You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flex.apache.org by Harbs <ha...@gmail.com> on 2017/03/09 09:18:40 UTC
[FalconJX] regex patterns
I’ve come across quite a few regex patterns which break compilation. Here’s some examples:
private static const _newLineTabPattern:RegExp = /[\u0009\u000a\u000d]/g;
static private const brRegEx:RegExp = /\u2028/;
private static const _newLineRegex:RegExp = /\u000A|\u000D\u000A?/g;
public static const anyPrintChar:RegExp = /[^\u0009\u000a\u000d\u0020]/g;
public static const attrRegex:RegExp = /\s+(\w+)(?:\s*=\s*(".*?"|'.*?'|[\w\.]+))?/sg;
public static const tagRegex:RegExp = /<(\/?)(\w+)((?:\s+\w+(?:\s*=\s*(?:".*?"|'.*?'|[\w\.]+))?)*)\s*(\/?)>/sg;
public static const stripRegex:RegExp = /<!--.*?-->|<\?(".*?"|'.*?'|[^>"']+)*>|<!(".*?"|'.*?'|[^>"']+)*>/sg;
The last three have to do with unsupported flags, but the first four break simply because the compiler evaluates the strings and they become spaces and line breaks, etc.
What can we do to prevent the compiler from killing patterns during compilation?
Harbs
Re: [FalconJX] regex patterns
Posted by Alex Harui <ah...@adobe.com>.
I put in some code to handle whitespace chars. Or do we need to handle
all chars in some unicode ranges?
-Alex
On 3/9/17, 10:42 AM, "Harbs" <ha...@gmail.com> wrote:
>Here’s one of them:
>
>static private const anyPrintChar:RegExp =
>/[^\u0009\u000a\u000d\u0020]/g;
>
>outputs:
>
>org.apache.flex.textLayout.elements.SpanElement.anyPrintChar = /[^
>
> ]/g;
>
>> On Mar 9, 2017, at 8:20 PM, Harbs <ha...@gmail.com> wrote:
>>
>> Yes. These are from TLF. The strings are evaluated by Java. There’s no
>>error except on the Javascript side because the string literals become
>>non-sensical.
>>
>> We had a similar issue when porting our app. I think we used
>>String.fromCharCode() to get around the problem there.
>>
>> Basically, we need some way to tell the compiler that any strings used
>>in RegExp objects should come through bit-for-bit identical and not be
>>evaluated as Strings. I have no idea what’s involved in doing that in
>>Java...
>>
>>> On Mar 9, 2017, at 7:43 PM, Alex Harui <ah...@adobe.com> wrote:
>>>
>>> Are these from TLF? What error are you getting? Falcon compiles TLF
>>>as
>>> part of the integration tests. Is it the JS output that is broken?
>>>
>>> Thanks,
>>> -Alex
>>>
>>> On 3/9/17, 1:18 AM, "Harbs" <ha...@gmail.com> wrote:
>>>
>>>> I’ve come across quite a few regex patterns which break compilation.
>>>> Here’s some examples:
>>>>
>>>> private static const _newLineTabPattern:RegExp =
>>>>/[\u0009\u000a\u000d]/g;
>>>> static private const brRegEx:RegExp = /\u2028/;
>>>> private static const _newLineRegex:RegExp = /\u000A|\u000D\u000A?/g;
>>>> public static const anyPrintChar:RegExp =
>>>>/[^\u0009\u000a\u000d\u0020]/g;
>>>> public static const attrRegex:RegExp =
>>>> /\s+(\w+)(?:\s*=\s*(".*?"|'.*?'|[\w\.]+))?/sg;
>>>> public static const tagRegex:RegExp =
>>>>
>>>>/<(\/?)(\w+)((?:\s+\w+(?:\s*=\s*(?:".*?"|'.*?'|[\w\.]+))?)*)\s*(\/?)>/s
>>>>g;
>>>> public static const stripRegex:RegExp =
>>>> /<!--.*?-->|<\?(".*?"|'.*?'|[^>"']+)*>|<!(".*?"|'.*?'|[^>"']+)*>/sg;
>>>>
>>>> The last three have to do with unsupported flags, but the first four
>>>> break simply because the compiler evaluates the strings and they
>>>>become
>>>> spaces and line breaks, etc.
>>>>
>>>> What can we do to prevent the compiler from killing patterns during
>>>> compilation?
>>>>
>>>> Harbs
>>>
>>
>
Re: [FalconJX] regex patterns
Posted by Harbs <ha...@gmail.com>.
Here’s one of them:
static private const anyPrintChar:RegExp = /[^\u0009\u000a\u000d\u0020]/g;
outputs:
org.apache.flex.textLayout.elements.SpanElement.anyPrintChar = /[^
]/g;
> On Mar 9, 2017, at 8:20 PM, Harbs <ha...@gmail.com> wrote:
>
> Yes. These are from TLF. The strings are evaluated by Java. There’s no error except on the Javascript side because the string literals become non-sensical.
>
> We had a similar issue when porting our app. I think we used String.fromCharCode() to get around the problem there.
>
> Basically, we need some way to tell the compiler that any strings used in RegExp objects should come through bit-for-bit identical and not be evaluated as Strings. I have no idea what’s involved in doing that in Java...
>
>> On Mar 9, 2017, at 7:43 PM, Alex Harui <ah...@adobe.com> wrote:
>>
>> Are these from TLF? What error are you getting? Falcon compiles TLF as
>> part of the integration tests. Is it the JS output that is broken?
