Chrome 100 may break some websites
With the version number about to reach 100, Google Chrome is experiencing some unexpected problems. And this could affect upwards of 2 billion users, with Firefox and Edge not far behind. The so-called problem is not that the browser will flash back, but lost its original role: to visit websites. Google said that the problem is mainly found in websites developed in cooperation with Duda, and has started to fix it.
The root cause of the problem is actually the version number. Most sites check the user agent string to determine the user's browser version. In Chrome, the current public version, for example, is represented by a string reads: Chrome/96.0.4664.110.
Normally, developers do not need to know the exact version of the browser. Duda will only read the two characters after "Chrome/" by default. So in the case of "Chrome/96.0.4664.110", Duda will only read "96". If "Chrome Canary" is already at "99.0.4791.0", then Duda will only read "99".
This reading method is used for security reasons, to prevent access by older, unsupported Chrome browsers. This is a simple design, but in this way, Chrome 100 is only recognized as Chrome 10.
For compatibility reasons, Duda will prevent Chrome browsers with versions lower than 40 from opening the site, which will result in Chrome 100 not being able to access the site properly.
To avoid this embarrassing situation, web developers had to find a way to correctly parse the three-digit version number before Chrome 100 was launched.
Finding a way to completely fix this vulnerability is tricky, and time is running out. Because Google is developing Chrome at an extremely impressive rate, Chrome version numbers are constantly being added.
Depending on the current release rate, Canary users could encounter the bug as soon as next month, and the billions of users running the stable version are only three to six months away. This also means that Chrome users will need to prepare for network outages or use another browser until the problem is resolved.
Chrome 100 is reminiscent of "Y2K"
Y2K is a common acronym for the year 2000 software bug. The letter Y in the acronym stands for "year" and K stands for kilo, and Y2K was named "millennium bug" because of its association with the millennium timestamp.
In the last century, the act of "storing information" was very expensive for both mainframe and personal computers, costing at least $10 per thousand bytes and in many cases more than $100 per thousand bytes.
For programmers, it was very important to minimize the amount of information stored, and to save 1 bit was 1 bit. so, the date format of data files in those days was six digits, in the form of DDMMYY, with two digits for the day, two digits for the month, and two digits for the year YY. for example, January 1, 1970, was 700101.
There is even a shorter representation, "YYDDD", where DDD is the number of days in a year. Since space on disks and tapes is also expensive, money is saved by reducing the size of the data file in which the date is stored.
However, some programs are unable to distinguish between 2000 and 1900 when faced with a two-digit year. This is where the "millennium bug" comes from.
Google is trying to eliminate the pitfalls of Chrome 100
In fact, on November 1, Google discovered this potential problem. In response, Google provided a test flag for Chrome (enter chrome://flags in the address bar and enable #force-major-version-to-100) that forces the browser's version number to show 100 so that sites can troubleshoot the problem and prepare for the change.
If that's not possible, they've also come up with a fallback: have the first two numbers of the Chrome browser version locked to 99, and then put the actual version number in the second set of numbers. This should work very well for older sites that have little chance of being updated.
This way, the content of the user agent string expressing the version number will be presented as "Chrome/99.100.XX", thus solving the Duda recognition problem.
Of course, this fallback does not cover all cases, as Duda may not be the only web toolkit that has problems. For now, Google wants to be able to definitively find all the sites across the web that were down at the time of Chrome 100's release and contact their developers about the issue.
If these issues can be resolved before Chrome 100 is released at the end of March, then nothing will need to change for web developers. Otherwise, Chrome will have to add new content to the user agent string.