"Tony Lewis" <[EMAIL PROTECTED]> writes: > Hrvoje Niksic wrote: > >> This patch imposes IMHO reasonable, yet safe, limits for reading server >> responses into memory. > > Your choice of default limits looks reasonable to me, but shouldn't > wget provide the user a way to override these limits?
Not if there is never a need to. Wget has a huge number of command-line options already; one should try not to add new ones without very good reasons to back them up. Overriding these limits would require *two* new cryptic command-line options that would clutter the code and documentation and in all likeliness would never be used, thus bringing no value to the user. Except for malicious server output, I see no use in accepting lines greater than 4k or HTTP response headers larger than 64k. If I am proven wrong, the limit can be raised, but I suspect that it will not be the case. (Note that the limit can safely be raised without any penalty for the usual, "short", responses.)
