AnqiCMS (AnqiCMS) Template Tag Manual
The automatic anchor text handling of in-site keywords is based on hits, the wider the vocabulary, the more frequently the same word appears in the text, the denser the generated internal links will be, and the readability of the text and the distribution of link weights will be diluted. This article explains why the density goes out of control, the three configuration ideas to check before going online, how to manage the keyword library hierarchically, and its relationship with sensitive word filtering and keyword replacement in sequence.
The title change may also change the address, depending on what the static address is derived from: when split by number or time slice, the address can remain unchanged, and a new address will be generated when an alias is created from the title. This article provides two judgment methods for these situations, as well as the order of handling old links - retain the original address, add a 301 redirect, update the sitemap and active push, and review the in-site anchor text, and also specifies which positions to check after the change.
The custom content model supports flexible content structures, but the cost of adding a field and modifying a field is not the same: adding a field mainly affects the input method and template value, while modifying a field will affect the historical data of published content, template value, and in-site links. This article provides four items to confirm before going online, batch approaches when necessary to make changes, and a checklist to run after the changes are made.
After container deployment, the patch of the Web server brought by the base image is still followed up by the site operator, and the responsibility will not shift because the image is built by someone else. It should be divided into three layers: application layer, image layer, and reverse proxy layer, with different repair versions and update rhythms at each layer. This article provides the division of responsibilities for the three layers, why an unchanged image tag is equal to no upgrade, and the operation sequence of rebuild and grayscale.
When the static rules fall to the Web server, the instruction responsible for finding files in order is often written incorrectly for the fallback target. Its semantics is to check the existence of files in the given order, using the first matching file to handle the request in the current context; only if all do not exist is an internal redirection made to the last parameter. This article explains two syntax forms, the slash convention for directory judgment, and the common mistakes when falling back to the back-end entry.
Scheduled release and batch import together, the key to scheduling is in state division: the documents imported in batch are first stored as draft or pending release, and then released peak-time with scheduled release. This article explains the meaning of each of the four document states: formal, draft, pending release, and recycling bin, how to verify the draft link with preview parameters, the troubleshooting sequence of missing fields in import, and how to recover deleted content through the recycling bin.
The candidate version is at the final stage of the release process, the purpose is to let people verify the upcoming packaged final version, not the version for daily use on production stations. This article uses the recent example of Joomla simultaneously releasing 6.2 candidate version and 5.x, 6.x dual branch security repair announcements to explain where the candidate version should run, which branch to upgrade first when running dual version lines in parallel, and the backup and rollback points that must be prepared before the upgrade.
Form and comment anti-robot spambot prevention, CAPTCHA only blocks the automatic submission layer, and cannot block scripts that have bypassed verification and low-cost manual submissions. The layered叠加 can be: submission verification (reCAPTCHA CAPTCHA, timestamp and honeytrap fields), content filtering (sensitive word filtering, keyword replacement, content review), publishing strategy (review before posting, pending publication status, frequency limit and login threshold). This article explains the limitations of each layer, as well as what to adjust when all three layers cannot block it.
The last modified time in the sitemap should be filled in according to the actual change: the protocol requires the use of W3C Datetime format, which can only be written as a date; changefreq and priority are defined as hints rather than instructions, and full-site refresh will only make the crawler treat this field as noise. This article explains the protocol meaning of this field, the four common ways it is filled incorrectly, how to obtain the real change time from the content status, and the time scope to be noted when配合 with timed publishing.
The single sitemap file has two concurrent limits: the number of addresses should not exceed 50,000, the uncompressed volume should not exceed 50MB, and any that reach the top must be split into multiple fragment files, and then a index file is used to list the fragments one by one. This article explains how to calculate the limit according to the original protocol, how to name the fragment directory stably, the division of labor between the index file and the sub-maps, as well as the pseudo-static rules and configuration to check when the background automatically generates the sitemap, to avoid bypassing the limit by trimming the number of addresses in an erroneous manner.
