Re: [PATCH] sloppy mime parsing patch

On Tuesday, April 8, 2003, at 10:55 AM, Stipe Tolj wrote:

> Aarno Syvänen wrote:
>>
>> Hi List,
>>
>> This is patching changing Kannel mime parsing so that it
>> accepts lf instead of crlf as a separator, but only if this is
>> configured (strict-parsing=no).
>>
>> I send the patch, because Paul vetoed the first one. Now
>> he can check does this one work properly.
>>
>> Patch has three parts:
>>
>> a) Kannel read configuration directive strict-parsing and
>>      parses mime multipart accordingly.
>> b) test_ppg accepts binary content  and sends mime
>>      content with different separators (crlf or lf).
>> c) There are two new octstr functions.
>
> Aarno the code delivers the "strictness" via several levels of
> function calls. Wouldn't it have been a lot of easier to pro-claim
> this as global variable (like ie. my smart-error message variable) and
> simply access it whereever the if condition is really needed?
>
> I think this would reduce the code changes drastically.
>
> Intention is that function parameters sould be semantically bind to
> the logic inside the function. If parameters are passed within several
> level, there is no semantical use of it within the "internal"
> functions and hence the only thing you do is to trasport a global
> scoped variable from a very high'ed to low'ed function call level.
>
> I hope this was not writen in too wired ways ;)

Another point of coding style. I wanted mime_parse have input
variable telling is parsing sloppy or strict. I think this kind of thing
should be *very* transparent, there are actually two different parsing
functions here.

Aarno