checking if there's a newer copy on the server. If that's the case, I'm not sure why you think it would be so slow. I find Firefox to be pretty fast, so ELinks should be at least as fast if it was doing the same server query, no? Overall, it seems to me like ELinks is a very fast browser, which should be no surprise since it's just text. When loading a page, I personally would definitely trade a split-second of time for the proper page. (I wouldn't want it to work as you describe, showing the old page and then updating it in the background. I'd prefer to wait.) Maybe this could be a setting for those that prefer speed and don't ever use ELinks to visit frequently-changing pages? (Notice I'm avoiding the whole issue of the complexity of implementing this...!) The split-second delay would be mildly annoying, but that is the best case, when the server is responsive. How is the performance with slow servers? In any case, I do not consider Firefox to be anywere near 'pretty fast'. Well, it's all relative, I suppose. I'd like everything to be instantaneous, but I find the slight delays acceptable. It may be the case that there are times when the network or server are slow, making loading of a cached page slow, and I just accept it as the way it is. That seems rare, though. From my perspective, it's also important to not show old, out-of-date pages. Another example is a site (www.huffingtonpost.com) that links from the main page to a second page with embedded named anchors; if the second page has been cached but I've reloaded the main page, the named anchor won't exist in the cached version of the second page, so I'll get an error dialog and have to reload the second page, go back to the main page, and re-click the link. Ugh. Doing what you describe might be feasible. It would be a little more code for the HTTP part and a little re-engineering of the cache code, but it doesn't sound too difficult. Whether it will actually get done also depends on whether the developers desire such behaviour, however. Sure, I understand that feature requests and fixes get done at the whim of the developers. I was just trying to raise the issue to the forefront. Frankly, for people who use ELinks to browse many web sites, the current state of things seems far from ideal. If ELinks is supposed to be used more to browse static documentation pages and things like that, I can see why it hasn't been noticed or deemed important. Reid ___ elinks-users mailing list elinks-users@linuxfromscratch.org http://linuxfromscratch.org/mailman/listinfo/elinks-users
Re: [elinks-users] Cache problems?
On Wed, Jul 19, 2006 at 02:05:35PM EDT, Miciah Dashiel Butler Masters wrote: [..] Do the browsers with which you are familiar actually check every time when you view a document (presumably using the HTTP If-Modified-Since header) whether there is a newer copy on the server? In mozilla it's customizable. Under Preferences Advanced Cache, you have the following options: Compare the page in the cache to the page on the network: . Every time I view the page . When the page is out of date . Once per session . Never I assume this to really mean.. :-) 'obtain and display' the page from the server in lieu of the one in the cache.. under the above circumstances.. rather than 'compare' which is not the same thing. Sorry for being splitting hairs. ELinks could do that, but it would be a little complex, and far too slow. Not really. I hit Ctrl-R all the time when accessing news sites where the contents change frequently and on a decent broadband connection *and* provided the server is responsive.. this does not result in any noticeable delay. IOW the reloading does not feel like it adds anything to the Ctrl-R action. OTOH doing the same thing in mozilla is quite slow on my machine.. not because of the time it takes to fetch a new version but appently because of the rendering (PIII 650MHz). This is one of the main reasons I have switched to elinks for pratcticall all web browsing.. couldn't put up with mozilla executing its million lines of code every time I displayed a new page. I can't even stand the behaviour with ignore_cache_control disabled, which only affects documents that explicitely signal that they should be reloaded from the server. I'm not sure what this is supposed to do. I set this flag to '0' as suggested in an earlier post and even though I cycled elinks I have not noticed a change in behavior. Such behaviour might be acceptable if done in the background, but then it would be a bit confusing (you load the document, you start to read it, then it suddenly updates while you're in the middling of reading it). The main problem that I have - even with pages whose contents do not change frequently such as an online docs/manuals is that the links on the page do not switch to their 'visited' color.. which sometimes makes it a bit confusing.. but no big deal, since I have gotten used to Ctrl-R'ing all the time without thinking. Since this mainly poses problem when hitting the back key - 'h' - maybe I should find a way to map this to a short script/macro that just does 'back'+'reload'? Hope the above is relevant. Thanks cga ___ elinks-users mailing list elinks-users@linuxfromscratch.org http://linuxfromscratch.org/mailman/listinfo/elinks-users
Re: [elinks-users] Cache problems?
