Archives, display and access responses
ManYue archive name collisions: distinguish similar appearance from identical characters
Check case and Unicode-normalisation candidates separately, preserving entry and content identities before making a clearly named reading copy.
On this page
Two images may appear to share a name while using different character sequences or only differing in case. This task concerns distinct raw paths that might collide in another environment, unlike duplicate entries with exactly the same path. Record metadata read-only first; neither all file systems nor ManYue necessarily use one name-comparison rule.
Fix the original archive before renaming anything
Choose a permitted small CBZ or ZIP and retain its original, source, size and working copy. List each member’s complete path and index, including directories: identical final names under different directories are not identical full paths. Do not first extract everything to an ordinary image folder or batch-rename a library to remove unfamiliar characters. If a listing tool merges names, record its inability to show full entries; one remaining image does not show that the archive originally had one member. Practise the record with an owned three-marked-image sample before examining real reading material.
Record case-only candidates separately
For example, A001.JPG and a001.jpg in one directory are different raw strings but may enter one group under case-insensitive comparison. This does not mean every device merges them or that their contents match. Assign independent observation identifiers, retaining original spelling, size and corner marks instead of writing lowercased names back into the archive. If a tool shows only friendly labels rather than actual paths, mark insufficient evidence. A comparison rule identifies potential collisions; it does not replace reading in the current environment, and uppercase does not imply a newer edition or better image quality.
Distinguish combining sequences from precomposed characters
The initial of é01.jpg can be one U+00E9 or U+0065 followed by U+0301. They may look alike when normally displayed while remaining different raw sequences. NFC normalisation brings this typical canonical-equivalence pair to one form; it does not make image content identical. Record code points only through an existing capable tool or a knowledgeable read-only computer check, not by claiming visual inspection revealed them. Retain the raw name and chosen normalisation form rather than replacing character evidence with a screenshot. If the raw sequence is unavailable, keep the suspicion provisional instead of inferring encoding from shelf order.
Keep the two comparison results in separate fields
For each path, keep separate fields for raw equality, equality after NFC and case-insensitive candidates. Do not combine different processes into the statement “duplicate files.” Specify whether full paths or final names are compared. Normalisation is not spelling correction, and case folding is not translation; do not also remove spaces or change punctuation and then claim the system considers names equal. In particular, do not silently substitute compatibility normalisation for NFC: circled digits and other distinctions may matter in names. This exercise identifies candidate groups for further checking. Content and platform behaviour need separate evidence; no deletion is triggered.
Check overwrite behaviour before taking out content
Only take known small images into an empty working directory if the tool can select each entry explicitly and save it under a different observation name. Use A/B identifiers and preserve original-path mappings. If it asks to overwrite or cannot distinguish entries, stop at the listing stage instead of replacing everything. Compare marks, dialogue and neighbouring pages; an already available fingerprint tool may assist byte comparison. Equal normalised names do not imply equal image bytes or appearance. If one seems older, choosing authoritative text still requires edition evidence; modification time and post-extraction numbering cannot supply it.
Record which images actually appear in ManYue
Import the working copy normally and locate candidate images by marks and content. Record app version, chapter, position and whether neighbouring pages can be read. Do not assume ManYue merges case, compares NFC or chooses the last entry; each claim needs current evidence. If only one image is found, write that A was located and B’s display remains unresolved rather than declaring B deleted. Confirm both records remain in the original archive, separating archive structure from app presentation. When search or directory labels look identical, image anchors are especially important so repeated visits to one image are not mistaken for checking two entries.
Preserve name mappings in a separate reading copy
If owned material genuinely needs organisation, make a separate reading copy whose clear names do not rely on case or combining-character differences, such as sequential page numbers plus observation identifiers. Keep an old-to-new path list and retain the original archive and extracted content. Manual renaming is not an app repair. Do not remove every accented letter, Han character or variant indiscriminately: many names do not collide, and changes may lose bibliographic information. Reimport the copy and check every mark and reading sequence; tidy filenames do not prove that no image was lost or replaced.
Exit and return to check identities and comparison rules
Exit the reader and return to both the original working copy and organised copy. Relocate candidate groups and neighbours from the record. Retain six items: raw path, observation identifier, comparison rule, character evidence, content status and app presentation, leaving unseen fields unknown. State the tools and environment covered instead of generalising all systems’ name behaviour from one observation. Completion means both naming differences and content identities remain traceable without silent overwrite. It does not classify every visually similar name as a duplicate comic. If another device behaves differently, make a new record rather than overwriting the earlier result.
Related guides and sources
ManYue: Audit repeated ZIP entry paths before extracting or changing files; ManYue volume filenames: distinguish numbering, editions and extras.
Checked on November 10, 2026: Apple public version information lists ManYue 2.0.5, requiring iOS / iPadOS 15.0 or later. See the official App Store for feature descriptions; specific controls must still be checked in the installed version.
Method reference: Unicode Standard Annex 15: Normalization Forms. Numbers, short passages and controls are original exercises. Sources explain concepts rather than prove automatic recognition or grading in the app. Manual notes may be kept externally. This article does not claim step-by-step testing on a physical device or private server; unresolved content should remain marked for review.