Courtesy of @chenmaop:
https://rawgit.com/posclegom/programthinkmirror/master/
这个是墙内可达的
对于repo过大不好下载有一个不太优雅的解决方案,就是利用 svn checkout https://github.com/posclegom/programthinkmirror/trunk/forum4 只选择性下载forum4文件夹,其余部分直接从btsync中复制过来

for my dear old friend

all scripts below, (in the form of master[n].sh)
presumes that the <BTSync folder of programthinkmirror>
is /root/blog/
you need to change that folder in all the scripts
to have them work properly if your BTSync reception
folder of programthinkmirror is different.


to update forum section only:

cd <hosting directory of programthinkmirror>
cd forum4
./year.sh
wc -l olddd
wc -l newww
make sure the two match
 if not, then programthink has added a new blog article
 in this case please refer to section below
./master0.sh
./master6.sh
./master4.sh


when programthink has added a new blog article:
(note: sometimes programthink lags in updating the
BTSync off-line blog folder, in this case one
can update the forum section first, then update
the blog when the updated off-line blog folder
becomes available later. That is, updating the nlist
without updating the mlist, and carry out all the
steps associated with updating the nlist/forum without
carrying out the final steps associated with updating
the mlist and the overall blog, e.g. master7.sh, etc.)

update src/mlist/mlist
update src/nlist/nlist
cp src/mlist/mlist .
cp src/nlist/nlist forum4/
cd forum4
mv ../src/pt17/shift.sh .
edit seq <bign> ... to seq <bign++> ...
uncomment seq line
./shift.sh (RUN ONLY ONCE!!!)
recomment seq line
mv shift.sh ../src/pt17/
./year.sh
vim olddd
add "0\n" line to the beginning of olddd
wc -l olddd
wc -l newww
wc -l nlist
wc -l ../mlist
make sure all 4 match
./master0.sh
./master6.sh
./master4.sh
cd ..
./master9b.sh
./master9.sh
./master7.sh
cp <BTSync folder of blog>/blog/html/index.html .
sed -i -e 's/http:/https:/g' index.html
(other manual cosmetic changes are applied to index.html at your discretion, please refer to index.html in this repository)
cd forum4
vim point
add initial comment count of new blog article to beginning of file
:wq


an additional note on src/:

the src/ contains both production and historical/deprecated codes,
folders containing production codes are:
pt10/
pt17/
pt18/
pt20/
pt21/

an additional note on >500 forums for the first time:

when an article's forum comments exceed 500,
do the following after updating the forum section:
./master9b.sh
./master9.sh
./master7.sh


an additional note on >1000 forums:

comments >500 are all appended into section C
pt17/sensor.sh reveals difference between last checkpoint
when difference exceed 500,
(use the lab/ directory for the following)
# append section B to current section C
# and make resulting file the new section C
vim nB.xml
:set encoding=utf-8
:wq
vim nC.xml
:set encoding=utf-8
:wq
cat nB.xml nC.xml > out.xml
vim out.xml
:set encoding=utf-8
:set nowrap
go to second line
1500x
150x
then x until the first <entry> tag
go to first line
$
remove ending </feed> tag
:wq
mv out.xml ../forum4/nC.xml
cd ../forum4
vi point
update *ONLY the specific line* of the file being worked on
:wq

note on code in labforum/ and src/pt21/
these codes are for verifying that the number of comments
for each forum are up-to-date with respect to actual counts
in the real forums on program-think.blogspot.com
how to use:
after a fresh forum update
cd forum4/
cp *.xml ../labforum/
cd ..
cd labforum
./master4.sh
./count.sh > count
diff count ../forum4/newww
(result from diff should be totally blank, any output
from diff indicates discrepancy between mirror forum
and actual forum and can be investigated by following
the line number on the display for diff, the line numbers
start from 1, so 1 should be subtracted from the line
number outputs in order to go to the relevant
forums)