这里记录我的一周分享,通常在周六发布。
梨梨原上草 @svuoalnalnis
公司把保洁外包给一间公司,常来的就那么几个面孔。时间长了,你就能体会她们之间的差别。
有时候一走进厨房,你就知道今天来上班的肯定是其中的某一个。而另外一些天,不用具体看脸也知道来摸鱼的是哪几个……
其实想对其中几个说,你们可能干啥都会做得很好,不限于保洁。只要时间允许,我的活也能干。
还真是这样。一生要是每本书都一字一句的读完,确实读不了几千本。但真正值得一字一句读完的书完全没有那么多,你需要略读上万本书(甚至可以更多)才能找出为数不多值得读一次又一次的那些书。一本书翻了几页放下,没什么,很多书没有读完,没什么。没有读完的那些书,书自己也有责任。
戚小诺 我问小九怎么表达他的愤怒,他说他的脑子里会有一个白色小人,一个黑色小人,黑色小人想要去干架:冲啊!我要干架!白色小人就在一边劝架:算了算了,不要打架,也许打不赢会受伤,爸爸妈妈还会担心。然后自己就会冷静很多。——头脑特工队没白看哈😂😂
memoryza 人们天生就是高效的改进者,而只有极少数会愿意从零开始
失败不可怕,一蹶不振才可怕
反馈是变得更好的关键。正如蓝迪波什所说:"人之所以会改变,是因为获得了反馈"。"失败"的本质是一种强反馈,它告诉我们这样或者此时不行,我们需要调整。在这个世界上,只要我们的生命还在继续(留得青山在),就没有真正的失败(不怕没柴烧),有的只是不断调整的过程。一个人的成长速度,取决于他得到反馈的频率和质量。因此,我们应该勇于改变并珍视每一次的反馈,正是它们引导着我们一步步实现人生目标。
真正重要的工作都不是一口气完成的,而是反复回到同一主题,慢慢打磨。像居里夫人发现放射性元素那年夏天,也去乡下休假、爬山。伟大成就需要的是以年为单位的积累,而不是每周都冲刺。别被高强度伪忙碌欺骗,真正的生产力,是以自然节奏、长期愿景,持续推进关键事务。
生活若总是愿意吃苦,就有吃不完的苦;若总是抱怨生活,就有抱怨不完的生活。总觉得前路还长,却不知不觉就进入四十几五十几。能改善生活质量的钱一定要尽早花,早花早享受。
调查期望不是让用户畅想这产品能具有什么功能,而是为了满足现有需求应该具有什么功能。
LLM 的注意力其实是个滑动窗口,不持续提醒,很容易跑偏,这一点就跟我们管理一个想法很多的员工一个道理。(哎!人的注意力也是个滑动窗口。)
Motivation: though reinstalling OS is rare, but I still want to automate the DNS update process.
Get API Token
- Go to https://dash.cloudflare.com/profile/api-tokens
- Click "Create Token"
- Use "Edit zone DNS" template
- Select your domain in "Zone Resources"
- Copy the generated token
Setup API Token in bashrc
Add to ~/.bashrc
:
export CF_API_TOKEN="your_api_token_here"
Reload: source ~/.bashrc
Enhanced Update Script
Create update-dns.sh
:
#!/bin/bash
# Usage function
usage() {
echo "Usage: $0 <record_name> [ip_address]"
echo " $0 <domain> <subdomain> [ip_address]"
echo "Examples:"
echo " $0 home.example.com # Uses Tailscale IP"
echo " $0 example.com # Root domain with Tailscale IP"
echo " $0 server.example.com 192.168.1.100 # Uses specified IP"
echo " $0 example.com home # Legacy format"
echo " $0 example.com @ 1.2.3.4 # Legacy root domain"
exit 1
}
# Check parameters
if [ $# -lt 1 ]; then
usage
fi
# Check if API token is set
if [ -z "$CF_API_TOKEN" ]; then
echo "Error: CF_API_TOKEN not set. Add it to ~/.bashrc"
exit 1
fi
# Parse arguments - detect format
FIRST_ARG="$1"
SECOND_ARG="$2"
THIRD_ARG="$3"
# Check if first argument contains multiple dots (FQDN format)
if [[ "$FIRST_ARG" == *.*.* ]] || ([[ "$FIRST_ARG" == *.* ]] && [[ "$SECOND_ARG" =~ ^[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+$ ]]); then
# FQDN format: script.sh home.example.com [ip]
RECORD_NAME="$FIRST_ARG"
CUSTOM_IP="$SECOND_ARG"
# Extract domain by removing first subdomain part
DOMAIN=$(echo "$RECORD_NAME" | cut -d'.' -f2-)
