Why Asia/Shanghai does not have LMT?

[shellinger]

According to tzdb-2019c of IANA, Asia/Shanghai uses LMT before 1901, as file "asia" says:
# Beijing time, used throughout China; represented by Shanghai.
Zone Asia/Shanghai 8:05:43 - LMT 1901
\t \t \t 8:00 Shang C%sT 1949 May 28
\t \t \t 8:00 PRC C%sT

But in moment-timezone-with-data.js, it uses the packed data as:
"Asia/Shanghai|CST CDT|-80 -90|010101010101010101010101010|-1c2w0 Rz0 11d0 1wL0 A10 8HX0 1G10 Tz0 1ip0 1jX0 1cN0 11b0 1oN0 aL0 1tU30 Rb0 1o10 11z0 1qN0 WL0 1qN0 11z0 1o10 11z0 1o10 11z0|23e6",

Apparently don't include the LMT part. How comes this inconsistence? Is it a bug?

BTW: What is the last number mean in the packed date (23e6 in above case). I can't find any explanation in all documents. Thanks!

[gilmoreorless]

This looks like it's caused by a drop in data size in recent releases (see #768). Specifically, the latest data files are only handling 32-bit timestamps (which only allows for going back to 1901-12-13).

Data files built in older releases contain the expanded range:

$ grep Shanghai data/packed/2019a.json | head -n1
"Asia/Shanghai|LMT CST CDT|-85.H -80 -90|0121212121212121212121212121|-2M0U5.H 1zWo5.H Rz0 11d0 1wL0 A10 8HX0 1G10 Tz0 1ip0 1jX0 1cN0 11b0 1oN0 aL0 1tU30 Rb0 1o10 11z0 1qN0 WL0 1qN0 11z0 1o10 11z0 1o10 11z0|23e6",

For your other question, the last value is the population field (in scientific notation, so 23e6 means 23,000,000). This is only used for ranking zones when using moment.tz.guess().