ulmo@ssyx.ucsc.edu (06/08/89)
These impressions are some that I have had of the Emacs editor so far. Does anyone have any patches or modifications for GNU Emacs which will make it case insensitive in all appropriate locations? (I think it's stupid Unix and C are case sensitive at all, and have often wondered how easy it would be to fix that on a system of my own). more comments: - Using ln to link a file and then edit it, Emacs will save the old version associated with the same old link, and the file will end up with a new inode, link structure, file descriptor, directory entry, etc. This isn't always intended or understood by the user (me), although I can see benefits and hardships both ways. - Versions. How easy is it to ask Emacs to use a new name for each version of a file I make (ITS-wise, where each version gets something.83383, or something better dealing with more intricate version #s would be nice too) and deal with it appropriately? I guess this is a fundamental problem with Unix: it calls many things Files (a concept I am absolutely sick of), and doesn't even bother to put file type information anywhere -- the file is anyone's guess as to what it is (so common practices are to put file type information in the calling environment (an unportable method) or in the file itself (something which gets in the way of data representation). Bleah. - Multitasking. Arggh! Why do I have to be stuck with one command line in an otherwise fairly powerful system? Why must I sit waiting for one command to finish for me to figure out what to type for another, or even ^Z out for god's sake? The worst thing is wanting to look something up, and realizing that there is no way to do that. - Incremental compilation (automatic). What's this thing about me having to manually type byte-compile? I guess I just haven't used Emacs for 20 years and gotten used to The Way it Should Be, so some of this is not exactly right on target. Also, is there a binary mode, complete with various Hex and Ascii translation guides? This would greatly simplify the task of people like me who like to look and modify a file in its full 8 bits. Bleah, it's late, but I think this all makes some sense.
tale@pawl.rpi.edu (David C Lawrence) (06/08/89)
In article <1989JE8.090959,4635;ULMO@SSYX.UCSC.EDU> ulmo@ssyx.ucsc.edu writes: > - Using ln to link a file and then edit it, Emacs will save the old > version associated with the same old link, and the file will end up > with a new inode, link structure, file descriptor, directory entry, > etc. This isn't always intended or understood by the user (me), > although I can see benefits and hardships both ways. I've been meaning to look at this one and see how the link could be preserved. The fellow who handles the /export partitions on our primary fileservers has exactly the same problem with the ring of files he has ln'ed. He wants to edit just one of the files (any file) and have the link preserved so that all of the others see the change. He prefers to use Emacs but has taken to vi'ing these files because it will preserve the link. Somewhat hand-in-hand with this is the file gets written out with new owner information (last time I tried this was a few months ago; we've been avoiding the problem in the meantime). % ls -l ~X/README -r--r--r-- 1 X 4258 Oct 26 1988 /appl/imagine1/X/README % sudo emacs ~X/README [change made, revision saved] % ls -l ~X/README -r--r--r-- 1 root 4258 Oct 26 1988 /appl/imagine1/X/README Perhaps a quick fix here would be for the original file owner to be noted and then have the file chown'ed back. This solution doesn't work, however, for the case where I am in the same group as a file owner and I edit a mode 664 file. Once again, vi happens to work as expected under these circumstances. Dave -- (setq mail '("tale@pawl.rpi.edu" "tale@itsgw.rpi.edu" "tale@rpitsmts.bitnet")) "I realize the Internet isn't the whole world, but it is the center of it." -- Greg Woods
lamy@ai.utoronto.ca (Jean-Francois Lamy) (06/08/89)
In article <TALE.89Jun8105435@imagine.pawl.rpi.edu> tale@pawl.rpi.edu writes: >I've been meaning to look at this one and see how the link could be >preserved. The fellow who handles the /export partitions on our >primary fileservers has exactly the same problem with the ring of >files he has ln'ed. He wants to edit just one of the files (any file) (setq backup-by-copying t) in your .emacs will solve that. Using hard links to share /export (I assume you are talking about SunOS4.x) is a dangerous proposition for that very reason. If emacs and vi can be made to do the "right" thing, g*d knows what all the other utilities will do with hard links. Jean-Francois Lamy lamy@ai.utoronto.ca, uunet!ai.utoronto.ca!lamy AI Group, Department of Computer Science, University of Toronto, Canada M5S 1A4
jr@bbn.com (John Robinson) (06/09/89)
In article <1989JE8.090959,4635;ULMO@SSYX.UCSC.EDU>, ulmo@ssyx writes: >Does anyone have any patches or modifications for GNU Emacs which will make it >case insensitive in all appropriate locations? case-fold-search for searches. file lookup will have to retain case sensitivity - it's in the filesystem. Not usre what other locations you have in mind. > (I think it's stupid Unix and >C are case sensitive at all, and have often wondered how easy it would be to >fix that on a system of my own). Maybe in GNU >- Using ln to link a file and then edit it, Emacs will save the old version Variables backup-by-copying, backup-by-copying-when-linked, backup-by-copying-when-mismatch should get you what you want. >- Versions. How easy is it to ask Emacs to use a new name for each version > of a file I make The variable is version-control >- Multitasking. Arggh! Why do I have to be stuck with one command line > in an otherwise fairly powerful system? Run a job-control capable shell in a shell window; explicit-shell-file-name variable. & at the end of a command line puts it in background. >- Incremental compilation (automatic). What's this thing about me > having to manually type byte-compile? Function byte-compile-directory will help. put dot after the defun and type ^X^E to redefine it, byt there is no "byte-compile this form and then eval it", though this could be consed up from what's there I think. >I guess I just haven't used Emacs for 20 years and gotten used to >The Way it Should Be, so some of this is not exactly right on target. It is still gonna be a few years until the Unix world catches up to where ITS, Tenex, Multics and Tops-20 were 15-20 years ago! Look to the FSF folks to maybe do this with GNU Unix, if it ever materializes. >Also, is there a binary mode, complete with various Hex and Ascii >translation guides? This would greatly simplify the task >of people like me who like to look and modify a file in its full 8 bits. emacs can edit anything, though it can be a bit cumbersome. There is a mode that makes it a little easier - it generates a ascii picture of the file in hex which you edit, then it turns the ascii back to binary when you write it out (I guess). Contact Keith Gabryelski, <ag@elgar.cts.com>. You'll become a believer sooner or later. -- /jr, nee John Robinson What a waste it is to lose one's mind--or not jr@bbn.com or bbn!jr to have a mind. How true that is. -Dan Quayle