4 ms·

Safari caused me by far the most issues, back when I was doing frontend in ~2017/2018, in terms of how it complied with web standards (I wasn't as fussed at the

by lol768 3y ago

Safari caused me by far the most issues, back when I was doing frontend in ~2017/2018, in terms of how it complied with web standards (I wasn't as fussed at the standards it didn't support - at least I knew!) and rendered the web application I was working on across desktop, iPad and iPhone. I spent more time chasing Safari bugs and rendering idiosyncrasies than I did dealing with IE11, frankly.

euroderf 3y ago
Satisfied Safari user here. Chrome had me chasing shiny, shiny but nowadays if something doesn't run right in Safari, life goes on, and Apple ecosystem integration more than compensates. Disclaimer: I am not a web dev.
toastal 3y ago
And bugfixes always took 6 months til the next OS update. Holding back Notifications and other PWA-related specs left a sour taste to me because it meant I couldn’t build a PWA …but you could buy Apple hardware a developer license and make a native app (even if it was mostly CRUD).
burai 3y ago
Most of the problems I faced in the projects I work are caused by developers working only and exclusively with Chrome and not bothering to check their work on Safari and Firefox. The issues that were found with Safari were easy to fix or work around. Nothing compares to the painful process during ~2005-2010 when it came to support IE6.
- blackoil 3y ago
  Safari testing requires a mac/iP*
  - MatthiasPortzel 3y ago
    Safari’s UI is based on mac-platform components, but the WebKit team provides builds of the WebKit rendering engine with a GTK-based UI. => https://webkit.org/downloads https://webkit.org/downloads
whalesalad 3y ago
Nothing compares to IE6. That was a hellscape.
joshstrange 3y ago
I’m sorry but just no. I’ve been developing on the web since IE6 and while Safari has a few quirks it doesn’t even start to compare to IE at any stage. The worst I deal with in Safari is lack of certain features (often things that weren’t standardized and Chrome just added) and some minor CSS quirks. In some of those cases when I dig into the “issue” it appears that Safari is handling it in a way that makes more sense. The only Safari issue that I have run into multiple times and disliked is the viewport height issue but new CSS units (dvh I think?) should fix that. On the other hand IE11 had a whole slew “Why isn’t this working?” Issues that made zero sense. I _loved_ the bugs that would not reproduce if you had dev tools (or what passed for IE dev tools) open or things related to cross-origin that just behaved differently for no good reason. QA found way more issues in IE browsers than Safari in my experience. And while IE11 was the least bad of the IEs I don’t think you get to use it as a goalpost. IE6-10 (and 11) caused me immense pain and frustration. Have we all forgotten using images to display a rounded corner? A 1px wide image to display a gradient? Or just images as buttons in general since CSS support was shit? The gross hacks I did for IE do not compare to what I need to do for Safari and anyone who says differently either didn’t live through that period (as a developer) or has a short memory.
- Klonoar 3y ago
  Gradients could be done via base64 css if you triggered the Outlook rendering engine under the hood. ;)
  - joshstrange 3y ago
    Haha, I don't think I ever used that trick but on a number of projects I know I dropped in a `generateGradient.php`-type file that had it's output cached so that in CSS we could just do `background: url(/generateGradient.php?start=XXXXXX&end=XXXXXX)` (and maybe some other params as time went on) to make changes easier.
    - Klonoar 3y ago
      https://rymc.io/blog/2010/you-got-your-base64-in-my-css/ https://rymc.io/blog/2010/you-got-your-base64-in-my-css/ If you want a time travel portal this morning. ;)
j1elo 3y ago
Yeah, they usually do their own thing. Not only in following specs, though. Like how if you use self-signed cert for HTTPS (e.g. for local development and tests while playing with WebRTC), all browsers will show a Warning page with an escape hatch that can be clicked to ignore the problem. You can grab your personal phone, or ask anyone to lend you their phone in a pinch, and check your marvelous web creation. Did I say "all browsers"? Sorry, not quite. iOS Safari outright refuses connecting. And it won't say why. I had to connect the damn iPhone with a cable to an Xcode machine, to open up browser console and see the error there. Turns out Safari is a special snowflake that requires doing some preparation ahead of time (no less than 3 steps!) 1) getting the Root CA into the phone, 2) install it in the cert storage, by the way you won't be able to do it if used a non-Apple app for step 1; and 3) enable full trust in the new CA. All that for a phone that I quickly borrowed from my flatmate. Thanks but no thanks! Not all developments are enterprise projects with proper laboratory devices.
- cubefox 3y ago
  The dramatic (RED ALERT DANGER !!!) certificate error messages are completely over the top. Safari is worse than others, but all are bad. The browser vendors are acting as if HTTPS with certificate issues is a hundred times worse than plain unencrypted HTTP, which it is not.

