encoding - Why is "​" being injected into my HTML? -


edit: can see issue here (look in source).

edit2: interesting, not issue in source. console (firebug well).

i have following markup in file called test.html:

​<!doctype html> <html> <head>     <title>test harness</title>     <link href='/css/main.css' rel='stylesheet' type='text/css' /> </head> <body>     <h3>test harness</h3> </body> </html>

but in chrome, see:

<!doctype html> <html> <head> </head> <body>     "&#8203;           "     <title>test harness</title>     <link href='/css/main.css' rel='stylesheet' type='text/css' />     <h3>test harness</h3> </body> </html>

it looks &#802 0 width space, causing it? using sublime text 2 utf-8 encoding , google app engine jinja2 (but jinja loading test.html). thoughts?

thanks in advance.

it issue in source. live example provided starts following bytes (i.e., appear before <!doctype html>): 0xe2 0x80 0x8b. can seen e.g. using rex swain’s http viewer selecting “hex” under “display format”. note validating page w3c markup validator gives information suggests there wrong @ start of document, message “line 1, column 1: non-space characters found without seeing doctype first.”

what happens in validator , in chrome tools – e.g. in firebug – bytes 0xe2 0x80 0x8b taken character data, implicitly starts body element (since character data cannot validly appear in head element or before it), implying empty head element before it.

the solution, of course, remove bytes. browsers ignore them, should not rely on such error handling, , bytes prevent useful html validation. how remove them, , how got there in first place, depends on authoring environment.

since page declared (in http headers) being utf-8 encoded, bytes represent zero width space (u+200b) character. has no visible glyph , no width, won’t notice in visual presentation though browsers treat being data @ start of body element. notation &#8203; character reference it, presumably used browser tools indicate presence of invisible character.

it possible software produced html document meant insert zero width no-break space (u+feff) instead. have been valid, since special convention, utf-8 encoded data may start character, known byte order mark (bom) when appearing @ start of data. using u+200b instead of u+feff sounds error software unlikely make, human beings may mistaken way if think of unicode names of characters.


Comments

Post a Comment

Popular posts from this blog

resizing Telegram inline keyboard -

telegram's inline keyboard great feature lots of different use cases. inline buttons added list of items this: inline_keyboard = [[inlinekeyboardbutton(text="button", callback_data="button"), inlinekeyboardbutton(text="reset",callback_data="reset")]] inline_keyboard_markup = inlinekeyboardmarkup(inline_keyboard) update.message.reply_text("hi", reply_markup=inline_keyboard_markup) the above code adds 2 buttons each half width of chat screen. i know the normal keyboard button there resize_keyboard parameter somehow can used. my question is there way resize inline buttons ? example make full width or quarter width. unfortunately, can't now. :( you can suggest @botsupport , might add feature next version.
Read more

command line - How can a Python program background itself? -

consider following python command-line program. doesn't work because incomplete, think illustrates problem: #!/usr/bin/python3 import multiprocessing pool = multiprocessing.pool() task in expensive_iterator_with_user_interaction(): pool.call_async(expensive_computation, (task,)) pool.close() print("dear user, may press ctrl-z , bg script.") pool.join() thus, problem scripts needs user input while it's launching expensive parallel processes. there moment when user interaction not needed anymore. then, ask user background process shell actions. however, can achieved program itself? (side note: core of problem entanglement of user input firing “expensive computations”. therefore, cannot separate both in 2 scripts, , letting 1 spawning other daemon.) the command screen may you. use screen -l create new session , run application. after inputting data, can press " ctrl-a " , press " d " on keyboard detach session. now
Read more

php - "cURL error 28: Resolving timed out" on Wordpress on Azure App Service on Linux -

Image
my wordpress installation on azure app service on linux using custom docker container has slow response times. pages take around 20-40 seconds load. i have troubleshooting plugin installed indicates problem "curl error 28: resolving timed out after n milliseconds" when making requests following urls https://api.wordpress.org/core/version-check/1.7/ http://api.wordpress.org/core/version-check/1.7/ https://api.wordpress.org/plugins/update-check/1.1/ http://api.wordpress.org/plugins/update-check/1.1/ https://api.wordpress.org/themes/update-check/1.1/ http://api.wordpress.org/themes/update-check/1.1/ curl works fine on scm-site command line. example works ok. curl -x post http://api.wordpress.org/core/version-check/1.7/ edit if ssh container , run php code, works fine. <?php $url = 'http://api.wordpress.org/core/version-check/1.7/'; $fields = array( 'version' => urlencode('4.8.1'), 'php' => urlencode(
Read more