perm filename EBUG18.MSG[E,ALS] blob sn#125795 filedate 1974-10-18 generic text, type T, neo UTF8
∂18-OCT-74  0037		R,LES
 In E, the "-FORM" command didn't get me to the bottom of the preceding page,
 but about 6 lines from the bottom.

∂18-OCT-74  0031		R,LES
 In E, the command ⊗XJFILL doesn't quite do the right thing.
 It preserves up to 2 spaces following a comma, rather than trimming
 to one space.  Of course, 2 spaces should be retained following a period.

∂17-OCT-74  2157		105,SGK AT TTY40   2157
 Please tell me, if you know, where the SEGNAM is set.

∂17-OCT-74  1947		105,SGK
 Guess what.  I can reproduce BUG IN WRITE CODE.  If you are interested in
 it I will be happy to show it to you.
␈ CC: ALS;REG

∂17-OCT-74  1854		105,SGK AT TTY40   1854
 The way E was handling VTs in the past works better than the current, which
 displays one thing, and does another.  As I have said before VT is not a command
 The command people want is <Control>VT which is the exact oposite of Return.
 I don't think VT was ever advertized as the inverse of Return.  Now VT inserst
 itself as it should (when you are in the line editor).  But by activating on
 it, you ruin the users display of his line.

∂17-OCT-74  1220		H,RF
 If a paragraph begins with a tab, but has no preceeding blank line,
 the the E command ⊗XJU 4,1,30 will give it the wrong number of spaces.
 If a line starts with tab and has a preceeding blank line, something
 weird happens:  The line seems to get extra spaces at front, but
 as soon as you enter the line (with α<sp>, for example), some of
 the text disappears.  It feels like a case of dangling pointers.

∂17-OCT-74  1157		H,RF
 Bugs in E:
 	⊗5⊗A⊗-⊗7⊗C causes a halt at user 406557 and the message
 	UNRECOGNIZED CONTROL CHARACTER -- <CTR><MET> C
 
 In the TEACH file, page 19, the information about ⊗XJFILL is wrong.
 
 In attach mode (perhaps also out of it) the command ⊗XJF 3<CR>
 refuses to word until I type two more <cr>s, and then does the wron
 thing, as if I had typed ⊗xjf ,,3<cr>
 

∂10-OCT-74  1214		FOL,REF
 Suggestion for E:
 	How about a /Quiet switch when invoking the editor, which represses
 the "NEED TO REFORMAT" and "INVALID DIRECTORY" questions and answers.
 	Thus , "ET FOO/N/Q" would edit FOO without a directory page, and
 without asking if the page should be removed, and "ET FOO/Q" would create
 a directory page for FOO if it didn't have one, without asking.
 
 	Also, a ⊗XSQUASH command, which accomplishes the same effect
 as is presently gotten by the sequence ⊗E ET FOO/N <cr> N ⊗E ET FOO <cr> Y
 
 i.e. -- reformatting the file to take up a minimum of space.
 
 	Further suggestion.  A command which would "move the underbar
 up and down" (exit from the line, then α<space> as many spaces into the
 line as we were when we invoked the command) would help in the production
 of a lot of similiar lines.  If this notion were to be carried further, the
 ability for the user to string together a series of E commands, give it
 a name (using the unused control characters) would be a simple but
 effect macro facility.
 
 	I'd appreciate hearing from you as to the feasibility and
 desirability of these suggestions.  Thanks.
 			Bob Filman 

∂09-OCT-74  1511		1,BH
 If the command ET FOO←BAR finds file BAR.XYZ it should create FOO.XYZ

∂08-OCT-74  1331		ACT,REG
 αβAαβ-αβC crashes E with an illegal uuo at 105.  Perhaps you could fix it?

∂03-OCT-74  0737		105,SGK
 The following bug has been introduced into E:

 When editing a line (it seems only a line which does not exist on DSK yet)
 typeing ↑. <Control Point> does not re enter the line editor properly.
 Normally, the line you were editing would be put back in the line editor,
 and your pointer would positioned as it was before.  I.E. ↑. has no
 noticable effect except that it leaves the line editor for a moment.
 At this time, ↑. seems to be reloading the line editor as would ↑XCAN <The extented
 command CAN>  More specifcally, ↑.'ing on a line which did not exist in
 the copy of the file on the DSK causes the line editor to be loaded with
 a copy of the TOP LINE(!) slightly modified.  
 	An easy way to see this happen is to create a file, type a few
 lines of text with different things in them so you can tell the difference
 between them.  Select a line which is not the top one, start line editing
 it and type ↑.  You will notice that the line editor will be restored with
 the top line of your file.
 	If this is not clear, let me know I will show it to you.

∂30-SEP-74  1053		1,BH
 Idea for E: have a mode in which αF etc. searches ignore text between
 semicolon and crlf, so it only finds relevant instances of symbols.