>>
>> Thanks,
>> -Alex
>>
>> On 3/9/17, 1:18 AM, "Harbs" <ha...@gmail.com> wrote:
>>
>>> I’ve come across quite a few regex patterns which break compilation.
>>> Here’s some examples:
>>>
>>> private static const _newLineTabPattern:RegExp = /[\u0009\u000a\u000d]/g;
>>> static private const brRegEx:RegExp = /\u2028/;
>>> private static const _newLineRegex:RegExp = /\u000A|\u000D\u000A?/g;
>>> public static const anyPrintChar:RegExp = /[^\u0009\u000a\u000d\u0020]/g;
>>> public static const attrRegex:RegExp =
>>> /\s+(\w+)(?:\s*=\s*(".*?"|'.*?'|[\w\.]+))?/sg;
>>> public static const tagRegex:RegExp =
>>> /<(\/?)(\w+)((?:\s+\w+(?:\s*=\s*(?:".*?"|'.*?'|[\w\.]+))?)*)\s*(\/?)>/sg;
>>> public static const stripRegex:RegExp =
>>> /<!--.*?-->|<\?(".*?"|'.*?'|[^>"']+)*>|<!(".*?"|'.*?'|[^>"']+)*>/sg;
>>>
>>> The last three have to do with unsupported flags, but the first four
>>> break simply because the compiler evaluates the strings and they become
>>> spaces and line breaks, etc.
>>>
>>> What can we do to prevent the compiler from killing patterns during
>>> compilation?
>>>
>>> Harbs
>>
>
Re: [FalconJX] regex patterns
Posted by Harbs <ha...@gmail.com>.
Yes. These are from TLF. The strings are evaluated by Java. There’s no error except on the Javascript side because the string literals become non-sensical.
We had a similar issue when porting our app. I think we used String.fromCharCode() to get around the problem there.
Basically, we need some way to tell the compiler that any strings used in RegExp objects should come through bit-for-bit identical and not be evaluated as Strings. I have no idea what’s involved in doing that in Java...
> On Mar 9, 2017, at 7:43 PM, Alex Harui <ah...@adobe.com> wrote:
>
> Are these from TLF? What error are you getting? Falcon compiles TLF as
> part of the integration tests. Is it the JS output that is broken?
>
> Thanks,
> -Alex
>
> On 3/9/17, 1:18 AM, "Harbs" <ha...@gmail.com> wrote:
>
>> I’ve come across quite a few regex patterns which break compilation.
>> Here’s some examples:
>>
>> private static const _newLineTabPattern:RegExp = /[\u0009\u000a\u000d]/g;
>> static private const brRegEx:RegExp = /\u2028/;
>> private static const _newLineRegex:RegExp = /\u000A|\u000D\u000A?/g;
>> public static const anyPrintChar:RegExp = /[^\u0009\u000a\u000d\u0020]/g;
>> public static const attrRegex:RegExp =
>> /\s+(\w+)(?:\s*=\s*(".*?"|'.*?'|[\w\.]+))?/sg;
>> public static const tagRegex:RegExp =
>> /<(\/?)(\w+)((?:\s+\w+(?:\s*=\s*(?:".*?"|'.*?'|[\w\.]+))?)*)\s*(\/?)>/sg;
>> public static const stripRegex:RegExp =
>> /<!--.*?-->|<\?(".*?"|'.*?'|[^>"']+)*>|<!(".*?"|'.*?'|[^>"']+)*>/sg;
>>
>> The last three have to do with unsupported flags, but the first four
>> break simply because the compiler evaluates the strings and they become
>> spaces and line breaks, etc.
>>
>> What can we do to prevent the compiler from killing patterns during
>> compilation?
>>
>> Harbs
>
Re: [FalconJX] regex patterns
Posted by Alex Harui <ah...@adobe.com>.
Are these from TLF? What error are you getting? Falcon compiles TLF as
part of the integration tests. Is it the JS output that is broken?
Thanks,
-Alex
On 3/9/17, 1:18 AM, "Harbs" <ha...@gmail.com> wrote:
>I’ve come across quite a few regex patterns which break compilation.
>Here’s some examples:
>
>private static const _newLineTabPattern:RegExp = /[\u0009\u000a\u000d]/g;
>static private const brRegEx:RegExp = /\u2028/;
>private static const _newLineRegex:RegExp = /\u000A|\u000D\u000A?/g;
>public static const anyPrintChar:RegExp = /[^\u0009\u000a\u000d\u0020]/g;
>public static const attrRegex:RegExp =
>/\s+(\w+)(?:\s*=\s*(".*?"|'.*?'|[\w\.]+))?/sg;
>public static const tagRegex:RegExp =
>/<(\/?)(\w+)((?:\s+\w+(?:\s*=\s*(?:".*?"|'.*?'|[\w\.]+))?)*)\s*(\/?)>/sg;
>public static const stripRegex:RegExp =
>/<!--.*?-->|<\?(".*?"|'.*?'|[^>"']+)*>|<!(".*?"|'.*?'|[^>"']+)*>/sg;
>
>The last three have to do with unsupported flags, but the first four
>break simply because the compiler evaluates the strings and they become
>spaces and line breaks, etc.
>
>What can we do to prevent the compiler from killing patterns during
>compilation?
>
>Harbs