Loadup-all.sh does not call Loadup-db.sh but copy-all.sh tries to copy the fuller.sysout and fuller.database files

[fghalasz]

Loadup-all.sh does not call Loadup-db.sh but copy-all.sh tries to copy the fuller.sysout and fuller.database files

[masinter]

I wanted to release fuller.sysout too. loadup-db.sh takes 5 minutes. Is that long?
Does it matter?

[rmkaplan]

It isn’t long when building a release, but I wouldn’t want to see it in the loadup-all script that I use quite often in building my own systems. The things I’m working on get analyzed incrementally, the other things change very slowly as PR’s are merged. I run the loadup-db scripts maybe once a month.

…

On Jul 22, 2023, at 10:32 PM, Larry Masinter ***@***.***> wrote: I wanted to release fuller.sysout too. loadup-db.sh takes 5 minutes. Is that long? Does it matter? — Reply to this email directly, view it on GitHub <#1289 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJP2ESWFJO32U2USQCTXRSZPRANCNFSM6AAAAAA2UIRVNM>. You are receiving this because you are subscribed to this thread.

[masinter]

I think I will continue to do my own loadups as I'm doing stuff.
@rmkaplan if you're happy doing your own loadups... ? I thought these changes were for you.

[rmkaplan]

This was about the expected/natural configuration of the release sysouts, not about sysouts made as part of internal development work. The release could have a slightly different version of the “all” that makes the full database and copies it over. I don’t think that copy-all runs for a local tmp build.

…

On Jul 23, 2023, at 8:09 AM, Larry Masinter ***@***.***> wrote: I thought these changes were for you.

[masinter]

scripts/loadup-all.sh does a copy-all.sh at the end. This keeps us from having to add tmp to LOADUPSDIRECTORIES

[rmkaplan]

Hmm, I didn’t know that, I thought quite some time ago we got rid of loadups/ in local set ups. Maybe it just went into .gitignore. I’ve been running all of my local stuff out of tmp/, not depending on the copy. Seems like we (or at least I) need some build scripts with a slightly different breakdown compared to the ones that the release process uses.

…

On Jul 23, 2023, at 8:55 AM, Larry Masinter ***@***.***> wrote: scripts/loadup-all.sh does a copy-all.sh at the end. This keeps us from having to add tmp to LOADUPSDIRECTORIES — Reply to this email directly, view it on GitHub <#1289 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJICU5ZDQ3OUTIKT6TDXRVCOFANCNFSM6AAAAAA2UIRVNM>. You are receiving this because you were mentioned.

[rmkaplan]

BTW, the .gitignore lists the different files in loadups/ separately. Shouldn’t it just say loadups/* ?

…

On Jul 23, 2023, at 9:03 AM, Ron Kaplan ***@***.***> wrote: Hmm, I didn’t know that, I thought quite some time ago we got rid of loadups/ in local set ups. Maybe it just went into .gitignore. I’ve been running all of my local stuff out of tmp/, not depending on the copy. Seems like we (or at least I) need some build scripts with a slightly different breakdown compared to the ones that the release process uses. > On Jul 23, 2023, at 8:55 AM, Larry Masinter ***@***.***> wrote: > > > scripts/loadup-all.sh does a copy-all.sh at the end. This keeps us from having to add tmp to LOADUPSDIRECTORIES > > — > Reply to this email directly, view it on GitHub <#1289 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJICU5ZDQ3OUTIKT6TDXRVCOFANCNFSM6AAAAAA2UIRVNM>. > You are receiving this because you were mentioned. >

[masinter]

if you look in loadups there are these files that aren't ignored:
README.md - explain what is in loadups
starter.sysout -- stable Interlisp for makeinit to run in
lisp.venuesysout full.venuesysout -- oldest medley 3.5 sysouts we have. Could be moved to obsolete.

[fghalasz]

So I propose that we fully separate loadup-all.sh and loadup-db.sh (maybe renamed to loadup-fuller.sh).

loadup-all.sh just does up to full.sysout including calling copy-all.sh without any of the fuller files.

loadup-db.sh creates the fuller files and copies them to loadups.

That way Ron can continue with loadup-all.sh for everyday work and call loadup-db.sh when needed.

From the github actions and release-medley.sh, we can call loadup-all.sh followed by loadup-db.sh so that the fuller stuff is included in the releases (if that's what we want).

[rmkaplan]

Seems reasonable.

…

On Jul 23, 2023, at 12:40 PM, Frank Halasz ***@***.***> wrote: So I propose that we fully separate loadup-all.sh and loadup-db.sh (maybe renamed to loadup-fuller.sh). loadup-all.sh just does up to full.sysout including calling copy-all.sh without any of the fuller files. loadup-db.sh creates the fuller files and copies them to loadups. That way Ron can continue with loadup-all.sh for everyday work and call loadup-db.sh when needed. From the github actions and release-medley.sh, we can call loadup-all.sh followed by loadup-db.sh so that the fuller stuff is included in the releases (if that's what we want). — Reply to this email directly, view it on GitHub <#1289 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQSTUJOAIE2S727UDZW5DYDXRV43HANCNFSM6AAAAAA2UIRVNM>. You are receiving this because you were mentioned.

[masinter]

@fghalasz I'm not sure in your scheme when the copy from tmp to loadups happen.
@rmkaplan It would simplify thigs if we stpped using tmp and just writing the files as they are made into loadups. We could gitignore loadups if we moved starter.sysout and loadups/README.md.

[fghalasz]

The copy to loadups from tmp happens within each of the scripts (loadup-all and loadup-db) for the files that are created within that script.

[nbriggs]

tmp/RDSYS and tmp/RDSYS.LCOM also need to be copied into library/ as part of the build. They are what give you "remote" access to a vmem files that was saved by uRaid's "v" command.

[masinter]

I think these issues have settled as far as matching the working style for a few of us, but we should start a "briefing blurb" that talks about how this whole thing holds together.

Once it's been stable a while.

[masinter]

this was a long discussion and I think we've have something now that makes this discussion not so interesting any more.