https://sourceforge.net/p/objectivelib/feature-requests/2005-12-06T20:44:27ZRecent changes to feature-requests2005-12-06T20:44:27Z2005-12-06T20:44:27ZWill Masonhttps://sourceforge.net/u/userid-25745/https://sourceforge.net734f0774247e1bb7a966bf38d0250da8c85478cc

<div class="markdown_content"><p>Look into building ObjectiveLib on the Power5 server at<br /> the SF compile farm. It runs SuSE, so it shouldn't be<br /> all that difficult.</p></div>

2005-10-17T19:59:50Z2005-10-17T19:59:50ZWill Masonhttps://sourceforge.net/u/userid-25745/https://sourceforge.net20fc8f2f28464603bafa0b5e3817589b0024d5e2

<div class="markdown_content"><p>A full crypto subsystem can be created. The Java crypto<br /> API can be used as a model, while the EVP library from<br /> OpenSSL can provide the crypto implementation.</p></div>

2005-10-16T23:03:23Z2005-10-16T23:03:23ZWill Masonhttps://sourceforge.net/u/userid-25745/https://sourceforge.net77635ad5fd65bd1f15786db339d0de559e79ed06

<div class="markdown_content"><p>Probably the best bet is to use the EVP facility of<br /> OpenSSL. This will be widely available.</p></div>

2005-08-08T19:06:29Z2005-08-08T19:06:29ZWill Masonhttps://sourceforge.net/u/userid-25745/https://sourceforge.net380078d45dabbcc0d82c9d28b4d5d08ac58acc0a

<div class="markdown_content"><p>Frequently used functors that tend to get created and<br /> thrown away a lot in OLAlgorithm should be cached. The<br /> most frequently used are OLLess and OLEqualTo.</p></div>

2005-07-22T21:18:45Z2005-07-22T21:18:45ZWill Masonhttps://sourceforge.net/u/userid-25745/https://sourceforge.net96b2f9d3efc9c823f648485becadb66f0c49660c

<div class="markdown_content"><p>A Cocoa/GNUstep-free version would be nice. Certain<br /> things could be conditionalized out of the build, like<br /> all the NSCoder stuff. Other stuff, like memory<br /> management, would be harder, but with the already<br /> consistent use of the memory management macros,<br /> hopefully it wouldn't be too painful.</p></div>

2005-07-12T00:17:05Z2005-07-12T00:17:05ZWill Masonhttps://sourceforge.net/u/userid-25745/https://sourceforge.net4f4d4be812e3b8fabae0f2b5e107e669a5c635b6

<div class="markdown_content"><p>Nuff said.</p></div>

2005-07-11T19:16:47Z2005-07-11T19:16:47ZWill Masonhttps://sourceforge.net/u/userid-25745/https://sourceforge.neta63aa4505ced554c309e4bc2efbc7f60e3bf655f

<div class="markdown_content"><p>Header guards will get rid of duplicate interface<br /> warnings on platforms that generate them. This includes<br /> all platforms except for NetBSD and Mac.</p></div>

2005-05-18T04:13:52Z2005-05-18T04:13:52ZWill Masonhttps://sourceforge.net/u/userid-25745/https://sourceforge.net9cf37bc678f69ce2bba92190223e6399ab0a2800

<div class="markdown_content"><p>A formatted text stream can be subclassed off of<br /> OLObject(Ou|In)Stream and can be implemented fairly<br /> easily. Two new methods can be added writeString: and<br /> writeFormat: that should handle everything. This will<br /> be easy. Famous last words...</p></div>

2005-03-10T16:32:49Z2005-03-10T16:32:49ZWill Masonhttps://sourceforge.net/u/userid-25745/https://sourceforge.net9ce6ede62a2e1ba198e87d73ce326bfe6dd06bda

<div class="markdown_content"><p>Since enumerators in Cocoa do a [[retain] autorelease]<br /> on objects when you return them, it needs to be clear<br /> in the OL docs that in OL iterators never, ever have<br /> ownership of the containers or of the items in them.<br /> There's been a lot of discussion of this in the<br /> cocoa-dev mailing list, and people seem to think that<br /> Cocoa's way of handling it is dumb. I agree.</p></div>

2005-02-27T02:37:29Z2005-02-27T02:37:29ZWill Masonhttps://sourceforge.net/u/userid-25745/https://sourceforge.nete289a9c8200756a60327ac84ee712ff0ef77ac5f

<div class="markdown_content"><p>A make command-line variable like "embeddable=yes"<br /> should be added to tweak the framework build so that it<br /> becomes embeddable. Build instructions should be modified.</p> <p>Option: When binary releases are created, make a new<br /> embeddable binary release.</p></div>