Moving the discussion to xemacs-beta, per policy. Thus the full-body quote. SJT> VETO SY> You're a bit late, I had already approved and committed this SY> to 21.5 ages ago. I suspected that. Based on the date, I was using a snorkel to breathe through the midterms (and 21.4.3) rather reviewing "feature" patches carefully. Oh, well. It probably won't do any harm, but is that really the way we want to run this railroad? Core lisp should not try to cater to user vagaries (space-saving fanatic or not, uses jka-compr or not) IMHO. SJT> 2001-05-13 Karl M. Hegbloom <karlheg(a)hegbloom.net&gt; SJT> * packages.el (locate-data-file): Use suffixes list to allow SJT> finding compressed copies of the data files. SJT> Karl, SJT> Sorry about the long delay. SJT> I think this is a bad idea. I'm not going to apply to 21.4, SJT> and I recommend that it not be applied to 21.5, either. SJT> (1) In your implementation, you left out the null suffix. SJT> That means an uncompressed file won't be found. SY> I added the null suffix before I committed. SJT> (2) You left out at least two common suffixes (".Z" and SJT> ".zip"), and there may be others (".lha", ".hqx") to worry SJT> about on other platforms. SY> But these can be added at a later time. People who use these SY> types of compressed files will be no worse off than they are SY> now. SJT> (3) It's probably a bad idea to return compressed files, as SJT> not all XEmacsen will know what to do with them, having SJT> potentially bad consequences. SY> But can't the XEmacsen that this patch was applied to handle SY> compressed files? As far as I know 21.1, 21.2, 21.4, and 21.5 SY> can handle it and the patch isn't being applied anywhere else. SY> What I'm saying is that is someone has an XEmacs that pukes on SY> compressed files, they aren't gonna have this patch anyway. No, AFAIK no XEmacs automagically handles compressed files _until jka-compr is enabled_. That means that code that does something like (load (locate-data-file "library.el")) will be badly hosed. (And don't tell me that no one in their right mind would load Lisp code that way; there will be similar examples that make more sense in practice if you put your mind to it.) SJT> (4) This convenience really would be most useful SJT> interactively, but the function in question isn't SJT> interactive. SY> Again, this can be added at a later time. The function not SY> being interactive doesn't make it unusable. Perhaps a little SY> harder to take advantage of the added functionality. The point here is that a _user_ will immediately recognize that something is wrong if a buffer is filled with raw gzip. Code will not, and will happily try to parse it with who knows what results. SY> I have no problem with any of these, although I can't think SY> right now how to go about doing (3). Check for whether jka-compr is loaded. I think featurep will do, although it may be possible to turn off jka-compr which would leave the feature in place. SY> Leave this patch in 21.5, don't commit to 21.4 or 21.1. I think it's risky for the reasons above. -- University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091 _________________ _________________ _________________ _________________ What are those straight lines for? "XEmacs rules."