Issue80

classification
Title Lost docstrings in 21.5
Type defect Module core code 21.5
Severity inelegant Platform N/A
Keywords
explanation
process
These optional controls are only of interest to committers and tracker administrators.
Status done/needs work   Reason fixed
Superseder   Submitted 2007-12-04.08:08:34
Priority normal   Assigned To
Nosy List aidan  

Created on 2008-01-19.04:58:04 by stephen, last changed 2008-05-13.11:58:26 by aidan.

Messages
msg744 [hidden] ([hidden]) Date: 2008-05-13.11:58:26
This was fixed by
http://hg.debian.org/hg/xemacs/xemacs?cs=15d36164ebd7 of
2007-12-09, which moves the declarations to top level and
uninterns the corresponding symbols at dump time if the
necessary features are not available.

Yes, it *might* be nice to modify make-docfile to understand
Lisp.
msg265 [hidden] ([hidden]) Date: 2008-01-19.06:43:06
  Message-ID: <18261.31456.373689.367546@parhasard.net>
Ar an ceathrú lá de mí na Nollaig, scríobh Aidan Kehoe: 

 >  Ar an ceathrú lá de mí na Nollaig, scríobh Stephen J. Turnbull: 
 > 
 >  > Aidan, looks like you touched this stuff last:
 >  > 
 >  > Finding pointers to doc strings...
 >  > Finding pointers to doc strings...done
 >  > Warning: doc lost for function unicode-error-translate-region.
 >  > Warning: doc lost for function frob-unicode-errors-region.
 >  > Warning: doc lost for function add-menu-item.
 >  > Warning: doc lost for function add-menu.
 >  >
 >  > I don't see what the problem is, anybody know?
 > 
 > The problem is that all four functions (and a few more defvars that the
 > warning isn’t picking up) are within (when (featurep FEXP) ...) clauses.
 > This means that when make-docfile.c looks through the ELC code, its
 > generally stupid matching sees a (byte-code ...) for the entire form instead
 > of separate #'defalias and #'defvar forms for each function and variable. 
 > 
 > One way to eliminate the error would be to move the featurep check at
 > runtime, and throw an 'unimplemented error if it’s not available.

Another way that wouldn’t require modifying any functions (though it would
require modifying defvar and defconst calls) would be to unintern the
symbols at the end of the file if the feature isn’t available.
msg98 [hidden] ([hidden]) Date: 2008-01-19.04:58:04
  Message-ID: <87ir3dwqr4.fsf@uwakimon.sk.tsukuba.ac.jp>
Aidan Kehoe writes:

 >  > One way to eliminate the error would be to move the featurep check at
 >  > runtime, and throw an 'unimplemented error if it's not available.
 > 
 > Another way that wouldn't require modifying any functions (though it would
 > require modifying defvar and defconst calls) would be to unintern the
 > symbols at the end of the file if the feature isn't available.

Another way would be to take up the comment in make-docfile.el and
rewrite the whole thing in Lisp.  (Not suggesting that at the present
time, just a reminder that maybe we should do something about the
make-docfile mess.)

I wonder if it wouldn't be possible to get the byte-compiler to do it
for Lisp.

_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta@xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
msg97 [hidden] ([hidden]) Date: 2008-01-19.04:58:04
  Message-ID: <877ijuyj99.fsf@uwakimon.sk.tsukuba.ac.jp>
Aidan, looks like you touched this stuff last:

Finding pointers to doc strings...
Finding pointers to doc strings...done
Warning: doc lost for function unicode-error-translate-region.
Warning: doc lost for function frob-unicode-errors-region.
Warning: doc lost for function add-menu-item.
Warning: doc lost for function add-menu.

I don't see what the problem is, anybody know?

_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta@xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
History
Date User Action Args
2008-05-13 11:58:26aidansetstatus: chatting -> done/needs work
severity: inelegant
platform: + N/A
nosy: + aidan
messages: + msg744
module: + core code 21.5
priority: normal
assignedto: aidan
reason: fixed
type: defect
2008-01-19 06:43:06aidansetmessages: + msg265
2008-01-19 04:58:04stephensetstatus: new -> chatting
messages: + msg98
2008-01-19 04:58:04stephencreate