On Wed, Jul 19, 2006 at 04:36:29PM -0400, cga2000 wrote: OTOH doing the same thing in mozilla is quite slow on my machine.. not because of the time it takes to fetch a new version but appently because of the rendering (PIII 650MHz). This is one of the main reasons I have switched to elinks for pratcticall all web browsing.. couldn't put up with mozilla executing its million lines of code every time I displayed a new page. I understand your grief. The very latest Firefox alpha or beta seems quite a bit better in that regard; you might want to give it a try. Since this mainly poses problem when hitting the back key - 'h' - maybe I should find a way to map this to a short script/macro that just does 'back'+'reload'? As someone mentioned earlier, you can use the keybinding manager to associate back with Follow the current link, forcing reload of the target. I think that's what you want. Unfortunately, I use the mouse to navigate pages, and I don't think there's any way at the moment to configure actions called by mouse clicks. reid ___ elinks-users mailing list elinks-users@linuxfromscratch.org http://linuxfromscratch.org/mailman/listinfo/elinks-users
Re: [elinks-users] Cache problems?
On Wed, Jul 19, 2006 at 04:43:30PM EDT, Reid Rivenburgh wrote: On Wed, Jul 19, 2006 at 04:36:29PM -0400, cga2000 wrote: OTOH doing the same thing in mozilla is quite slow on my machine.. not because of the time it takes to fetch a new version but appently because of the rendering (PIII 650MHz). This is one of the main reasons I have switched to elinks for pratcticall all web browsing.. couldn't put up with mozilla executing its million lines of code every time I displayed a new page. I understand your grief. :-) The very latest Firefox alpha or beta seems quite a bit better in that regard; you might want to give it a try. I've also heard that the SpiderMonkey - the new mozilla has marked improvements in the area. But slow rendering was only one of the issues. I have switched to a keyboard-only text-mode 'desktop' - interface, I should say and I am a much happier man. Since this mainly poses problem when hitting the back key - 'h' - maybe I should find a way to map this to a short script/macro that just does 'back'+'reload'? As someone mentioned earlier, you can use the keybinding manager to associate back with Follow the current link, forcing reload of the target. I think that's what you want. Unfortunately, I use the mouse to navigate pages, and I don't think there's any way at the moment to configure actions called by mouse clicks. Drop the mouse man..! Charles Darwin is categorical.. You'll never grow a third arm in your lifetime. Oh.. and as to developers being 'whimsical' relative to enhancements, I think not.. That's exactly why I barged into your thread.. The list is really the place where you can voice an opinion.. and since this caching behavior was something that struck me as being fundamentally different from the other browser.. Not sure it's a bad thing, though.. I briefly mentioned it but Miciah has a good point.. when the server is unresponsive it's OK I guess if you explicitly ask for a reload.. Maybe less so when the reload is done on a transparent basis every time you hit the back button (eg.). Thanks cga ___ elinks-users mailing list elinks-users@linuxfromscratch.org http://linuxfromscratch.org/mailman/listinfo/elinks-users
Re: [elinks-users] Cache problems?
