In `@build/debug.sh`:
- Around line 11-16: Replace the Chinese comment "加载Java 17 JVM参数" with an
English comment describing the same purpose (e.g., "Load Java 17 JVM options")
so comments are consistently in English; update the comment near the
jvm_options_file and jvm_options declarations to use the corrected English text.
- Line 18: 第 18 行条件判断使用了非 POSIX 的比较运算符 "=="（在 if [ x"$is_suspend" == x"true" ];
then），会在严格的 sh 实现中失败；将该比较改为单等号 "="（即使用 if [ x"$is_suspend" = x"true" ];
then）以确保与 POSIX sh（dash 等）兼容并提高可移植性，保留变量前的 x 前缀和引号以防空值或特殊字符。

In `@build/fix-groovy-compiler.sh`:
- Around line 3-15: The script contains Chinese comments and echo output (lines
around the header and description) that must be converted to English to match
repo conventions; update the top comment block and all echo strings (e.g., the
header lines currently printed with echo
"==========================================" and the "批量修复Groovy编译配置" message)
to clear, correctly spelled English, and ensure inline comments (above set -e,
SCRIPT_DIR, PROJECT_ROOT) are rewritten in English as well so all user-facing
and inline text in this script (fix-groovy-compiler.sh) uses English only.
- Around line 95-266: The script contains Chinese text and uses an unquoted
heredoc for the embedded Python block; replace all Chinese comments and
user-facing messages with English and switch the heredoc to a single-quoted form
(use <<'PYTHON_SCRIPT') so the shell will not expand variables inside the block,
and pass the FILE path into the Python script via an exported environment
variable (or as an argument) instead of relying on unescaped expansions; update
prints like "✅ 修复完成"/"⚠️  未找到需要替换的内容" and shell echoes "✅ 修复成功"/"⚠️ 
修复失败或无需修复，请手动检查" to clear English equivalents, ensure the Python logic
(functions/variables: pattern, pattern2, replace_plugin, lines/new_lines,
version detection using version_line and version_match) continues to work with
the quoted heredoc and the existing escaped ${project.java.version} strings, and
verify the script exits with the same status behavior.

In `@build/verify-java17.sh`:
- Around line 65-76: The current if-check uses a pipeline (mvn test ... | tee
/tmp/java17-test.log) so tee hides mvn's exit code; enable proper failure
detection by either enabling pipefail (set -o pipefail before the pipeline) or
capture mvn's exit with PIPESTATUS after the pipeline (e.g., inspect
${PIPESTATUS[0]}), then use that exit code to decide the if/else branch for the
mvn invocation shown in the mvn test -pl utils,abstraction ... 2>&1 | tee
/tmp/java17-test.log command and subsequent TEST_FAILED logic.
- Around line 1-98: The script contains Chinese comments and output strings;
replace all Chinese text with clear English equivalents (e.g., change header
echoes, section markers "[1/6]".."[6/6]", comment lines, error messages and
suggestions) while preserving logic and variable names like JAVA_VERSION,
SPRING_VERSION, HIBERNATE_VERSION, ASPECTJ_VERSION, GROOVY_VERSION and temp log
/tmp/java17-test.log; update messages for success/failure (e.g., "❌ 错误: 需要Java
17..." → "ERROR: Java 17 is required, current version: ...", "✓ 核心模块编译成功" → "OK:
core modules compiled") and translate all other echoed guidance and comments to
idiomatic, correctly spelled English.

In `@build/zstack`:
- Line 23: Replace the Chinese comment "# 加载Java 17 JVM参数" with an English
equivalent (e.g., "# Load Java 17 JVM parameters" or "# Load Java 17 JVM
options") in build/zstack; ensure the updated comment text for the JVM parameter
section uses correct English and scan the surrounding comments in the same file
for any other Chinese text to convert to English as well.

In `@build/zstack-debug`:
- Line 23: The comment at line with "# 加载Java 17 JVM参数" must be converted to
English to follow the project's language guidelines; replace the Chinese comment
with an appropriate English equivalent such as "Load Java 17 JVM parameters" or
"Load Java 17 JVM options" in the build/zstack-debug file so all comments and
messages are in correct English and consistent across the codebase.
- Around line 29-30: The script defines the JVM options into a misspelled
variable javaOptitons but then invokes java using javaOptions, so the JVM flags
are never passed; fix by correcting the variable name so both places use the
same identifier (either rename javaOptitons to javaOptions where it’s defined or
change the invocation to use javaOptitons) and ensure the value includes the
intended flags (-Xdebug -Xms256m -Xmx512m -XX:+HeapDumpOnOutOfMemoryError
-Xrunjdwp:transport=dt_socket,address=8787,server=y,suspend=y) so the JVM
receives the parameters.

In `@conf/jvm-options-java17.conf`:
- Around line 1-18: Replace all Chinese comments in this JVM options file with
clear English comments: change the file header comment to "Java 17 module system
configuration", the reflective-access note "开放反射访问（226处setAccessible需要）" to
"Open reflective access (required for setAccessible usages)", the internal-API
export note "导出内部API（sun.misc.Signal需要）" to "Export internal APIs (required for
sun.misc.Signal)", the Hibernate note "Hibernate需要" to "Required by Hibernate",
and the AspectJ note "AspectJ需要" to "Required by AspectJ"; keep the option lines
(--add-opens, --add-exports, --add-opens java.base/java.time, --add-opens
java.base/java.lang.invoke) unchanged.

In `@docs/ARCHITECTURE.md`:
- Around line 1-1199: The document "ZStack 后端架构索引" is in Chinese but the repo
requires English-only content; update this ARCHITECTURE.md by providing an
English version (either replace the Chinese text or add a full English
translation alongside the Chinese), ensuring all headings, descriptions, table
entries, code comments/examples and appended notes (e.g., "文档版本", "最后更新") are
converted to proper English, and keep structural identifiers (like
class/interface names: CloudBus, DatabaseFacade, ThreadFacade, FlowChain,
GlobalConfig, etc.) unchanged so readers can locate code references.

In `@docs/BUILD-ACCELERATION.md`:
- Around line 1-141: The document "ZStack 构建加速指南" is written in Chinese but the
repo requires English-only documentation; please convert this file to English
(or add a complete English version alongside the Chinese) so all headings (e.g.,
"ZStack 构建加速指南"), section titles, prose, and any inline comments are in correct
English and free of spelling errors; keep command examples and flags unchanged,
preserve technical details (mvnd/mvn usage, profiles, flags like -P premium,
-Dmevoco.skip.obfuscate=true, .mvn/jvm.config settings, and plugin notes such as
gmavenplus fork) and, if you retain the Chinese version, add a clear bilingual
header indicating both languages and ensure filenames/links referenced in the
text remain consistent.

In `@longjob/src/main/java/org/zstack/longjob/LongJobFactoryImpl.java`:
- Around line 54-57: The catch block in LongJobFactoryImpl that handles
InstantiationException | IllegalAccessException | NoSuchMethodException |
InvocationTargetException currently calls e.printStackTrace(); replace this with
a proper logger call (e.g., logger.error("Failed to instantiate LongJob via
reflection for " + it.getName(), e)) so the exception is recorded via the class
logger; update the catch to log the exception and any contextual info (the
constructor/type or variable job) instead of printing the stack trace.