#staff | Logs for 2019-10-15
« return
[08:09:41] <FatPhil> I can't get the "a patent" link to work in https://soylentnews.org
[08:10:14] <FatPhil> looks like the code has stripped all the & out of the (invalid, because not entity encoded) href.
[08:10:38] <FatPhil> When I edit it to have & everywhere, the same thing seems to happen
[08:11:00] <FatPhil> it works in preview, just not when I save my edits
[08:32:57] <chromas> it looks like the uspto site is just a pile of balls
[08:33:39] <janrinok> I can see it fine - what do you mean getting the link to work exactly?
[08:34:20] <chromas> it loads for me but there should be a 'permalink' for it instead of requiring 376 parameters
[08:34:49] <janrinok> true - and I cannot get it to display any of the 14 pages of images.
[08:35:04] <janrinok> ... but that looks like it is just a crap site.
[08:35:35] <chromas> I don't see any images
[08:35:43] <chromas> The link that says "images" takes me to a one-page pdf
[08:38:47] <janrinok> as, for me it opens page 1 of 14 but with nothing of interest displayed. The header is there, the margin links are there, but no images.
[08:39:27] <chromas> Maybe that's why google made its own patent search thinger
[08:40:28] <chromas> I see. There's a sidebar to go to different sections, and each section is its own pdf
[08:41:22] <janrinok> that site is an example of how not to write a usable page
[08:41:35] <chromas> at least it doesn't have flash-based navigation
[08:48:46] <Bytram> coffee++
[08:48:46] <Bender> karma - coffee: 8
[08:49:19] <janrinok> chromas, ....yet
[08:49:32] <janrinok> ~gday Bytram !!!
[08:49:47] <janrinok> oooh, bots not working...
[08:51:25] <Bytram> ~gday janrinok
[08:52:01] <janrinok> have you been called in yet today - or are you hiding?
[08:52:15] <Bytram> have the late shift
[08:52:16] <chromas> exec died earlier
[08:52:22] <Bytram> RIP
[08:54:42] <Bytram> FatPhil: Did you read my report on the missing journal entry? I already ran into that problem and reported that using "&" in a URL is problematic; quick workaround is to just use "&" in URLs, or better, percent-encode them.
[08:55:53] <Bytram> The percent encoding of ampersand would be %26
[08:56:18] <Bytram> I haven't tried that one, personally, but that is what came up in the discussion
[09:25:31] <FatPhil> But it breaks when you *don't* use & too - that's how I found the problem in the first place.
[09:27:28] <FatPhil> oooh, now that's weird - my edits did indeed fix the problem, I was just being shown a cached version of the page after my edit, which could also be considered a bug.
[09:29:29] <Bytram> FatPhil: Nope, that is SOP... takes a few minutes for changes on the site to make it to the (caching) front-ends
[09:29:31] <FatPhil> The problem's in the original submission, so may have been caused by the bot (upstart) repeating the valid &'s from the source article.
[09:30:04] <FatPhil> It's still a bug even if it's SOP.
[09:31:02] <Bytram> Are you proposing that we bounce the frontends every time we poswt a new story or update an existing one?
[09:31:35] <FatPhil> An update to a story should redirect to a URL which forces the freshest page to be rendered.
[09:32:25] <Bytram> In theory, the preview that you see on the editing page should match what is generated by the site; and it generally does. ...
[09:32:46] <Bytram> so the existing URL to the story is now invalid?
[09:33:27] <FatPhil> Rando readers don't need to see the updates in real time, but the person who updated the story bloody well ought to.
[09:33:35] <Bytram> now, sending a message of some kind to the front ends to update their copy of a story... that could make sense, but let's not go around changing URLs for each story update.
[09:33:52] <Bytram> just look at the preview that is on the editing page
[09:34:24] <FatPhil> i.e. just trust that it's done the right thing on update?
[09:34:29] <Bytram> yeop
[09:34:47] <FatPhil> If "trust" is SOP, then why even bother previewing? trust yourself too!
[09:35:34] <Bytram> I struggled with that concept for a while... but once I accepted that the front-ends only check for updated "stuff" periodically, I realized I just needed a little patience and it all seems to work out.
[09:36:36] <Bytram> What do you mean? Trust the preview! If that ain't what you want, fix it. If the preview *is* what you want, then save/update and wait for it to make it's way through.
[09:36:45] <Bytram> s/'//
[09:41:39] <FatPhil> OK, now I know that I must trust the preview and only the preview, that will be my SOP (I always preview anyway, even for single-whitespace changes, so it's no change of procedure, only of where trust is placed)
[09:42:03] -!- audioguy [audioguy!~audioguy@Soylent/Staff/Developer/audioguy] has joined #staff
[09:42:03] -!- mode/#staff [+v audioguy] by Ares
[09:52:17] <chromas> The preview isn't always right, though
[09:52:36] <chromas> For example, a poo glyph works in the dept field but becomes a "?" on save
[09:55:54] <chromas> Never seen &s be a problem. URLs should only be in a tags anyhow
[09:56:10] * chromas looks into upstart's & stripping
[10:01:58] <chromas> hm, they actually put & in their url. well it's not like the verge knows much about computers. remember their how-to-buld-a-pc video?
[10:03:43] <Bytram> chromas: Yep, "&" should only be in tags... be sure to tell that to "AT&T"
[10:03:57] <Bytram> ;)
[10:07:19] <chromas> Aha! The bot submits the urls as-are, so it's all rehash's fault and I don't have to do nuffin
[10:07:41] <chromas> though it does appear waterfox is smart enough to convert them to & when clicking