On Wed, Jul 19, 2006 at 05:31:42PM EDT, Reid Rivenburgh wrote: On Wed, Jul 19, 2006 at 05:10:42PM -0400, cga2000 wrote: I've also heard that the SpiderMonkey - the new mozilla has marked improvements in the area. But slow rendering was only one of the issues. I have switched to a keyboard-only text-mode 'desktop' - interface, I should say and I am a much happier man. Nice. I don't think I could get by with text-only in my main browser FYI, I think there's a Firefox extension that is aimed at keyboard-only use. Never used it, though. I'll look it up.. Although I'm sceptical.. Microsoft did a very good job of making gui's navigable via the keyboard but you cannot change the fact that the gui designer was thinking *mouse* when he designed the layout.. etc. Just and example.. in many guis I have seen you may have to drill down three.. four.. five..? levels of dialog/popup boxes the size of a postage stamp (some of them you'll even have to reach for that scrollbar so you can get to what you're looking for..) before you eventually get to the particular flag or whatever that you want to change.. While this is perfectly suitable with the inherent slowness of the mouse, it becomes rather annnoying when you use a considerably faster interface like the keyboard. Drop the mouse man..! Charles Darwin is categorical.. You'll never grow a third arm in your lifetime. Oh it's no problem, I use my tail to control the mouse! :) :-) Not that I'm particularly happy with the keyboard either. The traditional typewriter keyboard was bad enough but what IBM did to it is beyond words.. Only solace I find in a keyboard-only environment is its relative simplicity (as compared with keyboard+mouse) .. Even a fairly large number of inconsistant keyboard shortcuts across applications is in my case a tad more efficient than switching action mode every other second. Oh.. and as to developers being 'whimsical' relative to enhancements, I think not.. That's exactly why I barged into your thread.. The list is really the place where you can voice an opinion.. and since this caching behavior was something that struck me as being fundamentally different from the other browser.. Not sure it's a bad thing, though.. I briefly mentioned it but Miciah has a good point.. when the server is unresponsive it's OK I guess if you explicitly ask for a reload.. Maybe less so when the reload is done on a transparent basis every time you hit the back button (eg.). I guess whims was a bad word to use. so far they haven't objected... I just wanted to explain that I understand it's open source and that the developers can do what they like. I don't think they may.. or can. Whatever they may state in their disclaimers they just cannot pull the plug on a piece of software that has some degree of following... And in fact they very rarely do so.. one of the nice things about OSS is that when it happens someone else will step in.. if the software is worth the effort.. that is.. (?) As to *being able* to do what they like.. it's pretty much the same thing.. too much change that turns out to be unsatisfactory and the user community will just not take it.. someone forks the project.. etc. .. like .. I'm a vim user/fan.. now Bram Moolenaar et al. all of a sudden see the light and the next version of Vim turns out to be an MS Word clone... how 'bout that? I do hope someone will implement/improve CSS rendering some time soon, though.. Not that I spend much time on commercial web sites.. but a growing number are getting hard to read. Yeah.. that's how you eventually get them to pay attention.. .. moan consistently on the list with a bunch of other troublemakers until they really get sick of it.. OK.. fixed in CVS - GIT, I mean.. now p*ss off.. :-) Sure, having a discussion on this list is one way to convince them that your way is the right way! (So here goes:) As for the cache issue, I think the current method is not good from a usability standpoint. It sounds like he and I use ELinks in different ways so he doesn't run into this problem often, plus he places a lot of value on speed. I think if Mozilla or Firefox worked the same way, the majority of users would be up in arms! .. interesting point.. but then they have been conditioned by years of IE ( Netscape) abuse in so many ways.. So that may or may not be an argument.. I found this behavior a bit disconcerting at first but after a couple of weeks I've grown used to it. As mentioned in an earlier post, and provided I fully understand the issue, this would appear to be something that can be customized in mozilla.. don't know about FF, though. So if I was in charge of ELinks development, I'd make the Mozilla compare automatically (or whatever it is) be the default but put in Mozilla's other settings as options. Then everyone could be happy. .. and call the option something else rather than
Re: [elinks-users] Cache problems?