elif [[ "$FIRST_ARG" == *.* ]] && [ -z "$SECOND_ARG" ]; then
# Root domain format: script.sh example.com
RECORD_NAME="$FIRST_ARG"
DOMAIN="$FIRST_ARG"
CUSTOM_IP=""
elif [[ "$FIRST_ARG" == *.* ]] && [ -n "$SECOND_ARG" ] && [[ ! "$SECOND_ARG" =~ ^[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
# Legacy format: script.sh example.com home [ip]
DOMAIN="$FIRST_ARG"
SUBDOMAIN="$SECOND_ARG"
CUSTOM_IP="$THIRD_ARG"
if [ "$SUBDOMAIN" = "@" ]; then
RECORD_NAME="$DOMAIN"
else
RECORD_NAME="$SUBDOMAIN.$DOMAIN"
fi
else
echo "Error: Invalid arguments format"
usage
fi
# Get IP address
if [ -n "$CUSTOM_IP" ]; then
NEW_IP="$CUSTOM_IP"
echo "Using provided IP: $NEW_IP"
else
if command -v tailscale >/dev/null 2>&1; then
NEW_IP=$(tailscale ip -4)
echo "Using Tailscale IP: $NEW_IP"
else
echo "Error: tailscale not found and no IP provided"
exit 1
fi
fi
# Validate IP format
if ! echo "$NEW_IP" | grep -E '^[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}$' >/dev/null; then
echo "Error: Invalid IP address format: $NEW_IP"
exit 1
fi
echo "Updating DNS record: $RECORD_NAME -> $NEW_IP"
# Get Zone ID
echo "Getting Zone ID for $DOMAIN..."
ZONE_RESPONSE=$(curl -s -H "Authorization: Bearer $CF_API_TOKEN" \
"https://api.cloudflare.com/client/v4/zones?name=$DOMAIN")
ZONE_ID=$(echo "$ZONE_RESPONSE" | jq -r '.result[0].id')
if [ "$ZONE_ID" = "null" ] || [ -z "$ZONE_ID" ]; then
echo "Error: Could not find zone for domain $DOMAIN"
echo "Response: $ZONE_RESPONSE"
exit 1
fi
echo "Zone ID: $ZONE_ID"
# Get Record ID
echo "Getting Record ID for $RECORD_NAME..."
RECORD_RESPONSE=$(curl -s -H "Authorization: Bearer $CF_API_TOKEN" \
"https://api.cloudflare.com/client/v4/zones/$ZONE_ID/dns_records?name=$RECORD_NAME&type=A")
RECORD_ID=$(echo "$RECORD_RESPONSE" | jq -r '.result[0].id')
if [ "$RECORD_ID" = "null" ] || [ -z "$RECORD_ID" ]; then
echo "Error: Could not find A record for $RECORD_NAME"
echo "Available records:"
echo "$RECORD_RESPONSE" | jq -r '.result[] | "\(.name) (\(.type))"'
exit 1
fi
echo "Record ID: $RECORD_ID"
# Update DNS record
echo "Updating DNS record..."
UPDATE_RESPONSE=$(curl -s -X PUT \
"https://api.cloudflare.com/client/v4/zones/$ZONE_ID/dns_records/$RECORD_ID" \
-H "Authorization: Bearer $CF_API_TOKEN" \
-H "Content-Type: application/json" \
--data "{\"type\":\"A\",\"name\":\"$RECORD_NAME\",\"content\":\"$NEW_IP\"}")
SUCCESS=$(echo "$UPDATE_RESPONSE" | jq -r '.success')
if [ "$SUCCESS" = "true" ]; then
echo "✅ DNS record updated successfully!"
