Comment 17 for bug 311818
Revision history for this message
| poy (poy) wrote Re: case-sensitivity in Unix systems | #17 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
1b1ed1a
(Get the code!)
almost everything in this patch is wrong. in ascending order of importance:
- the naming of "stricmps" & "strnicmps" doesn't make sense; it sounds like you just wanted "strcmp" & "strncmp".
- i don't understand the whole CaseStringHash / CaseStringEq deal. why not just let the STL handle it?
- searching is still case insensitive, yet the bloom table has been made case sensitive; weird...
- and to the actual issue. it is beyond me why after all the discussion and rebuttals, you still went out of your way to re-propose Windkracht8's patch of producing invalid file lists.
you seem to be discussing an "artificial restriction"; as i wrote in my previous comment, this is not the point of this thread. you should have been arguing about this months ago when ADC was still being established. now it is here and can't be hacked around in such unintelligent ways. up until now, no client has had the guts of implementing the crazy idea of producing invalid file lists, so there is probably not much validation going on in client & bot software around; but don't be surprised if after your file lists get in the wild, such validation mechanisms do appear, rendering your file lists essentially useless.
it looks like you are willing to go with the 4th of my possibilities, which was ammending the "recommendation" file. properly expose your arguments to the ADC managers and hope they will side with you, then. but just making your client invalid without proper consulting, and daring to ask for the same changes to be brought over other clients, shows either a lack of comprehension or great insolence.
furthermode, i have enumerated several possibilities that could work, with the last one being my favorite as it involved no protocol-level change. it even got picked up again by Jacek in his last comment. why not implement it instead?