On Wed, Jul 19, 2006 at 09:44:41PM -0400, cga2000 wrote: On Wed, Jul 19, 2006 at 05:31:42PM EDT, Reid Rivenburgh wrote: On Wed, Jul 19, 2006 at 05:10:42PM -0400, cga2000 wrote: I've also heard that the SpiderMonkey - the new mozilla has marked improvements in the area. But slow rendering was only one of the issues. I have switched to a keyboard-only text-mode 'desktop' - interface, I should say and I am a much happier man. Nice. I don't think I could get by with text-only in my main browser FYI, I think there's a Firefox extension that is aimed at keyboard-only use. Never used it, though. I'll look it up.. Although I'm sceptical.. Microsoft did a very good job of making gui's navigable via the keyboard but you cannot change the fact that the gui designer was thinking *mouse* when he designed the layout.. etc. Just and example.. in many guis I have seen you may have to drill down three.. four.. five..? levels of dialog/popup boxes the size of a postage stamp (some of them you'll even have to reach for that scrollbar so you can get to what you're looking for..) before you eventually get to the particular flag or whatever that you want to change.. While this is perfectly suitable with the inherent slowness of the mouse, it becomes rather annnoying when you use a considerably faster interface like the keyboard. This is why why have keyboard accelerators. Drop the mouse man..! Charles Darwin is categorical.. You'll never grow a third arm in your lifetime. Oh it's no problem, I use my tail to control the mouse! :) :-) Not that I'm particularly happy with the keyboard either. The traditional typewriter keyboard was bad enough but what IBM did to it is beyond words.. What did IBM do? Only solace I find in a keyboard-only environment is its relative simplicity (as compared with keyboard+mouse) .. Even a fairly large number of inconsistant keyboard shortcuts across applications is in my case a tad more efficient than switching action mode every other second. I prefer to be able to form my environment to myself, you know, like a human. Using GNU Screen, my Web browser, E-mail, IRC client, shells, editors, music player, c. are each exactly one keypress away. Each program is customised (mostly) with comfortable (for me) keybindings. Oh.. and as to developers being 'whimsical' relative to enhancements, I think not.. That's exactly why I barged into your thread.. The list is really the place where you can voice an opinion.. and since this caching behavior was something that struck me as being fundamentally different from the other browser.. Not sure it's a bad thing, though.. I briefly mentioned it but Miciah has a good point.. when the server is unresponsive it's OK I guess if you explicitly ask for a reload.. Maybe less so when the reload is done on a transparent basis every time you hit the back button (eg.). I guess whims was a bad word to use. so far they haven't objected... How offensive, implying that we have whims! In fact, developers have itches and occasional fancies, but none of this whim nonsense. I just wanted to explain that I understand it's open source and that the developers can do what they like. I don't think they may.. or can. Whatever they may state in their disclaimers they just cannot pull the plug on a piece of software that has some degree of following... And in fact they very rarely do so.. one of the nice things about OSS is that when it happens someone else will step in.. if the software is worth the effort.. that is.. (?) Guess how ELinks got started? See http://elinks.cz/history.html. The impetus for ELinks was largly that it looked like Links had been abandoned. As to *being able* to do what they like.. it's pretty much the same thing.. too much change that turns out to be unsatisfactory and the user community will just not take it.. someone forks the project.. etc. ELinks was forked also because Links's developer was very strict with regard to patches. .. like .. I'm a vim user/fan.. now Bram Moolenaar et al. all of a sudden see the light and the next version of Vim turns out to be an MS Word clone... how 'bout that? It is getting there, but at least Vim still runs on the terminal. I do hope someone will implement/improve CSS rendering some time soon, though.. Not that I spend much time on commercial web sites.. but a growing number are getting hard to read. The developers hope for that too. However, most of them work, most of them are students, and better CSS support requires DOM support and a box model and stuff that all requires a lot of coding effort. We would all like to get this done, but don't get your hopes too high. Yeah.. that's how you eventually get them to pay attention.. .. moan consistently on the list with a bunch of other troublemakers until they really get sick of it.. OK.. fixed in CVS - GIT, I mean..