echo " $RECORD_NAME now points to $NEW_IP"
else
echo "❌ Failed to update DNS record"
echo "Response: $UPDATE_RESPONSE"
exit 1
fi
Make executable: chmod +x update-dns.sh
Usage Examples
# New FQDN format (recommended)
./update-dns.sh home.example.com # Uses Tailscale IP
./update-dns.sh server.example.com 192.168.1.100 # Uses custom IP
./update-dns.sh example.com # Root domain with Tailscale IP
# Legacy format (still supported)
./update-dns.sh example.com home # Subdomain with Tailscale IP
./update-dns.sh example.com @ 1.2.3.4 # Root domain with custom IP
# Multiple subdomains
./update-dns.sh web.example.com
./update-dns.sh api.example.com
./update-dns.sh db.example.com 10.0.0.5
Features
- ✅ API token stored securely in
.bashrc
- ✅ Auto-discovers Zone ID and Record ID
- ✅ Supports custom IP or auto-detects Tailscale IP
- ✅ Supports root domain updates with
@
- ✅ Input validation and error handling
- ✅ Clear success/failure messages
- ✅ Usage help and examples
Alternative: Using flarectl
# Install
go install github.com/cloudflare/cloudflare-go/cmd/flarectl@latest
# Update (simpler)
export CF_API_TOKEN="your_token"
flarectl dns update --zone yourdomain.com --name subdomain --content $(tailscale ip -4) --type A
Background
Yesterday, my PopOS system broke unexpectedly. The refresh install option failed, and attempting to reinstall while preserving the old /var
and /home
partitions was unsuccessful.
Solution: Fresh Custom Installation
Performed a clean custom installation with:
/boot/efi
- formatted
/
- formatted
/recovery
- formatted
This approach worked successfully, then migrated data from old partitions.
Partition Migration Process
Tools Used
parted --list # Get partition information
blkid # Get partition UUIDs
Note: Couldn't stop GDM or enter single-user mode (system would hang). Performed migration on running system instead.
1. Home Partition Migration (/dev/nvme0n1p8)
sudo mv /home /home.bak # Backup current home
sudo mkdir /home
sudo mount /dev/nvme0n1p8 /home
ls /home # Confirm correct partition
sudo blkid /dev/nvme0n1p8 # Get UUID
sudo vi /etc/fstab
Added to /etc/fstab
:
UUID=32e4ed56-6ed9-4f14-9585-ffff54f997b2 /home ext4 defaults 0 2
Rebooted to verify system stability.
2. Backup Partition Setup (/dev/nvme0n1p10)
sudo parted --list
sudo blkid /dev/nvme0n1p10
sudo vi /etc/fstab
Added to /etc/fstab
:
UUID=8aba0dd5-7058-4036-8a08-fcd1e0e002bc /backup ext4 defaults 0 2
sudo mount -a # Test mount configuration
3. Var Partition Migration (/dev/nvme0n1p7)
# Mount old var partition and backup data
sudo mkdir /mnt/var
sudo mount /dev/nvme0n1p7 /mnt/var
sudo tar -czpvf /backup/old-var-backup.tar.gz -C /mnt var
# Prepare new var partition
sudo umount /mnt/var
sudo mkfs.ext4 /dev/nvme0n1p7 # Format old var partition
sudo mkdir /mnt/new-var
sudo mount /dev/nvme0n1p7 /mnt/new-var
# Copy current var data to new partition
sudo cp -av /var/* /mnt/new-var/
# Update fstab and switch to new partition
sudo blkid /dev/nvme0n1p7
sudo vi /etc/fstab # Add var partition entry
sudo mv /var /var.old && sudo mkdir /var && sudo mount -a
Key Lessons Learned
- Fresh installation with custom partitioning was more reliable than trying to preserve old partitions during install
- System migration on a running system worked fine when rescue mode wasn't accessible
- Always backup critical data before formatting partitions
- Test each partition mount configuration before proceeding to the next step
Status: Migration completed successfully. System running normally with all data preserved.
Need to share changes from a GitHub PR (even closed ones)? Just add .diff
or .patch
to the PR URL:
Original URL:
https://github.com/owner/repo/pull/123
For diff format:
https://github.com/owner/repo/pull/123.diff
For patch format:
https://github.com/owner/repo/pull/123.patch
Both work for open, closed, or merged PRs. Perfect for code reviews, investigations, or sharing changes with others.
For LLM/AI analysis: Use .diff
format - it's cleaner and more standardized than .patch
which includes extra email headers.
Problem
Rider's WinForms Designer locks files, preventing builds. It is super annoying!
Solution
- File → Settings (Ctrl+Alt+S)
- Tools → Windows Forms Designer
- Uncheck "Enable Windows Forms Designer"
- Apply and restart Rider
Emergency Fix
Kill locked processes:
taskkill /IM dotnet.exe /F
Alternative
Use Visual Studio for form design, Rider for code.