做 HTTPS 請求時,客戶端往往得自己拼出一段完整的網址。問題來了:只要路徑裡出現中文,程式立刻 crash。追下去才發現,罪魁禍首是中文沒有做轉碼。
- (NSString *)stringByAddingPercentEscapesUsingEncoding:(NSStringEncoding)encoding - (NSString *)stringByReplacingPercentEscapesUsingEncoding:(NSStringEncoding)encoding
前者負責把字串轉成百分號編碼,後者負責還原。實測有效。
NSString *string1 = @"https://www.cloudsafe.com/文件夾"; NSString *string2 = [string1 stringByAddingPercentEscapesUsingEncoding:NSUTF8StringEncoding]; NSString *string3 = [string2 stringByAddingPercentEscapesUsingEncoding:NSUTF8StringEncoding]; NSString *string4 = [string2 stringByReplacingPercentEscapesUsingEncoding:NSUTF8StringEncoding]; NSString *string5 = [string3 stringByReplacingPercentEscapesUsingEncoding:NSUTF8StringEncoding]; NSString *string6 = [string4 stringByReplacingPercentEscapesUsingEncoding:NSUTF8StringEncoding]; NSString *string7 = [string5 stringByReplacingPercentEscapesUsingEncoding:NSUTF8StringEncoding];
輸出結果:
string1: https://www.cloudsafe.com/文件夾 string2: https://www.cloudsafe.com/%E6%96%87%E4%BB%B6%E5%A4%B9 string3: https://www.cloudsafe.com/%25E6%2596%2587%25E4%25BB%25B6%25E5%25A4%25B9 string4: https://www.cloudsafe.com/文件夾 string5: https://www.cloudsafe.com/%E6%96%87%E4%BB%B6%E5%A4%B9 string6: https://www.cloudsafe.com/文件夾 string7: https://www.cloudsafe.com/文件夾
從 string2 與 string3 可以看出:對已經編碼過的字串再編碼一次,百分號本身也被編掉了(%E6 變成 %25E6),這就是二次轉碼的由來。
我的做法是:URL 拼完之後,對整個字串連續轉碼兩次。
NSMutableString *address = [[NSMutableString stringWithString:SetNiChengStringWithoutUserNameAndNiCheng] mutableCopy]; address = [[address stringByAppendingString:app.name] mutableCopy]; address = [[address stringByAppendingString:@"/"] mutableCopy]; address = [[address stringByAppendingString:_nameTextField.text] mutableCopy]; address = [[address stringByAddingPercentEscapesUsingEncoding:NSUTF8StringEncoding] mutableCopy]; address = [[address stringByAddingPercentEscapesUsingEncoding:NSUTF8StringEncoding] mutableCopy];
如果後端只是單純把路徑存起來,收到資料後可以完全不轉回中文,就讓它維持編碼後的樣子。等到客戶端要讀取這段資料時,自己解一次碼即可正常顯示:
str = [str stringByReplacingPercentEscapesUsingEncoding:NSUTF8StringEncoding];
這樣一來,中文路徑在傳輸層全程都是安全的 ASCII 字元,顯示層則交給客戶端還原,責任切得很乾淨。