SEDIT record definitions should be in exports.all

[MattHeffron]

The record definitions needed to call some SEDIT functions are not in exports.all.
For example, to use SEDIT edit environments, my old code had to (in the COMS):

[DECLARE%: EVAL@COMPILE 
   (* ;; "This hackery is necessary since these DATATYPE definitions are not in EXPORTS.ALL")

   (P (for DT in '(SEDIT::EDIT-CONTEXT SEDIT::EDIT-ENV SEDIT::EDIT-NODE)
         do
         (EVAL (SYSRECLOOK1 DT))
         (UNMARKASCHANGED DT 'RECORD]

[masinter]

The way I've seen this done is (DECLARE%: ***@***.*** DONTCOPY (FILES (LOADCOMP) SEDIT-DECLS) in the COMS of the file that needs the record definition. (I"m not sure about the LOADCOMP vs just FILES. Also I'm not sure about SEDIT-DECLS vs SEDIT-EXPORTS. Is there a SEDIT-specific exports? https://LarryMasinter.net https://interlisp.org

…

On Mon, Jun 29, 2026 at 11:05 AM Matt Heffron ***@***.***> wrote: *MattHeffron* created an issue (Interlisp/medley#2680) <#2680> The record definitions needed to call some SEDIT functions are not in exports.all. For example, to use SEDIT edit environments, my old code had to (in the COMS): [DECLARE%: ***@***.*** (* ;; "This hackery is necessary since these DATATYPE definitions are not in EXPORTS.ALL") (P (for DT in '(SEDIT::EDIT-CONTEXT SEDIT::EDIT-ENV SEDIT::EDIT-NODE) do (EVAL (SYSRECLOOK1 DT)) (UNMARKASCHANGED DT 'RECORD] — Reply to this email directly, view it on GitHub <#2680?email_source=notifications&email_token=AAIQTK7WFVMD2U6HEBMKAFD5CKVWTA5CNFSL4Z3JMQ5C6L3HNF2C22DVMIXUS43TOVSS6NBXG4YDIOBTGY2DHJTSMVQXG33OVJZXKYTTMNZGSYTFMSSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIQTK3PXQJGR3HPYYJU2ZL5CKVWTAVCNFSNUABFKJSXA33TNF2G64TZHMZDQMJVGE4TSNRUHNEXG43VMU5TINZXGA2DQMZWGQZ2C5QC> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[MattHeffron]

For consistency, I think the declarations should be in exports.all, like for other system definitions.

SEDIT-EXPORTS is not related to exports.all; it is exporting symbols from the SEDIT package.

[rmkaplan]

In Tedit I have set up a systematic way of maintaining declarations among files that also allows the declarations to reside on the file with the other code that they are most associated with. This piggybacks on the generic EXPORT mechanism. I also have a systematic way of setting up an editing environment. You can look at the functions MAKE-TEDIT-EXPORTS.ALL UPDATE-TEDIT EDIT-TEDIT on TEDIT. Basically, all of the public-within-Tedit declarations are under EXPORT commands. They would go into the system EXPORTS.ALL if the Tedit files were on EXPORTFILES on FILESETS...but they aren't. So instead, when a declaration changes, I run (MAKE-TEDIT-EXPORTS.ALL) which copies all of the exports to a file TEDIT-EXPORTS.ALL. Anybody who wants to work with the internals of Tedit can load TEDIT-EXPORTS.ALL and get all of the exported declarations without loading any of the other Tedit files. (EDIT-TEDIT could load that too, but that isn't particularly necessary for someone to actually work on Tedit. Because EDIT-TEDIT does a LOADFROM and LOADCOMP of everything, whether exported or not.) Another convention is the separate file internal/TEDIT-DEBUG, which sets up the environment for editing but also brings in all sort of little widgets for showing and fiddling the various Tedit internal structures. If we started to work on SEDIT, that may be something else to set up.

…

On Jul 13, 2026, at 3:33 PM, Matt Heffron ***@***.***> wrote: MattHeffron left a comment (Interlisp/medley#2680) <#2680 (comment)> For consistency, I think the declarations should be in exports.all, like for other system definitions. SEDIT-EXPORTS is not related to exports.all; it is exporting symbols from the SEDIT package. — Reply to this email directly, view it on GitHub <#2680?email_source=notifications&email_token=AQSTUJL3T5YU4JDQRZMX25L5EVPTBA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIOJWGM2DEMZWGAYKM4TFMFZW63VKON2WE43DOJUWEZLEUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#issuecomment-4963423600>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJOBJF4DXI4JFHW5VRL5EVPTBAVCNFSNUABFKJSXA33TNF2G64TZHMZDQMJVGE4TSNRUHNEXG43VMU5TINZXGA2DQMZWGQZ2C5QC>. You are receiving this because you are subscribed to this thread.