by lol768 3y ago

euroderf 3y ago

Satisfied Safari user here. Chrome had me chasing shiny, shiny but nowadays if something doesn't run right in Safari, life goes on, and Apple ecosystem integration more than compensates. Disclaimer: I am not a web dev.

toastal 3y ago

And bugfixes always took 6 months til the next OS update. Holding back Notifications and other PWA-related specs left a sour taste to me because it meant I couldn’t build a PWA …but you could buy Apple hardware a developer license and make a native app (even if it was mostly CRUD).

burai 3y ago

Most of the problems I faced in the projects I work are caused by developers working only and exclusively with Chrome and not bothering to check their work on Safari and Firefox. The issues that were found with Safari were easy to fix or work around. Nothing compares to the painful process during ~2005-2010 when it came to support IE6.

blackoil 3y ago
Safari testing requires a mac/iP*
- MatthiasPortzel 3y ago
  Safari’s UI is based on mac-platform components, but the WebKit team provides builds of the WebKit rendering engine with a GTK-based UI. => https://webkit.org/downloads https://webkit.org/downloads

whalesalad 3y ago

Nothing compares to IE6. That was a hellscape.

joshstrange 3y ago

I’m sorry but just no. I’ve been developing on the web since IE6 and while Safari has a few quirks it doesn’t even start to compare to IE at any stage. The worst I deal with in Safari is lack of certain features (often things that weren’t standardized and Chrome just added) and some minor CSS quirks. In some of those cases when I dig into the “issue” it appears that Safari is handling it in a way that makes more sense. The only Safari issue that I have run into multiple times and disliked is the viewport height issue but new CSS units (dvh I think?) should fix that. On the other hand IE11 had a whole slew “Why isn’t this working?” Issues that made zero sense. I _loved_ the bugs that would not reproduce if you had dev tools (or what passed for IE dev tools) open or things related to cross-origin that just behaved differently for no good reason. QA found way more issues in IE browsers than Safari in my experience. And while IE11 was the least bad of the IEs I don’t think you get to use it as a goalpost. IE6-10 (and 11) caused me immense pain and frustration. Have we all forgotten using images to display a rounded corner? A 1px wide image to display a gradient? Or just images as buttons in general since CSS support was shit? The gross hacks I did for IE do not compare to what I need to do for Safari and anyone who says differently either didn’t live through that period (as a developer) or has a short memory.

Klonoar 3y ago
Gradients could be done via base64 css if you triggered the Outlook rendering engine under the hood. ;)
- joshstrange 3y ago
  Haha, I don't think I ever used that trick but on a number of projects I know I dropped in a `generateGradient.php`-type file that had it's output cached so that in CSS we could just do `background: url(/generateGradient.php?start=XXXXXX&end=XXXXXX)` (and maybe some other params as time went on) to make changes easier.
  - Klonoar 3y ago
    https://rymc.io/blog/2010/you-got-your-base64-in-my-css/ https://rymc.io/blog/2010/you-got-your-base64-in-my-css/ If you want a time travel portal this morning. ;)

j1elo 3y ago

Yeah, they usually do their own thing. Not only in following specs, though. Like how if you use self-signed cert for HTTPS (e.g. for local development and tests while playing with WebRTC), all browsers will show a Warning page with an escape hatch that can be clicked to ignore the problem. You can grab your personal phone, or ask anyone to lend you their phone in a pinch, and check your marvelous web creation. Did I say "all browsers"? Sorry, not quite. iOS Safari outright refuses connecting. And it won't say why. I had to connect the damn iPhone with a cable to an Xcode machine, to open up browser console and see the error there. Turns out Safari is a special snowflake that requires doing some preparation ahead of time (no less than 3 steps!) 1) getting the Root CA into the phone, 2) install it in the cert storage, by the way you won't be able to do it if used a non-Apple app for step 1; and 3) enable full trust in the new CA. All that for a phone that I quickly borrowed from my flatmate. Thanks but no thanks! Not all developments are enterprise projects with proper laboratory devices.

cubefox 3y ago
The dramatic (RED ALERT DANGER !!!) certificate error messages are completely over the top. Safari is worse than others, but all are bad. The browser vendors are acting as if HTTPS with certificate issues is a hundred times worse than plain unencrypted HTTP, which it is not.